THE INFLUENCE OF WORK STRESS AND WORK SUPPORT ON BURNOUT

IN PUBLIC HOSPITAL NURSES

Rebecca Spooner-Lane, BBehSc, B(Hons)Psy

Thesis submitted in
fulfilment of the requirements
for the award of the Degree of
Doctor of Philosophy.
July 2004

School of Learning and Professional Studies

Faculty of Education
Queensland University of Technology
ii
 iii

KEYWORDS

Burnout, emotional exhaustion, depersonalisation, personal accomplishment, work

stress, work stressors, lack of professional recognition and support, patient care
uncertainty, job conditions, interpersonal conflict, role overload, role conflict, role
ambiguity, social support, work support, emotional support, instrumental support,
coworker support, supervisor support, nurses.
 iv

ABSTRACT

Lazarus and Folkman’s (1984) transactional stress-strain-coping theory

provides the theoretical background for the present thesis. This theory proposes that

strain (i.e., burnout) occurs when demands (i.e., work stressors) exceed coping

resources (e.g., social support). The current thesis explores the influence of social

support on the stress-burnout relationship in nurses. A sample of Australian nurses

working across three public hospitals in Brisbane’s metropolitan district were

recruited to investigate the nature and level of burnout experienced by nurses.

Burnout is characterised by emotional exhaustion, depersonalisation and reduced

personal accomplishment. The present research addresses gaps in the empirical

literature by investigating the key work stressors experienced by Australian nurses

and by establishing nurses’ referent levels of work stress, social support, and burnout.

In addition, the research explores the complex relationships between work stress,

social support and burnout.

The majority of nursing studies have failed to consider how support from

within the nurses’ work environment mitigates burnout. The present research builds

upon previous nursing literature by examining the ‘main’ and ‘buffering’ effect

hypotheses. Studies have consistently found support for the main effect model,

however the hypothesis that social support buffers the negative effects of stress has

resulted in highly conflicting findings. Some theorists (Cohen & Wills, 1985; Cutrona

& Russell, 1990) propose that the buffering effects of social support will only be

found if there is an adequate match between the needs elicited by the stressful event

and the type of support an individual receives. The present study extends the stressor-

support matching theory by exploring the matching between specific types of

 v

stressors to specific types (i.e., emotional and instrumental) and sources (i.e.,

supervisor and coworkers) of support. Cutrona (1990) suggests that the controllability

of a stressor is the primary dimension in determining an appropriate match between

stressors and types of support. Cutrona proposes that controllable stressful events

elicit needs for instrumental support and uncontrollable events elicit needs for

emotional support. Heeding Cutrona’s advice, independent raters classified nurses’

work stressors as controllable or uncontrollable stressful events prior to investigating

the stressor-support matching theory.

Three sequential studies were undertaken to explore the variables of interest to

this research program. In Study 1, focus groups were conducted with 68 nurses (11

males, 34 females) from two public hospitals. The qualitative data was subjected to

content analysis. The findings revealed that Australian nurses are exposed to a range

of job-specific stressors (i.e., Job Conditions, Job Uncertainty, Interpersonal Conflict

and a Lack of Professional Recognition and Support) and generic role stressors (i.e.,

Role Overload, Role Conflict and Role Ambiguity). The findings prompted the

research to utilise Wolfgang’s (1988) Health Professions Stress Inventory to measure

nurses’ job-specific stressors and Osipow and Spokane’s (1987) Occupational Roles

Questionnaire to measure nurses’ role stressors in Study 2.

The findings from Study 1 also confirmed that the way nurses perceive work

support is consistent with current social support literature. Nurses indicated that their

two main sources of support were their coworkers and their supervisor. Furthermore,

nurses discussed social support from a multidimensional perspective, recognising

different types of support that were broadly classified as emotional and instrumental

support. Based on these findings, the researcher developed a work support measure

specifically for the purpose of this research. Items were taken from established social
 vi

support scales and were slightly modified to ensure that they were contextually

relevant to nurses.

In Study 2, 273 nurses (38 males, 235 females) completed a multi-measure

questionnaire. While there was sufficient research evidence to indicate that the

Occupational Roles Questionnaire (Osipow & Spokane, 1987) and the Maslach

Burnout Inventory (Maslach, Jackson, & Leiter, 1996) possess adequate levels of

reliability and validity, less was known about Wolfgang’s Health Professions Stress

Inventory and the work support scales designed for this research program. Factor

analysis of the Health Professions Stress Inventory revealed a four-factor solution:

Lack of Professional Recognition and Support, Patient Care Uncertainty, Job

Conditions, and Interpersonal Conflict. Cronbach’s coefficient alphas ranged from .62

to .83. Factor analysis of the Coworker Support Scale revealed a two-factor solution,

representing emotional and instrumental support. Cronbach’s coefficient alphas for

the Emotional Coworker Support and Instrumental Coworker Support were .92 and

.88 respectively. Contrary to expectations, factor analysis of the Supervisor Support

Scale revealed a one-factor solution. It was therefore deemed appropriate to examine

Supervisor Support as a unidimensional construct in subsequent analyses. Cronbach’s

coefficient alpha for the Supervisor Support scale was .96. Overall, the results from

Study 2 provided supporting evidence to suggest that the measures used in the current

research program were psychometrically sound.

In Study 3, the data collected in Study 2 was subjected to further statistical

analysis. The findings from Study 3 indicated that overall, the sample of Australian

nurses reported low to moderate levels of work stress, moderate levels of work

support and moderately high levels of burnout. For Emotional Exhaustion, predictor

variables accounted for 42.2% of the total variance. Sociodemographic factors

 vii

explained a small but significant proportion of the variance (2.7%). Work stressors

however, were the main predictors of Emotional Exhaustion, explaining 41.5% of the

total variance. Role Overload, Job Conditions and Role Conflict were the main

determinants of Emotional Exhaustion, with Role Overload explaining most of the

variance. For Depersonalisation, the predictor variables accounted for 34.2% of the

total variance. Sociodemographic factors (11.5%) and work stressors (33.6%) both

explained a significant proportion of the variance. Role Conflict and Patient Care

Uncertainty were the main determinants of Depersonalisation, with Role Conflict

explaining most of the variance. For Personal Accomplishment, Role Conflict and

Role Ambiguity explained 20.5% of the total variance, with Role Conflict explaining

most of the variance. Sociodemographic factors and job-specific stressors were not

significant predictors of Personal Accomplishment. Evidence for main effects of work

support on burnout were limited. There was no evidence to suggest that work support

had significant main effects on Emotional Exhaustion. Supervisor Support had a

small, but significant main effect on Depersonalisation (β = -.15, p < .05) and

Personal Accomplishment (β = -.24, p < .01). There was no evidence of main effects

for Emotional and Instrumental Coworker Support. Furthermore, the present research

found no significant evidence to support the buffering effect of work support on

burnout. Theoretical and practical implications of these findings are discussed.

 viii

TABLE OF CONTENTS

CHAPTER 1: INTRODUCTION
Background 1
The Changing Nature of Nursing in Australia 1
Burnout Among the Human Service Professions 5
The Influence of Social Support on Burnout 8
Research Program Outline 10
Study 1 10
Study 2 13
Study 3 15
Contribution to the Research Area 16
Substantive Contributions 16
Theoretical Contributions 17
Methodological Contributions 18
Summary 18
CHAPTER 2: A REVIEW OF THE OCCUPATIONAL STRESS
LITERATURE WITHIN THE NURSING PROFESSION
Definition of Occupational Stress 20
Work Stress in Nursing 25
Prevalence of Work Stress 26
Sources of Work Stress for Nurses 27
Job-specific stressors 27
Role stressors 34
Levels of Work Stress Across Nursing Wards 38
Levels of Work Stress Based on Gender 40
Summary 41
CHAPTER 3: A REVIEW OF THE LITERATURE ON BURNOUT
Background 42
Definition and Measurement of Burnout 44
Burnout – An Outcome of Job Stress 46
Burnout as a Process 48
Prevalence of Burnout among Nurses 55
Antecedents of Burnout 60
The Relationship Between Job-Specific Stressors and Burnout 61
The Relationship Between Role Stressors and Burnout 64
 ix

The Relationship Between Sociodemographic Factors and Burnout 71

Summary 74
CHAPTER 4: A REVIEW OF THE SOCIAL SUPPORT LITERATURE
Background 75
Definition of Social Support 76
Structural Support 76
Functional Support 78
Types of Support 80
Sources of Support 83
Support from Supervisors and Coworkers 84
Models of Social Support 87
Main effect model 88
Indirect effect model 90
Buffering effect model 93
Methodological Issues 96
Stressor-Support Matching Hypothesis 99
Optimal Matching Theory 103
Summary 107
CHAPTER 5: METHODOLOGICAL CONSIDERATIONS
Sampling Issues 109
Sample Size 110
Sample Representativeness 112
Sample Selection for Study 1 112
Sample Selection for Studies 2 and 3 115
Measurement Issues – Reliability and Validity 119
Qualitative Research – Study 1 119
Quantitative Research – Studies 2 and 3 122
Self-Report Data 123
Choice of Instruments 125
Description of Variables Investigated 125
Work Stress 126
Job-specific stress 126
Role Stress 129
Work Support 131
Burnout 135
Rationale for Research Method 138
Study 1 138
 x

Studies 2 and 3 141

Overview of Data Analyses 144
Study 1 144
Inter-Rater Reliability 146
Studies 2 and 3 146
Summary 149
CHAPTER 6: STUDY 1
A Qualitative Study of Nurses’ Perceptions of Work Stress and Social Support 150
Overview 150
Principal Research Questions and Associated Sub-Problems 151
Rationale for Collecting Qualitative Research 152
Rationale for Investigating Nurses’ Work Stressors 152
Rationale for Investigating Social Support 153
Method 154
Subjects 154
Procedure 154
Results 156
Classification of Comments into Work Stress Categories 156
Step 1: Range of Work Stressors Identified by Australian Nurses 158
Step 2: Broad Work Stress Categories and Associated Stressors 160
Step 3: Classification of Work Stress Subcategories 162
Step 4: Comparing Australian Nurses’ Sources of Stress to the NSS 163
Step 5: Description of Australian Nurses’ Main Sources of Work Stress 164
Work overload 164
Lack of respect and recognition from doctors 165
Difficult/demanding patients 166
Difficult/demanding relatives 167
Lack of support 168
Classification of Comments into Social Support Categories 169
Sources of Support 169
Identification of Major Sources of Work Support 170
Types of Support 171
Step 1: Identification of Supportive Behaviours 172
Step 2: Classification of the Social Support Dimensions 173
Step 3: Identification of the Main Types of Support for Nurses 174
Step 4: Description of the Main Types of Support for Nurses 174
Emotional Support 174
 xi

Instrumental Support 175

Discussion 176
Chronic Sources of Stress Among Australian Nurses 177
Australian Nurses’ Sources of Stress Compared to the NSS 181
Measures of Nurses’ Work Stress 185
Conceptualisation of Social Support 187
Limitations 190
CHAPTER 7: STUDY 2
An Assessment of the Psychometric properties of the HPSI and the 191
Work Support Scales
Rationale for Study 2 191
Method 193
Subjects 193
Materials 193
Work Stress 194
Job-specific stress 194
Role Stress 195
Work Support 195
Burnout 197
Procedure 197
Data Analysis 198
Results 199
Factor Structure of the HPSI 199
Factor Structure of the Work Support Scales 207
Coworker support 207
Supervisor support 208
Discussion 213
Summary 217
Limitations 218
CHAPTER 8: STUDY 3
The Influence of Social Support on the Stress-Burnout Relationship 219
Overview 219
Rationale for Study 3 220
Research Hypotheses 222
Method 224
Subjects 224
Materials 224
 xii

Work Stress Controllability 225

Procedure 228
Data Analysis 228
Results 229
Descriptive Statistics for the Independent and Dependent Variables 229
Level of Work Stress Among Australian Nurses 232
Sources of Stress for Australian Nurses 234
Major job-specific stressors 234
Major role stressors 236
The Nature of Burnout Among Australian Nurses 238
Intercorrelations between the burnout dimensions 238
Average level of burnout 239
Comparison with Other Human Service Professions 242
Prevalence of Burnout 243
Correlates of Burnout 244
The Relationship Between Job-Specific Stressors and Burnout 246
The Relationships Between Role Stressors and Burnout 246
The Relationship Between Sociodemographic Variables and Burnout 247
Level of Social Support for Nurses Working at Public Hospitals 249
The Relationship Between Social Support and Burnout 250
The Relationship Between Work Stressors and Social Support 252
Work Stress Determinants of Burnout 255
Antecedents of Emotional Exhaustion 255
Antecedents of Depersonalisation 256
Antecedents of Personal Accomplishment 257
The Influence of Social Support on the Stress-Burnout Relationship 258
The Influence of Work Support on Emotional Exhaustion 259
The Influence of Work Support on Depersonalisation 263
The Influence of Work Support on Personal Accomplishment 267
Discussion 272
Work Stress Among Hospital Nurses 273
Burnout – A Multidimensional Construct 277
Level of Burnout Among Nurses 277
Correlates of Burnout 280
Relationships between work stress and burnout 280
Relationships between sociodemographic factors and burnout 282
Levels of Work Support 283
 xiii

The Influence of Work Support on the Stress-Burnout Relationship 283

Relationships between work support and burnout 283
Relationships between work support and work stress 284
Main Predictors of Burnout 284
Main Effects of Work Support on Burnout 288
Buffering Effect of Work Support on Burnout 290
Limitations 291
Summary 292
CHAPTER 9: DISCUSSION AND CONCLUSION OF THE RESEARCH PROGRAM
Overview of the Research Program 294
Overview of the Main Research Findings 297
Theoretical Implications for Burnout 302
Theoretical and Methodological Implications for Social Support 305
Implications for the Health Care Industry 312
Organisational Interventions 313
Individual Interventions 316
Limitations 318
Future Research Recommendations 319
Conclusion 323
REFERENCES 325
 xiv

LIST OF TABLES

2.1 Studies Between 1993-2003 that have used Gray-Toft and 33

Anderson’s (1981) Nursing Stress Scale

3.1 Burnout Levels Among the Human Service Professions Based on 59

Studies Published Between 1979 and 1998

5.1 Summary Statistics for the Demographic Variables 117

5.2 Supervisor Support Scale 134

6.1 Sources of Stress for a Sample of Australian Nurses 158

6.2 Classification of Australian Nurses’ Sources of Work Stress 161

6.3 Main Sources of Work Stress for a Sample of Australian Nurses 163

6.4 Australian Nurses’ Major Sources of Stress Compared to the NSS 164

6.5 Sources of Workplace Support for a Sample of Australian Nurses 170

6.6 A Range of Supportive Behaviours Identified by Australian Nurses 172

6.7 Classification of the Social Support Dimensions 173

6.8 Main Types of Work Stress for a Sample of Australian Nurses 174

7.1 Pattern Matrix for Oblique Factors of the HPSI 200

7.2 Pearson Product-Moment Correlations Between the HPSI Subscales, 205

ORQ and the MBI (N = 273)

7.3 Pattern Matrix for Oblique Factors of the Coworker Support Scale 208

7.4 Factor Matrix of the Supervisor Support Scale 209

7.5 Pearson Product-Moment Correlations Between the Work Support 210

Scales and the HPSI and ORQ Subscales (N = 273)

7.6 Pearson Product-Moment Correlations Between the Work Support 210

Scales and the MBI (N = 273)

 xv

8.1 Nurses’ Perceptions of Work Stress Controllability 227

8.2 Descriptive Data for the Independent and Dependent Variables 229

(N = 273)

8.3 Mean Levels of Work Stress for the Total Sample Population and 232

Across Gender

8.4 The Highest Rated Job-Specific Stressors for Australian Nurses 235

8.5 The Highest Rated Role Stressors for Australian Nurses 239

8.6 Pearson Product-Moment Correlations between the MBI Subscales 237

(N = 273)

8.7 Means, Standard Deviations, and Reliability Coefficients for a 238

Sample of Australian Nurses and a MBI Normative Sample

8.8 Mean Scores for the MBI Subscales for Nursing Studies 241

Between 1991 and 2001

8.9 Burnout Levels Across Occupational Fields 242

8.10 Categorisation of the MBI Scores 243

8.11 Correlations Between the Work Stress Subscales and the MBI 245

Subscales

8.12 Pearson’s Product-Moment Correlations Between Burnout 248

and the Sociodemographic Variables

8.13 Mean Levels of Coworker and Supervisor Support 249

8.14 Pearson’s Product-Moment Correlations Between the 251

Work Support Scales and the MBI Subscales

8.15 Pearson’s Product-Moment Correlations Between the Work Stress 252

and Work Support Variables

8.16 Standard Multiple Regression of Work Stress on Emotional Exhaustion 255

 xvi

8.17 Standard Multiple Regression of Work Stress on Depersonalisation 257

8.18 Standard Multiple Regression of Work Stress on 258

Personal Accomplishment

8.19 Hierarchical Regression Analysis for Emotional Exhaustion 261

(Standardised Regression Coefficients β)

8.20 Hierarchical Regression Analysis for Depersonalisation 265

(Standardised Regression Coefficients β)

8.21 Hierarchical Regression Analysis for Personal Accomplishment 269

(Standardised Regression Coefficients β)

 xvii

LIST OF FIGURES

Figure 2.1 Lazarus and Folkman’s (1984) transactional model of stress 22

Figure 3.1 Maslach’s (1982) model of burnout 50

Figure 3.2 Golembiewski and Munzenrider’s (1988) model of burnout 51

Figure 3.3 Leiter’s (1993) model of burnout 53

Figure 4.1 Potential effects of social support on work stress and health 88

Figure 8.1 Total mean item scores for the Health Professions Stress 234

Inventory

Figure 8.2 Total mean item scores for the Occupational Roles 236

Questionnaire

Figure 8.3 Four-step hierarchical regression process for Emotional 260

Exhaustion

Figure 8.4 Four-step hierarchical regression process for Depersonalisation 264

Figure 8.5 Main effect of Supervisor Support on Depersonalisation 267

Figure 8.6 Three-step hierarchical regression process for Personal 268

Accomplishment

Figure 8.7 Main effect of Supervisor Support on Personal 271

Accomplishment

Figure 9.1 A conceptual model of burnout among nurses 303

 xviii

LIST OF APPENDICES

Appendix A Information sheet and survey used in Study 2 381

Appendix B Information sheet, consent form and coding sheet 392

used in Study 1

Appendix C Rating sheet used to classify nurses’ work stressors 403

into controllable and uncontrollable events

 xix

LIST OF ABBREVIATIONS

SCARC Senate Community Affairs References Committee

AIHW Australian Institute of Health and Welfare

NSS Nursing Stress Scale

HPSI Health Professions Stress Inventory

MBI Maslach Burnout Inventory

MBI-HSS Maslach Burnout Inventory – Human Services Survey

ORQ Occupational Roles Questionnaire

OSI Occupational Stress Inventory

CNC Central Nurse Coordinator

NPC Nursing Practice Coordinator

CINAHL Cumulative Index for Nursing and Allied Health Literature

 xx

STATEMENT OF ORIGINAL AUTHORSHIP

The work contained in this thesis has not been previously submitted for a degree or

diploma at any other higher education institution. To the best of my knowledge and

belief, the thesis contains no material previously published or written by another

person except where due reference is made.

Signed: …………………………………………………………

Date: ……………………………………………………………
 xxi

DEDICATION

To Nurses.
 xxii

ACKNOWLEDGEMENTS

The challenges I faced throughout the writing of my thesis were overcome

because of the unwavering support I received from my principal supervisor Professor

Wendy Patton. Thank you for sharing your research knowledge and experiences with

me, for believing in me, and for accepting to be a part of my journey. Words are

unable to express my gratitude. Thank you to my second supervisor Dr Richard

Goddard for your enthusiasm, advice and guidance. My sincere thanks goes to Dr

Roland Simons who patiently taught me how to use the statistical analysis software

program used in this study. Thank you to all the nurses who participated in this study.

I hope that in the not too distant future you receive the recognition and praise you

deserve. Finally, thank you to my husband Christopher Lane. Without your support,

your encouragement, and above all else, your love, this would not have been possible.
 1

CHAPTER 1

Introduction

Background

The current thesis investigates how work-related support influences the level

of stress and burnout experienced by Australian nurses working in public hospitals.

Using a sample of nursing staff from three public hospitals in Queensland’s Brisbane

metropolitan district, the thesis provides a comprehensive examination of the nature

and level of stress and burnout experienced by nurses and investigates the complex

relationships between a range of chronic work stressors, social support and burnout.

During the time the current research was being undertaken, it was well

publicised in the media that public hospitals were facing a nursing shortage crisis.

Nurses acknowledged feeling stressed and burnt out and as a result, significant

numbers of nurses were choosing to work part-time or were voluntarily leaving the

profession. The following section serves to highlight the pertinent pressures currently

faced by Australian nurses working in the public hospital system and provides support

for further research into stress and burnout in this profession.

The Changing Nature of Nursing in Australia

It is well recognised that the nature of nursing in Australia is rapidly changing.

In a recent review of nursing in Australia, Johnson and Preston (2001) reported that

health care has become far more complicated and difficult for nurses. As one senior

nurse stated in their study:

The people now in general wards were in intensive care wards fifteen years

ago and many people cared for in hospital fifteen years ago are now cared for

in the community. The people who are now in intensive care would have died

fifteen years ago (p.6).

 2

It would seem that substantial budget cuts and economically driven health

agendas have taken their toll on Australia’s health system and, certainly, on nurses

(Richards, 2000). As the nursing workforce constitutes the largest group in the health

care system – over 55% of the entire health workforce – it has often been the most

affected by fiscal constraint (Senate Community Affairs References Committee

[SCARC], 2002).

Despite reduced funding, hospital admissions are increasing at a significant

rate. The increase in admissions is mainly attributable to Australia’s population

growth and its ageing population with its concomitant increase in chronic illnesses

and disabilities. According to the Australian Institute of Health and Welfare [AIHW]

(2002), between 1993-94 and 2000-01, hospital admissions increased by 33.2%. An

upward trend in admissions has been accompanied by significant declines in the

number of available beds and the length of hospital stays. Between 1993-94 and 2000-

01, there was a 14.2% reduction in available beds in public hospitals, resulting in a

decrease from 3.4 to 2.7 beds per 1000 population (AIHW, 2002). Also during this

time, the average length of hospital stays decreased by 19.6% from 4.6 days to 3.7

days (AIHW, 2002). Shortened hospital stays have been credited to new technology

and scientific developments that have enabled the rapid assessment, treatment and

discharge of patients from hospitals (SCARC, 2002). For nurses, however,

technological improvements have led to a growing need for a highly specialised

workforce capable of using and understanding this technology (SCARC, 2002).

Indeed, some specialties would not be able to function efficiently without an

appropriate specialist nurse workforce. In addition, shorter stays in hospital have

resulted in a greater turnover of patients with more complex care needs. The effect on
 3

nurses is an augmented workload and the provision of more diverse care which

requires greater alertness and vigilance (Iliffe, 2000).

Changes to the health care system, the way in which services are delivered,

and changes to the skills required of those working within the system, are

significantly affecting the nursing workforce at a time when there is a severe shortage

of experienced nurses and related problems in retaining those still in the nursing

workforce (SCARC, 2002). The AIHW (1999) reported that nursing registrations and

enrolments fell from 270,720 in 1993 to 257,662 in 1999, a decrease of 4.8%. In

Queensland, however, there were 36,817 registered nurses in 2001, an increase of

5.4% since 1996 (SCARC, 2002).

Significant numbers of nurses are leaving the profession or are choosing part-

time or casual work. The nursing workforce has a large number of part-time

employees and this is readily increasing. The proportion of nurses in part-time work

(that is less than 35 hours per week) increased from 41.2% to 44.1% between 1990 to

1999, and in 1999 enrolled nurses were much more likely to work part-time (60%)

than registered nurses (52.5%) (AIHW, 1999). This has put an even greater strain on

full-time nursing staff as they are regularly required to orientate casual staff to the

ward and take on additional work for new staff who are unable to perform higher-

level functions (Iliffe, 2000). Iliffe suggested that reduced staffing levels are

compromising the nurses’ quality of care and have increased the propensity for

mistakes to be made.

An ageing nursing workforce is also exacerbating the nursing shortage crisis.

The average age of nurses increased from 39.1 years to 40.4 years between 1994 and

1997. The Australian Nursing Council Incorporated (2002) noted that the ageing

nursing population reflected the expansion of the nursing workforce which occurred
 4

during the 1970s and 1980s. Those nurses are now in their 40s and 50s. It is

predicted that over the next 10 to 15 years, 30% of the workforce will be

contemplating retirement. There is also concern that there are insufficient graduates to

replace older nurses as they reach retirement. Commencements of domestic students

in all nursing courses (undergraduate, postgraduate, and research students) decreased

from 11,653 in 1994 to 8,423 in 2000 (SCARC, 2002). Iliffe suggested that poor

working conditions are having a devastating effect on attracting people to the

profession.

In 1999, over 2000 nursing members from the Victorian Branch of the

Australian Nursing Federation participated in a study designed to examine the impact

of understaffing on nurses. The study’s findings revealed that 56% of the sample

considered low nurse/patient ratios were responsible for extensive workloads

(Considine & Buchanan, 1999). Furthermore, 65% of nurses surveyed reported

working unpaid overtime such as working through meal breaks or extending shift

times to complete work. Considine and Buchanan estimated that nurses’ unpaid

labour was contributing the equivalent of 300-450 full-time nursing positions per

week.

In summary, the shift to a ‘cost-control’ approach to managing public

hospitals has resulted in significant changes in the nature of work undertaken by

nurses. Changes in the objective conditions at work have had major implications for

nurses’ subjective experiences of work, with increasing numbers of nurses feeling

stressed and as a consequence, are opting to work part-time or leave the profession.

The present research is therefore both timely and vitally important as very few

empirical studies have explored the implications of these stressful work conditions on

Australian nurses’ health and well-being. Furthermore, there have been no studies to
 5

date that have examined burnout among Australian nurses. The section below briefly

examines the burnout syndrome and provides evidence as to why nurses are

susceptible.

Burnout Among the Human Service Professions

Burnout describes a characteristic bundle of strain symptoms found primarily

in helping professions (Büssing & Glaser, 1999). Burnout is not a symptom of work

stress; it is the end result of unmanaged work stress (Altun, 2002). It is characterised

by feelings of emotional exhaustion, depersonalisation and reduced personal

accomplishment (Cordes & Dougherty, 1993; Vandenberghe & Huberman, 1999).

Emotional exhaustion involves feelings of being emotionally overextended and

exhausted by one’s work (Maslach, Jackson, & Leiter, 1996). Depersonalisation

refers to the development of impersonal and unfeeling attitudes towards recipients of

one’s service, care, treatment, or instruction (Maslach et al., 1996). Feelings of

reduced personal accomplishment occur when an individual experiences a decline in

his or her feelings of competence and successful achievement in working with people

(Maslach, 1998).

Burnout has been found to occur among individuals who work with people in

emotionally charged situations. It is prominent in the human service professions (such

as nursing, paramedics, teaching, social work and counselling) where extensive and

direct face-to-face contact with other individuals is involved (Cordes & Dougherty,

1993). It is considered to be a response to the chronic stress of dealing with other

people, particularly when they are troubled or are having problems (Maslach &

Schaufeli, 1993). Burnout is likely to occur in people who feel overworked and

unappreciated (Altun, 2002). Freudenberger (1974) added that “the dedicated and the
 6

committed” were the ones most prone to suffering burnout, because they “work too

much, too long, and too intensely” (p. 161).

It would seem that nurses are especially vulnerable to burnout. A nurse’s

principal mission is orientated towards nurturing and caring for people in the human

health experience (Duquette, Kérouac, Sandhu, Ducharme, & Saulnier, 1994). They

confront patients with loneliness, pain, agony, incapacity, disease, and death and they

provide presence, comfort, support and help to patients around the clock (Duquette et

al., 1994). A major appeal of working in the health care profession is the challenge of

dealing with difficult and sometimes life-threatening situations, such as helping

people who are experiencing major health problems. Such work can be rewarding,

for instance when patients recover because of the professional’s efforts (Bakker,

Killmer, Siegrist, & Schaufeli, 2000). However, it is this constant exposure to

stressful situations that forms an excellent breeding ground for burnout in nurses (De

Rijk, LeBlanc, Schaufeli, & De Jonge, 1998).

Earlier empirical studies investigating burnout in nurses have shown that

burnout is positively correlated with the amount of time nurses spend with their

patients (Cronin-Stubbs & Brophy, 1985), with the intensity of the emotional

demands posed by their patients (Lewinson, Conley, & Blessing-Moore, 1981), and

with exposure to patients with a poor prognosis (Hare, Pratt, & Andrews, 1988). More

recent studies, however, have demonstrated that burnout among nurses is associated

with important work-related factors, such as high workload (Duquette et al., 1994;

Greenglass, Burke, & Fiksenbaum, 2001; Landsbergis, 1988), poor support (Corrigan

et al., 1994; Duquette et al., 1994), interpersonal conflict (Payne, 2001), death and

dying (Payne, 2001) and inadequate preparation (Payne, 2001).

 7

In recent years there has been a surge of interest in the level of burnout

experienced by hospital nursing staff (Bakker et al., 2000; Duquette, Kérouac,

Sandhu, Ducharme, & Saulnier, 1995; Greenglass et al., 2001; Payne, 2001; Peeters &

Le Blanc, 2001). According to two European epidemiological studies, burnout affects

approximately 25% of all nurses (Landau, 1992; Saint-Arnaud, Gingras, Boulard,

Vezina, & Lee-Gosselin, 1992). Left unchecked, burnout can have grave implications

not only for the nurse, but also for their patients. It has been postulated that burnout is

correlated with a range of self-reported psychological and physical strain indicators

such as tension and irritability (Duquette et al., 1995), fatigue (Costantini, Solano,

DiNapoli & Bosco, 1997), headache and sleep disorders (Constantini et al., 1997).

Burnout has been implicated in the reduction in quality of care and service delivery,

absenteeism and job turnover (Cox, Kuk, & Leiter, 1993; Maslach & Jackson, 1981,

1982; Perlman & Hartman, 1982; Pines and Maslach, 1978; Van Yperen, Buunk, &

Schaufeli, 1992). Ultimately, this compromised standard of care impacts on the

effectiveness and success of a hospital (Akroyd, Caison, & Adams, 2002).

In summary, burnout is a specific kind of occupational stress reaction

prevalent among human service professionals. It occurs as a result of the demanding

and emotionally charged relationships between caregivers and their recipients

(Maslach & Schaufeli, 1993). Nursing is an inherently stressful occupation (Schaefer

& Moos, 1993) and researchers have found that the nursing population is at a high

risk of burnout (Cordes & Dougherty, 1993; Duquette et al., 1994; Peeters & Le

Blanc, 2001).

In the present study, the researcher examined Australian nurses’ perceptions of

work stress in an attempt to identify stressful work conditions that significantly

predict burnout. Furthermore, the researcher investigated an important coping

 8

resource available to nurses in their work environment – social support. The role

social support plays in reducing or preventing burnout will now be briefly discussed.

The Influence of Social Support on Burnout

The occupational stress literature identifies social support as a useful coping

resource in managing stressful situations within the workplace and reducing the

harmful consequences of stress on well-being (Fenlason & Beehr, 1994; Kong,

Wertheimer, & Myers, 1994; Scheck, Kinicki, & Davy, 1997). Nursing researchers

have focused mainly on the coping process among patients, spouses and significant

others during chronic illness (Armstrong, 1987; Burckhardt, 1987; Gulick, 1995;

Kammer, 1994; Lambert & Lambert, 1987; Lauver & Tak, 1995). Less research

however, has been conducted on the process of how nurses cope in the workplace

(Albrecht & Halsey, 1991; Hendel, Fish, & Aboudi, 2000; Simoni & Paterson, 1997).

Social support describes those individuals and groups one turns to either on a regular

basis or in a time of need (Scheck et al., 1997). Leavy (1983, p.5) broadly defined

social support as “the availability of helping relationships and the quality of those

relationships.”

Researchers have identified a range of supportive behaviours provided to

individuals in need. These supportive behaviours have been further categorised into

two key types of support – emotional and instrumental support. Emotional support is

information obtained from others that one is respected and accepted (Scheck et al.,

1997). This type of support suggests to individuals they are valued by another. In

contrast, instrumental support includes a wide range of activities concerning the

practical help one is given from others (House, 1981; Scheck et al., 1997; Thoits,

1982). Such activities include help with work responsibilities, advice in solving

problems and sharing of knowledge. In addition, there are two main sources of
 9

support - work based support (supervisors/coworkers) and non-work based support

(e.g., partner, family, friends).

A review of the literature demonstrates that there are two main processes in

which social support influences the stress-burnout relationship. Some studies have

indicated a ‘main’ or ‘direct’ effect for social support, suggesting that support reduces

burnout regardless of the intensity of the work stressors experienced (e.g., Beehr,

1985; Cohen & Wills, 1985; Sullivan & Bhagat, 1992). Others have found a

‘buffering’ or ‘moderating’ effect in which social support interacts with work

stressors to affect burnout. Specifically, the relationship between stress and burnout is

thought to be stronger for those individuals with low levels of support and weaker for

those individuals with high levels of support (Ganster, Fusilier, & Mayes, 1986;

LaRocco, House, & French, 1980).

Although empirical studies have consistently found support for the main effect

model, there has been mixed support for the buffering effect. A problem inherent in

earlier studies is their failure to consider different stressful events pose different

coping requirements. The stressor-support matching model proposes that stress

buffering occurs only when there is a match between the needs elicited by the

stressful event and the functions of support that are perceived to be available (Cohen

& McKay, 1984; Cohen & Wills, 1985). The present thesis builds upon previous

stressor-support matching models (Cohen & Wills, 1985; Cutrona, 1990; Cutrona &

Russell, 1990) by examining not only the types of support (i.e., emotional support and

instrumental support) provided to nurses, but also the sources of support (i.e.,

coworker support and supervisor support) available to nurses within a hospital

environment. The thesis proposes that if the right kind of support from the right
 10

source of support is matched to a specific type of work stress, then burnout may be

prevented.

In addition, the researcher heeded the advice of Cutrona (1990; Cutrona &

Russell, 1990) who proposed that in order to discover optimal matches between

different types of stressful events and types of support, the controllability of the

stressful event should be considered. According to Cutrona’s conceptual model, when

an event is uncontrollable, that is, nothing can be done to prevent the event or lessen

its consequences, the most beneficial support will serve to minimise the intensity of

the individual’s negative emotional reactions to the event (i.e., emotional support). By

contrast, when an event is controllable, that is the individual can prevent its

occurrence or consequences, the most beneficial support will foster effective

instrumental action (i.e., instrumental support). In the present thesis, stressful work-

related factors are classified as controllable or uncontrollable events by nursing

experts before examining the effect of different types of support from various sources

of support on burnout. In the following section, a broad overview of the research

program is given.

Research Program Outline

The research program consisted of three successive studies. The thesis

comprised both qualitative and quantitative research methods. The qualitative

research component preceded the quantitative research component. The studies are

briefly outlined below.

Study 1

The first study commenced mid 2000. Study 1 involved the qualitative

component of the research program in which focus groups were conducted with 68

nurses (12 males, 56 females) from two public hospitals in Queensland, Australia.
 11

Focus groups are defined as “in-depth, open-ended group discussions that explore a

specific set of issues on a predefined and limited topic” (Robinson, 1999, p.905).

Study 1 was designed to address two principal research questions and two associated

sub-problems. The principal research questions were:

(1) What are the chronic sources of work stress commonly experienced among

Australian nurses?

(2) To what extent is the way in which nurses’ perceive work support

consistent with how social support is conceptualised and operationalised in

recent empirical literature?

The sub-problems examined in Study 1 related to principal research question 1. The

associated sub-problems included:

(1a) To what extent are the sources of stress experienced by Australian nurses

consistent with the seven stress factors measured by Gray-Toft and

Anderson’s (1981a) NSS?

(1b) Do Australian nurses from various wards share similar perceptions of

work stress?

An examination of recent nursing literature revealed that there is very little

published research investigating the common sources of work stress among Australian

nurses working in a public hospital system. Since similar findings regarding the

nature of nurses’ work stress have been found across nursing studies, researchers have

tended to presume that all nurses are exposed to the same work stressors. Wheeler

(1998) proposed, however, that consistent findings across studies may be the result of

an over-reliance on Gray-Toft and Anderson’s (1981a) Nursing Stress Scale (NSS).

Study 1 sought to determine the common stressors that Australian nurses regularly

confront at work by depicting the work stressors that were prevalent in most nursing
 12

wards. In addition, Study 1 sought to determine whether the work stressors identified

by a sample of Australian nurses were consistent with the stressful events measured

by the NSS.

Content analysis of the qualitative data in Study 1 provided insight into the

wide variety of work stressors experienced by nurses. Major sources of stress

included job-specific work stressors and generic role stressors. The job-specific

sources of stress identified by nurses most closely aligned to the four work stress

factors assessed using Wolfgang’s (1988a) Health Professions Stress Inventory

(HPSI): Professional Recognition, Patient Care Responsibilities, Job Conflicts and

Professional Uncertainty. The major sources of role stress were consistent with those

measured by Osipow and Spokane’s (1987) Occupational Roles Questionnaire: Role

Overload, Role Conflict, and Role Ambiguity. An investigation of chronic sources of

stress for Australian nurses was a necessary preliminary step before collecting

quantitative data in Study 2 as it enabled the researcher to determine the most

appropriate measure for assessing work stressors for a sample of Australian nurses.

A second gap in the social support literature is that most researchers have

failed to clearly operationalise the social support construct before designing measures

of social support. Several measures of social support exist in the stress-coping

literature. These measures have been primarily designed to assess the structure of an

individual’s social network or the functions of social support. Before constructing a

suitable measure of work support for nurses in Study 2, the researcher took the

opportunity to examine nurses’ perceptions of support at work in Study 1.

The results suggested that nurses’ perceptions of work support are consistent

with contemporary views that social support is a multidimensional construct. For

instance, content analysis of the qualitative data revealed that the main supportive
 13

behaviours identified by nurses could be broadly categorised as either emotional or

instrumental support. Furthermore, the nurses’ primary sources of support were their

nursing colleagues and their nursing supervisor (e.g., Central Nurse

Coordinator/Nursing Practice Coordinator).

Study 2

The second study of the research program was conducted late 2000 and early

2001. Using a cross-sectional methodology, 273 nurses from three public hospitals in

Brisbane’s metropolitan district completed a self-report questionnaire. The

questionnaire comprised of four measures that were designed to assess: job-specific

stress (i.e., professional recognition, patient care responsibilities, job conflicts and

professional uncertainty), role stress (i.e., role conflict, role overload and role

ambiguity), work support (i.e., coworker and supervisor support), and burnout (i.e.,

emotional exhaustion, depersonalisation, and reduced personal accomplishment).

Prior to examining the relationships between work stress, social support and

burnout in Study 3, it was deemed appropriate to assess the psychometric soundness

of Wolfgang’s (1988a) HPSI and the work support scales. Although the HPSI was

designed specifically to measure work stress among health care professionals, the

HPSI has not been commonly used in the nursing stress literature. Three studies

(Akhtar & Lee, 2002; Eells, Lacefield, & Maxey, 1994; Gupchup & Wolfgang, 1994)

to date have examined the factor structure of HPSI and two of these studies (Akhtar &

Lee, 2002; Eells et al., 1994) have used a sample comprising of nurses. Similarly, the

work support scales were constructed specifically for the purpose of this research and

therefore their psychometric properties were not known. It was important to establish

that these measures demonstrated adequate levels of reliability and validity in order to
 14

be confident that research findings were accurate and not the result of measurement

error.

Factor analysis of the HPSI was conducted using principal axis factoring. The

results indicated that the factor structure was different from the factor structure

identified in previous studies (Akhtar & Lee, 2002; Eells et al., 1994; Gupchup &

Wolfgang, 1994). The researcher renamed the four stress factors: Lack of

Professional Recognition and Support, Patient Care Uncertainty, Job Conditions, and

Interpersonal Conflict. The four-factor solution explained 45.2% of the total variance.

The HPSI yielded adequate internal reliability coefficient alphas which ranged

between .62 and .83. Intercorrelations between the four factors ranged between .21

and .56. These low to moderate correlations suggested that none of the work stress

factors were redundant. Evidence of the HPSI’s construct validity was also apparent.

Examination of the nursing literature revealed that there were no established

measures that assessed social support which were consistent with the findings

obtained in Study 1. A measure of work support was therefore developed for the

purpose of the present research. The work support scales were designed to assess

nurses’ perceptions of the types of support they receive at work from their supervisor

and their coworkers. To construct this measure, the researcher took specific items

from existing social support scales that demonstrated adequate psychometric

characteristics. The chosen items closely resembled the key supportive behaviours

identified by the sample of nurses who participated in the focus group discussions in

Study 1. The survey items were slightly modified to ensure that they were

contextually relevant to nurses.

The factorial validity of the Coworker Support Scale and the Supervisor

Support Scale were assessed by factor analysis using principal axis factoring. Some
 15

interesting findings were revealed. First, factor analysis of the Coworker Support

Scale identified two factors labelled emotional support and instrumental support. The

two-factor solution explained 68.56% of the total variance. Internal consistency was

high with a Cronbach’s coefficient alpha of .92 for Emotional Coworker Support and

.88 for Instrumental Coworker Support. These factors correlated highly and shared

52% of same variance. Despite high inter-scale correlations, the emotional and

instrumental coworker subscales were differentially related to the work stressors and

burnout dimensions. It was therefore concluded that the emotional and instrumental

support subscales were not redundant.

In contrast to the two-factor Coworker Support Scale, factor analysis of the

Supervisor Support Scale revealed a one-factor solution which accounted for 71.34%

of the total variability. The Supervisor Support Scale demonstrated very high internal

reliability with a Cronbach coefficient alpha of .96. It was therefore not feasible to

examine the different types of supervisor support when exploring the effects of

supervisor support on the stressor-strain relationship.

Overall, however, Wolfgang’s (1988a) HPSI and the work support scales

demonstrated adequate levels of reliability and validity. The data obtained from these

measures was therefore used for further analyses in Study 3.

Study 3

Following confirmation of the reliability and validity of the job-specific stress

measure and the work support scales, the data was subjected to further statistical

analysis. Study 3 establishes the nursing sample’s overall levels of work stress, social

support and burnout. In addition, main sources of work stress for Australian nurses

are identified. The sample’s level of burnout is compared to appropriate normative

data provided in the Maslach Burnout Inventory Manual (Maslach et al., 1996), as
 16

well as burnout studies on nurses and other human service professionals. Study 3 also

identifies the main correlates of burnout. Furthermore, the amount of variance

attributed to sociodemographic factors, work stressors, and work support in

explaining nurses’ burnout is explored. Finally, the effects of different types and

sources of support on the stress-burnout relationship are examined.

Contribution to the Research Area

The current research program aims to make substantive contributions to the

existing knowledge in the areas of occupational stress, social support and burnout in

the nursing and human services literature. In addition, the research program

endeavours to make theoretical contributions to the study of burnout and social

support and methodological contributions in the areas of nursing stress and social

support.

Substantive Contributions

First, the research program extends the nursing literature by examining a

sample of Australian nurses’ perceptions of work stress. Few recent empirical studies

(Bryant, Fairbrother, & Fenton, 2000; Healy & McKay, 2000; Lumby, 1996) have

explored Australian nurses’ perceptions of work stress. The research will provide

insight into the varied work stressors commonly confronted by a sample of Australian

public hospital nurses. Furthermore, the research will determine the major sources of

stress and the average level of stress experienced by a sample of public hospital

nurses in Australia.

Second, there is no research into burnout among Australian nurses to date. The

present thesis endeavours to determine the prevalence of burnout in a sample of

Australian nurses. The research findings will also provide a benchmark for assessing

burnout among Australian nurses in future studies.

 17

The research program also addresses the lack of research into the effects of

social support on nurses’ health and well-being. Specifically, the study further

contributes to our understanding of how nurses view the support received from within

their work environment. Furthermore, the level of support nurses receive at work will

be established.

With respect to the human services literature, the research provides an

indication of how Australian nurses’ burnout levels compare to nurses overseas and

other human service professionals. In addition, this research expands current

knowledge of specific individual and organisational factors that serve to both enhance

and reduce susceptibility to burnout. The research findings will contribute to the

development of appropriate stress-reduction strategies that may be used in public

hospitals and other health care facilities.

Theoretical Contributions

Previous nursing literature has examined work stressors and burnout as

relatively separate constructs. The present thesis reconciles conceptually distinct, but

related, phenomena by using Lazarus and Folkman’s (1984) stress-strain-coping

framework as the basis for this research. In particular, the study examines individual

factors (i.e., sociodemographic factors), organisational factors (i.e., work stressors),

and coping resources (i.e., work support) in relation to burnout, enabling the

researcher to offer a model of burnout for nurses that is theoretically grounded.

With respect to social support, the research contributes to the existing social

support literature by further exploring the way in which social support is

conceptualised and operationalised. In addition, the research extends the stressor-

support matching hypothesis by exploring not only types, but also sources of support.

Furthermore, when examining the stressor-support matching hypothesis, the

 18

controllability of nurses’ work stressors is taken into consideration before examining

the match between specific stressful situations and support requirements.

Methodological Contributions

Previous nursing researchers have commonly relied on Gray-Toft and

Anderson’s (1981a) NSS without establishing whether it is the most appropriate tool

measuring job-specific stress in their particular nursing sample. In the present study,

Australian nurses’ perceptions of work stress are considered before choosing an

adequate measure of job-specific stress for nurses. Similarly, the majority of social

support researchers have relied on Caplan, Cobb, French, Van Harrison and Pinneau’s

(1975) Social Support Instrument to measure work support despite its inability to

sufficiently measure different types of support. Based on a sample of nurses’

perceptions of support at work, a measure of work support was constructed for the

purpose of this research. The psychometric soundness of the job-specific stress

measure and the work support scales are explored.

Summary

Chapter 1 has presented a brief overview of the research program. The

researcher outlined information relevant to nursing in Australia today thus providing

the contextual background in which this research may be understood. The three

studies that comprise this research program were briefly discussed. The substantive,

theoretical and methodological contributions that this research makes in the areas of

nurses’ work stress, social support, and burnout have been delineated.

In Chapters 2 to 4, a detailed review of the literature pertinent to this thesis is

provided. Chapter 2 examines the occupational stress literature. Chapter 3 explores

the literature in relation to burnout and Chapter 4 details the social support literature.

Chapter 5 discusses the major methodological considerations pertinent to this research

 19

before summarising the main research questions and associated research hypotheses.

Chapters 6 to 8 will each present one of the three studies that comprise this research

program. Chapter 9 provides an overall discussion of the research program including

implications and concludes with recommendations for future research.

 20

CHAPTER 2

A Review of the Occupational Stress Literature within the Nursing Profession

Chapter 2 provides a comprehensive review of the current occupational stress

literature within the nursing profession. First, a definition of work stress is provided

that stems from a transactional theoretical framework. Next, a brief review of the

prevalence of work stress in the nursing industry is given. This is followed by a

discussion on the major sources of stress associated with nursing in a hospital setting.

Methodological issues concerned with measuring nurses’ stress are also examined.

Finally, this chapter explores studies examining nurses’ perceptions of work stress

based on nursing wards and gender.

Definition of Occupational Stress

The growing awareness of the implications of work stress on nurses is evident

by the sheer number of studies investigating occupational stress in nursing. In recent

years, nursing studies have primarily examined the causes and consequences of work

stress without providing a clear definition of the stress concept. Furthermore, little or

no attention has been given to the theoretical framework underpinning the stress

process.

In the scientific literature, stress has been defined as a response to something

in the environment or as the stimulus that caused the response (Helps, 1996). Closer

examination of the nursing literature would suggest that studies primarily research

stress from a stimulus-based perspective (Wheeler, 1997a). A stimulus-based model

asserts that stress is an aspect of the environment (a stimulus) that causes a strain

reaction in the individual exposed to the stressful stimulus (Bartlett, 1998). Stress is a

characteristic, event, or situation in the environment. Most nurse stress studies suggest

that nurses’ work stress is caused by external stimuli (e.g., noise, abusive patients)
 21

which cause varying levels of strain (i.e., pressure) with which the nurse is unable to

cope (Wheeler, 1997a).

The value of defining nursing stress from a stimulus-based perspective is its

recognition of a wide variety of work-related events or situations. Contemporary

researchers criticise this model, however, for its inability to explain and account for

the complexities of the stress process (Bartlett, 1998). By simply focusing attention on

one component (i.e., the stimulus), other components that make up the stress process

become artificially separated (Dewe, 1989). One result of this arbitrary separation is

different situations are being labelled stressful and, irrespective of whether they are

stressful or not, they are presumed to have some intrinsically stressful properties

(Dewe, 1989). The fact that a nurse is bombarded by stimuli does not necessarily

mean that he/she is distressed by them (Wheeler, 1997a). It is argued that situational

characteristics alone may be inadequate to predict an individual’s response. Stress

must be first perceived to exist before it can result in strain in individuals.

A more preferable approach to defining stress is one that takes into

consideration that stress is relational in nature, involving some sort of transaction

between the individual and the environment. Lazarus and Folkman’s (1984)

transactional model defines the stress process as “a particular relationship between the

person and the environment that is appraised by the person as taxing or exceeding his

or her resources and endangering his or her well-being” (Lazarus & Folkman, 1984,

p.19). Dewe (1989) proposed that a transactional definition of stress differs from a

stimulus or response-based definition of stress in three main ways: (1) stress does not

reside solely in a situation or in a response; (2) stress is a process; a transaction

between the individual and the environment; and (3) coping and adaptation are

explicit parts of that transaction that help shape the stressful experience.
 22

A transactional approach focuses on the individual’s perception of the

demands of the environment (i.e., stressors) and his/her resources to cope (e.g., social

support), as the important factors determining well-being (i.e., absence of strain)

(Payne, 2001). It asserts that individuals constantly appraise their environment and

that on an individual level, the perception of a mismatch between perceived demands

and the ability to cope with them constitutes stress.

As illustrated in Figure 2.1, a transactional model of stress involves two

important interacting processes: appraisal and coping.

This figure is not available online. Please consult the hardcopy

thesis available from the QUT Library.

Figure 2.1. Lazarus and Folkman’s (1984) transactional model of stress.

_____________________________________________________________________
Note. From Intervention in Occupational Stress (p.6), by R. R Ross and E. M. Altmaier, 1994, London:

Sage.

The appraisal process begins when a person evaluates a particular event,

situation, or demand (Ross & Altmaier, 1994). The process of cognitive appraisal

describes how a person construes an event and it is divided into two types: primary

appraisal and secondary appraisal (see Figure 2.1). Primary appraisal determines the

meaning an event has for the individual because it constitutes an assessment of the

relevance, significance, and implications of that event for that particular individual.

Primary appraisal is therefore about determining whether or not the individual has

anything at stake in the encounter.

 23

It is predicted that stress occurs when the environment is evaluated as harmful,

threatening, or challenging. An appraisal of harm means that damage has already

occurred, while an appraisal of threat refers to harm that will likely happen in the

future. Challenge refers to a condition of high demand in which the emphasis is on

mastering the demands, overcoming obstacles and growing and expanding as an

individual. In threat, the focus is on protecting against harm. In challenge, the

emphasis is on the positive outcome possibilities.

Secondary appraisal follows primary appraisal. Secondary appraisal involves

an assessment of the individual’s ability to cope with the harm, threat, or challenge. It

involves judging what resources the individual has access to and what options are

available in dealing with the event (Bartlett, 1998). The coping response process

refers to cognitive or behavioural efforts to deal with, tolerate or reduce the excess

demand. For instance, coping with work stressors may be eased by having resources

available in the work environment (Ross & Altmaier, 1994). Based on Lazarus and

Folkman’s (1984) model, it could be inferred that if a nurse’s colleagues can assist

him/her in deciding that the situation is not threatening but challenging, for example,

then less stress will result. Similarly, if a nurse believes that support is available at

work, then perceived coping resources are increased and therefore the threat

associated with the stressors will not be as harmful. The influence of work support is

described further in Chapter 4.

The transactional model is sometimes referred to as the

cognitive/transactional/phenomenological model of stress. The cognitive element

assumes that thinking, conscious awareness, memory, meaning and significance of

events are “the central mediators and the immediate causal agents in determining

stress and coping” (Bailey & Clarke, 1989, p. 21) to the perceiving individual.
 24

Applying this theory to the nurses, one may say that if he/she does not feel a ‘sense’

of threat in the stimulus situation, then there is no stress (Wheeler, 1997a). The

transactional component emphasises the continuous interaction of a person’s

cognitive perception or appraisal of the surrounding and constantly changing

environment which the individual shares with others. The phenomenological element

is suggestive of the personal, individualistic and idiosyncratic dimensions that are

involved in perceiving a situation. A person’s reaction to stress is thought to be

highly individualistic, that is, two nurses can react in completely different ways to the

same stressor (Wheeler, 1997a).

According to Ross and Altmaier (1994), the difficulty and the attractiveness of

this model is its flexibility. The model allows us to understand stress as the

combination of personal issues and concerns, which change over time, as well as the

resources and responses that a person can call upon in times of stress, which also

change over time. These responses, in turn, affect the initial situation or stressor, and

may cause us to appraise it, or think about it differently. Thus the stress response is a

transactional one, where the balance of demands and resources defines stress: if the

demands are greater than the resources, stress occurs. The reverse is also indicated: if

resources are available to meet the demand, then the appraisal might be one of

challenge rather than harm or threat, and thus would be less stressful to the individual.

A limitation of Lazarus and Folkman’s model of stress is that although it can

predict individual differences in the experience and reaction to stress, it cannot predict

which aspects of the work environment will be stressful. According to Lazarus

(1995), it is not that useful to identify conditions of work which adversely affect most

workers because stress is ultimately an individual phenomenon. Lazarus (1995, p.9)

states that “to describe and understand stress in the workplace requires that
 25

…individual patterns be studied to generate knowledge about the kinds of persons

who are more or less vulnerable to divergent sources of stress.” The current thesis

however, takes the position of Brief and George (1995) who acknowledge that stress

essentially occurs at an individual level, but believe it is useful to try to discover

working conditions which are likely to adversely affect most workers exposed to

them. By searching for those conditions of work likely to adversely affect most

workers exposed to them, a taxonomy of stressful job conditions may be constructed

and interventions to mitigate them can be devised.

In summary, a transactional model proposes that the individual, work

stressors, coping resources, and strains need to be considered jointly in explaining the

stress-strain-coping process because they are interdependent. Work stressors may be

defined as the “antecedent conditions within one’s job or the organisation which

require adaptive responses on the part of the employees” (Jex & Beehr, 1991, p.312).

The negative reaction to stressors is ‘strain’. The response to stressors is also affected

by individual differences between employees such as sociodemographic factors and

coping resources such as social support. The transactional model underpins the

present research program’s investigation by taking into consideration the influence of

social support on the stress-burnout relationship among nurses. In the section below,

the work stressors associated with nursing, as well as the methodological limitations

of recent nursing studies, are discussed.

Work Stress in Nursing

The nursing profession and the stress commonly associated with it has been

the subject of considerable research for decades. Researchers have been primarily

concerned with understanding the levels and sources of stress among various nursing
 26

departments or wards. In the section that follows, the prevalence and the major work

stressors associated with nursing will be discussed.

Prevalence of Work Stress

There is a widely held belief that nursing is one of the most inherently

stressful occupations (Rees & Cooper, 1992). However, empirical evidence providing

justification for this assertion is largely based on anecdotal evidence. According to

Wheeler (1997b), the assumption that stress is more common and more severe among

nurses than other professions may be the result of a high number of earlier studies

(Cassam & Hackett, 1972; Foxall, Zimmerman, Standle, & Bene, 1990; Hague, 1987;

Hipwell, Tyler, & Wilson, 1989; Oskins, 1979; Topf, 1989; Yu, Mansfield, Packard,

Vicary, & McCool, 1989) that concentrated on nurses’ stress in intensive care units

and other highly specialised areas (e.g., coronary care), rather than on more general

nursing wards. This was particularly the case during the 1970s and 1980s. A factor

common to these studies is that they focused on specialised critical care units which

accommodate patients who are very ill, highly dependent on acute care and who may

not be seen as suitable for general medical/surgical wards on the basis of their poor

state of health (Wheeler, 1997b). These studies may have led to the common, but

perhaps unfounded, view that nurses have higher levels of stress than other human

service occupations (Wheeler, 1997b).

Furthermore, it is difficult to comment on the prevalence of stress in nursing

due to methodological weaknesses in several of the studies on nurse stress, such as the

use of small-scale survey designs, the use of non-randomised samples, the absence of

objective measures and a preponderance of subjective evidence of stress (Wheeler,

1997b). In an attempt to identify the sources, frequency and degree of stress

experienced by British nurses, Wheeler (1994) and Wheeler and Riding (1994)
 27

conducted a study using 125 first-level general nurses and midwives. To determine

the prevalence of nurses’ work stress, a specific single item measure was used.

Respondents were asked: “In general, how stressful do you find the job you are in?”

Ratings were made on a five-point scale: not at all stressful, mildly stressful,

moderately stressful, very stressful, and extremely stressful. Responses were scored

from 0 to 4. They found that most nurses rated their degree of stress as either mild

(41.6%) or moderate (36%) with a much smaller percentage rating it as extreme

(16.9%). This finding suggests that although nurses are exposed to a wide variety of

potentially stressful events, not all nurses experience a high degree of stress. This

finding provides some support for Lazarus’ (1995) assertion that assuming adequate

coping resources are available, not all work stressors are perceived to be stressful.

Nevertheless, there can be little doubt, however, that nursing is, by its very

nature, an occupation subject to a high degree of stress (McGrath, Reid, & Boore,

1989). Furthermore, nursing is characterised by a number of stressors not commonly

experienced by most other professions. These include not only dealing with situations

involving death and dying, but also more mundane stressors such as working long

hours, working shifts and on weekends. It is the complex and demanding nature of

this profession that has encouraged much recent research on nurses’ sources and

severity of workplace stress. In the section below, studies examining the major

sources of work stress for nurses will be briefly discussed.

Sources of Work Stress for Nurses

Job-specific stressors. Studies of occupational stress in nurses have uncovered

a number of stressors specific to this profession. In the course of their careers, nurses

experience confrontation with severe or emergency illnesses (Grout, Steffan, &

Bailey, 1981), patient death, conflict with physicians (Gray-Toft & Anderson, 1981b;
 28

Jacobson, 1978; Kennedy & Grey, 1997; Walters & Hams, 1989), patients with

behaviour problems (Hartrick & Hills, 1993; Kennedy & Grey, 1997; Pagel &

Wittmann, 1986) work overload (Gray-Toft & Anderson, 1981b; Hingley & Cooper,

1986; Janssen, De Jonge, & Bakker, 1999; Kennedy & Grey, 1997; Schaufeli, 1990;

Schaefer & Moos, 1993; Wheeler & Riding, 1994) and inadequate preparation to meet

the emotional needs of patients and their family (Descamp & Thomas, 1993).

Much of the research conducted on Australian nurses’ work stressors has been

instigated by a national union for nurses, that is, the Australian Nursing Federation

(ANF). With 120,000 members, the ANF is the largest professional nursing

organisation in Australia. Members are employed in a wide range of enterprises in

urban, rural and remote locations in both the public and private sectors, including

hospitals, health, and community services, schools, universities, the armed forces,

statutory authorities, local government, off-shore territories, and industry. The union

represents Australian nursing internationally through links with other national and

international nursing organisations, professional associations and the International

Labour Organisations. The ANF is responsible for initiating activities nationally to

raise political awareness, and political action, if necessary, among members and the

general public in the pursuit of improved public policy on health and related issues.

In the late 1990s, the Victorian Branch of the ANF commissioned Considine

and Buchanan from the Australian Centre for Industrial Relations Research and

Training to investigate the issues associated with the working conditions of Victorian

nurses. A stratified random sample of 2161 union members participated in the study.

Considine and Buchanan (1999) reported that nurses’ major sources of stress were

related to hours and rostering, particularly the regular working of unpaid over time;

inadequate staffing, absence of trained and experienced staff; excessive workloads

 29

and increasing levels of responsibility. The majority of nurses indicated that they

worked overtime and only 19% indicated that they were always paid for it.

It is argued that these research findings concerning nurses’ working conditions

should be viewed with some caution. First, there is no published evidence outlining

the survey items, how the survey items were developed, and whether the survey met

principles of scientific rigor (e.g., reliability and validity). Second, the sample

population is not representative of the total nursing population as the surveys

investigating nurses’ working conditions were sent only to union members in public

and private acute and aged care facilities, public psychiatric facilities, high

dependency units and hospitals involved in Home Care Programs. It is for this reason

that the findings cannot be generalised to the wider nursing population in Australia.

Closer examination of the nursing literature indicates that there are few

empirical studies examining the main sources of stress for Australian nurses and what

impact this has had on nurses’ well-being. A literature search using the Cumulative

Index for Nursing and Allied Health Literature (CINAHL) database over the last

decade revealed a limited number of empirical studies investigating Australian nurses’

work stress (Bryant et al., 2000; Healy & McKay, 2000; Lumby, 1996).

Bryant et al. (2000) investigated workplace stress levels and

personal/workplace demographics using a sample of 170 urological nurses. A

questionnaire was designed by the authors based on a review of stress in nursing

literature. Content analysis of open-ended questions which sought details regarding

reasons for, and responses to, workplace stress yielded key response categories.

Excessive workload was the most prominently identified cause of work stress among

the sample of urological nurses. Excessive workload was reported four times more

frequently than any other stressor, and was reported by nearly 90% of the sample.
 30

When taken together, the empirical and qualitative findings suggested that a heavy

workload and poor staffing levels are the highest predictors of workplace stress. This

study also provides support for Brief and George’s (1995) assertion that although the

perception of stress is an individual phenomenon, employees exposed to the same

work environment perceive stressors in a similar manner.

Healy and McKay (2000) examined the relationships between nursing work-

related stressors and coping strategies in a volunteer sample of 129 registered nurses

from Melbourne metropolitan and Victorian regional institutions. Work stressors were

assessed using Gray-Toft and Anderson’s (1981a) NSS. Their results were consistent

with Bryant et al.’s (2000) findings. Healy and McKay revealed that workload was

the highest perceived stressor in the nurses’ working environment. The workload

subscale of the NSS taps into issues arising from the physical environment, such as

actual workload, inadequate staffing levels and insufficient time to complete work

tasks. Conflict with other nurses and lack of staff support were the least reported

stressors.

Lumby (1996) however, used a pedagogical approach to investigate nurses’

perceptions of work stress using 40 nursing students from intensive care, emergency

care, coronary care, operating theatre and high dependency units across several

teaching hospitals. Students were divided into four groups and each group was asked

to produce a map that identified themselves within micro and macro structures of their

world of work. The students were given no specific guidelines on how their final map

should be arranged. The group maps provided a medium for illustrating the nurse as

being the centre of a centrifuge of demands, inadequacies, frustrations and power

games leaving them feeling inadequate and with very little control despite an

immense responsibility. The groups included conditions of work such as shift work,
 31

poor skill mix, inadequate rostering, increased agency staff, the physical and

emotional demands of their work, the devaluing of their role in relation to the roles of

other health professionals (whose work they coordinate), increased use of technology

and noise, and safety aspects including the effect of their work on their physical and

emotional state. Major stressors related to lack of support by administrators at all

levels. Doctors were also identified as not providing support and indeed many groups

claimed that there was an increase in verbal abuse by some doctors. Other stressors

related to financial changes that affected adequate resourcing, including: out of date

and malfunctioning equipment, changes in types of patients and their length of stay.

The nurses reported that there was an increase in elderly patients being admitted, as

well as sicker patients, and the length of stay for patients was becoming shorter. This

means that the patients at any one time require a higher acuity. The lack of

involvement of experienced nurse clinicians in decision-making at all levels (apart

from that involving direct patient care) was depicted as a source of extreme frustration

within the group.

Over the last two decades, it would appear that whilst some of the main

sources of stress for nurses have changed, others such as work overload and lack of

recognition from doctors have persisted. For instance, in an earlier study conducted by

Linder-Pelz, Pierce, and Minslow (1987) using a sample of 983 Australian nurses

working across 14 hospital wards, non-responsiveness of management was ranked by

the nurses as the most severe work stressor for nurses. This was followed by parking

problems, work overload, nurses’ ambiguous and unsatisfactory status in the medical

care team, followed by a feeling of low professional esteem. The least stressful events

included environmental interferences with the real work of nursing, problems with

training requirements, and job prospects.

 32

Empirical studies investigating work stress among Australian nurses have had

diverse objectives and have used different methodologies, thus making comparisons

across studies problematic. As a result, making generalisations about the chronic

sources of work stress among Australian nurses is difficult. Most empirical studies,

(Cronin-Stubbs & Brophy, 1985; Dewe, 1989; Foxall et al., 1990; Gray-Toft &

Anderson, 1981b; Power & Sharp, 1988; Wheeler, 1994a, 1994b, 1998) however,

have consistently found that patient death and dying, workload, interpersonal conflict,

and role preparation are the key situational contexts in which stress occurs

(Farrington, 1995). For example, Hipwell et al. (1989) found that death and dying;

work overload; uncertainty over treatment; and stress from conflict with others (e.g.,

doctors) were the main sources of nurses’ stress.

Similar work stressors have been found in both Australian and international

studies, for example: Cohen-Mansfield (1989) (United States of America); Dewe

(1989) (New Zealand); Foxall et al. (1990) (United States of America); Gray-Toft and

Anderson (1981b, 1985) (United States of America); Healy and McKay (2000)

(Australia); Hipwell et al. (1989) (United Kingdom); Wheeler and Riding (1994,

1998) (United Kingdom). These consistent findings across studies seem to suggest

exposure to certain nursing stressors is universal and that nurses perceive work

stressors in similar ways. This finding also provides some support for Brief and

George’s (1995) assertion that there are some work stressors that adversely effect

most workers when exposed to them.

Wheeler (1998), however, offered an alternative explanation. Wheeler pointed

out that the reason for the presence of similar stressors across these studies may be

due to the consistent use of the NSS developed by Gray-Toft and Anderson (1981a).

A number of international studies have used this research instrument (or some
 33

variation of it) over the last decade to measure nurses’ levels of stress in a variety of

wards (Refer Table 2.1). Wheeler proposes that it is therefore likely that these studies

have identified roughly the same stressors.

Table 2.1

Studies Between 1993-2003 that have used Gray-Toft and Anderson’s (1981) Nursing

Stress Scale

Author(s) Year N Type(s) of Nursing Location

Lee 2003 362 Primary Care Hong Kong
Happell, Pinikahana, & Martin 2003 51 Forensic Australia
Escot, Artero, Gandubert, Boulenger, & 2001 37 Oncology France
Ritchie
Stordeur, Dhoore, & Vandenberghe 2001 625 General Belgian
Cole, Slocumb, & Mastey 2001 119 Critical Care United
States
Kilfedder, Power, & Wells 2001 510 Psychiatric Scotland
Payne 2001 89 Hospice United
Kingdom
French, Lenton, Walters, & Eyles 2000 2280 General Ontario
Agüir, Pons, Echegaray, Ramos, & 2000 201 General Spain
Sánchez
Healy & McKay 2000 129 General Australia
DePew, Gordon, Yoder, & Goodwin 1999 106 Burns and Special Care United
States
Kennedy 1999 125 Geriatric United
States
Hillhouse & Adler 1997 260 General United
States
Hillhouse & Adler 1996 297 General United
States
Collins 1996 113 General United
States
Duquette, Kérouac, Sandhu, Ducharme & 1995 1545 Geriatric France
Saulnier
Tyler & Cushway 1995 245 General United
Kingdom
Tyler & Ellison 1994 60 Theatre, Liver/Renal/ United
Haematology/Oncology, Kingdom
Elective Surgery
Descamp & Thomas 1993 72 General United
States
Wright, Blache, Ralph, & Luterman 1993 31 Medical/Cardiac Intensive United
Care, Burns, Neurotrauma States

The NSS consists of 34 items that describe situations that have been identified

from the literature and from interviews with nurses, physicians, and chaplains as

causing stress for nurses in the performance of their duties. The scale was validated
 34

using a relatively small sample of 122 American nurses from five departments

(medicine, surgery, cardiovascular surgery, oncology, and hospice) at a large private,

general hospital. These departments were chosen because their patients represented a

range of medical conditions requiring different types of nursing care and exposure to

various sources of stress. The NSS provides a total stress score as well as scores for

each of the seven subscales that measure the frequency of stress experienced by

nurses in a hospital environment. Seven major sources of stress were identified by

Gray-Toft and Anderson. These included: death and dying, conflict with physicians,

inadequate preparation, lack of support, conflict with other nurses, work load, and

uncertainty concerning treatment. Gray-Toft and Anderson reported internal

consistency coefficients ranging from .79 to .89 for each subscale.

A significant limitation of recent empirical studies that use the NSS to

examine work stress among Australian nurses is that it is not known whether the work

conditions and the associated demands placed upon Australian nurses working in

public hospitals are similar to American nurses working in private hospitals. Bruhn

(1991) argued that it is not productive to use a measure merely because it is used or

quoted widely. It is important to know what one wants to learn, before deciding how

to measure it (Bruhn, 1973, 1991). This crucial factor appears to suggest the need for

further clarification of the sources of stress among Australian nurses working in the

public hospital context. One of the main purposes of Study 1 is to address this gap in

the nursing stress literature. In Study 1, qualitative data will be collected by

conducting focus groups with nurses to further our understanding of the key work

stressors Australian nurses are exposed to working in a public hospital.

Role stressors. Although nursing studies predominantly investigate job-

specific stress, such as those measured by Gray-Toft and Anderson’s (1981a) NSS,
 35

other empirical studies have taken a more generic approach by measuring role stress.

Role stress is considered to be characteristic of most occupations, especially white

collar occupations such as nursing. It may be viewed as the consequence of disparity

between an individual’s perception of the characteristics of a specific role and what is

actually being achieved by the individual currently performing the specific role

(Lambert & Lambert, 2001). The concept of role stress, originally advanced by Gross,

Mason and McEachern (1958), refers to a condition where role obligations (i.e.,

expected, and actual behaviours associated with a job position) are vague, irritating,

difficult, conflicting, or impossible to meet (Hardy & Hardy, 1988). Role stress is

now recognised to comprise of three distinct types: role conflict, role ambiguity, and

role overload (Iwata & Suzuki, 1997; Kahn, Wolfe, Quinn, Snoek, & Rosenthal,

1964).

Role conflict occurs for an individual when a person in the work environment

communicates an expectation about how he or she should behave and this expectation

makes it difficult or impossible to fulfil another behavioural expectation or set of

expectations. More simply, stress is caused by the inability to meet or difficulty in

meeting the various expectations of behaviour (Van Sell, Brief, & Schuler, 1981).

Role ambiguity results when there is inadequate, unclear, or confusing information

about expected role behaviours (Van Sell et al., 1981). Role overload is a final form

of role conflict, analogous to work overload. Most studies fail to differentiate between

work overload and role overload. Quick, Quick, Nelson, and Hurrell (1997) however,

suggest that role overload can be distinguished from work overload in that workload

is based on actual tasks, whereas role overload is based on the behaviours that are

expected of the individual. Role overload occurs when too many behaviours are
 36

expected of an individual or the behaviour expected is too complicated or difficult for

the individual to execute (Quick et al., 1997).

Research has indicated that role stress is inherent in the nursing role (Baba,

Galperin, & Lituchy, 1999; Beehr, King, & King, 1990; Chang & Hancock, 2003;

Gil-Monte, Valcarcel, & Zornoza, 1995; Gray-Toft & Anderson, 1985; Lambert &

Lambert, 2001). It is reasoned that changes in the health care structure and the nature

of patient treatment such as managed care, reductions in length of hospital stays,

increased patient acuity, new technologies and a focus on cost effective quality of care

approach to nursing may have contributed to higher levels of role stress for nurses

(Garrett & McDaniel, 2001). A literature review conducted by Lambert and Lambert

(2001) found that most studies investigating role stress in nursing have found work

environment factors to be involved. For example, the experience of role stress is

related to having little control in one’s job, high job demands and lack of support from

peers (Chapman, 1993; Cheng, Kawachi, Coakley, Schwartz, & Colditz, 2000; Fong,

1993; Glass, McKnight, & Validmarsdottir, 1993; Webster & Hackett, 1999). Other

studies have found that being required to work on different wards, lack of essential

resources including nursing staff, and work overload to be major factors (Foxall et al.,

1990; Frisch, Dembeck, & Shannon, 1991; Hatcher & Laschinger, 1996; Murray,

1998; Snape & Cavanagh, 1993).

Empirical studies have consistently found that nurses who are stressed have

higher absenteeism rates, lower work satisfaction and are more likely to leave the

organisation (Callaghan & Field, 1991; Larson, 1987). They have more conflicts with

coworkers (MacNeil & Weisz, 1987) which only escalates the problem as aggression

from colleagues has been found to be a major source of stress for nurses (Farrell,

1999). They also suffer psychologically, in terms of feelings of inadequacy, self-

 37

doubt, lowered self-esteem, irritability, depression, somatic disturbance, sleep

disorders and burnout (Foxall et al., 1990). In the following chapter, the relationships

between work stressors (i.e., job-specific stressors and role stressors) and burnout

(i.e., emotional exhaustion, depersonalisation, and reduced personal accomplishment)

among nurses will be examined in detail.

In summary, there is a lack of published research regarding the main sources

of stress among Australian nurses working in public hospitals. In Study 1, focus

groups with nurses from a variety of hospital wards will be conducted to further our

understanding of the work events that Australian nurses perceive as stressful. The

focus group discussions are designed to address the following research question and

its associated sub-problem:

What are the chronic sources of work stress commonly experienced among

Australian nurses?

To what extent are the sources of stress experienced by Australian nurses

consistent with the seven stress factors measured by Gray-Toft and Anderson’s

(1981a) NSS?

A benefit of conducting focus groups with nurses is that the researcher is able

to collect rich qualitative data without the constraints associated with using an already

established nurse stress measure that assesses a predefined and limited number of

stressful events. The data obtained from the focus groups will assist the researcher in

choosing the most appropriate tool(s) to measure Australian nurses’ occupational

stress in Study 2. The findings will also extend current nursing literature by

confirming whether the sources of perceived stress by Australian nurses are consistent

with those that are experienced by overseas nurses.

 38

Levels of Work Stress Across Nursing Wards

A further issue which has been of appeal to researchers examining stress in

nursing has been to assess the topic in relation to specific areas of clinical practice, for

example critical care nursing or psychiatric nursing (Kirkcaldy & Martin, 2000). A

prominent theme often underlying this research is that certain hospital environments,

particularly intensive care units (ICUs) and emergency departments, are potentially

more stressful to nurses than others. For instance, Foxall et al. (1990) found that ICU

and hospice nurses differed in terms of frequency and sources of job stress from

medical and surgical nurses. Their findings indicated that death and dying situations

were the most stressful to the former groups, with work overload and staffing

situations proving the most stressful to medical and surgical nurses (Kirkcaldy &

Martin, 2000).

While some earlier studies suggest that critical care nurses experience more

stress than general medical or surgical nurses, this view is not commonly supported

by more recent studies. Although studies have consistently found that levels of stress

do not differ between wards, some sources of stress appear to be specific to

departments. For instance, Tyler and Ellison (1994) examined occupational stress

using the NSS in four areas of high dependency nursing: theatre, liver/renal,

haematology/oncology and elective surgery. They found that the level of stress

experienced by nurses was similar across all four departments, but their sources of

stress varied. In particular, theatre nurses experienced less stress through patients’

death and dying. Managing workload was most stressful for nurses in theatre and

haematology. Inadequate preparation was a greater stressor for liver and elective

surgery wards than for theatres and haematology.

 39

Recent studies investigating Australian nurses’ stress have not examined

whether nurses from different wards perceive and experience work stressors in a

similar way. In an earlier study, however, Linder-Pelz et al. (1987) found that

intensive care was not the most stressful type of nursing work. Intensive care nurses

followed spinal, maternity, and medical wards in the proportion of its nurses who

were excessively stressed. However, the mean stress scores in these four areas did not

differ significantly. Further, while lack of management responsiveness, unsatisfactory

team status and poor job prospects were relatively severe stressors in the intensive

care unit, they were not unique to these units. These findings provide further evidence

that nurses have similar perceptions of stress. Although not the focus of Study 1, the

researcher will take the opportunity to examine the following sub-problem:

Do Australian nurses from various hospital wards share similar perceptions of

work stress?

Kirkcaldy and Martin (2000) revealed that overall, the results of research have

failed to find consistent or conclusive evidence that department type is itself a direct

cause of stress in nursing. Current evidence suggests that similarities among nurses

working in different environments are more striking than differences between them,

with the pervasiveness of certain stressors such as workload or death and dying

outweighing the influence of department type or specialism. However, it is important

to note that many of the studies (e.g., DePew et al., 1999; Linder-Pelz et al., 1987;

Wright et al., 1993) from which this inconsistent conclusion has been drawn have

often featured a small number of nursing groups, frequently two or three, and have

featured small samples often below 100 participants.

 40

Levels of Work Stress Based on Gender

Whilst the emphasis of nursing studies has been the investigation of the main

sources of nurses’ work stress, some studies have also examined perceptions of work

stress according to gender. Overall, empirical research suggests that nurses’ levels and

sources of work stress are not significantly influenced by gender (Baba et al., 1999;

Kirkcaldy & Martin, 2000; Tyler & Ellison, 1994). It has been suggested that any

variation attributed to gender may not be found in nursing studies since the sample

size of males is typically small. For instance, in Australia, only 10% of the nursing

workforce are males. However, wider empirical literature has also been unable to

account for very much variance in occupational stress based on gender. Martocchio

and O’Leary (1989) undertook a meta-analysis of 15 studies that examined gender

differences in occupational stress and found no differences in experienced and

perceived stress. Emslie, Hunt, and Macintyre (1999) propose that when men and

women in the same professions are compared, gender accounts for little of the

variance in stress in comparison to working conditions. Furthermore, Matuszek,

Nelson, and Quick (1995) reported that men and women operating at roughly the

same level within organisations generally experience the same stressors. Greenglass

(1991) suggested that gender differences in work stress are confounded by

occupational role and hierarchical position and that when these variables are

controlled for, no significant gender differences are apparent. In light of these

findings, it could be inferred that men and women differ very little in the way they

appraise potentially stressful events. It is hypothesised in the present research study

therefore that:

Hypothesis 1: Nurses’ levels of work stress will not be significantly different

based on gender.
 41

Summary

In summary, the present thesis utilises a transactional approach to explore the

stress process. Although the study recognises that stress occurs at an individual level,

the researcher supports the view that it is important to consider nurses’ perceptions of

stress at a group level in order for appropriate work stress interventions to be

developed to address this issue. A multitude of work stressors have been identified in

the nursing literature. Whilst some stressors are considered pertinent to the nursing

profession (e.g., death and dying, interpersonal conflict, lack of support), other

stressors are applicable to most occupational roles (e.g., role overload, role conflict,

and role ambiguity). The present research program will consider nurses’ work

stressors in relation to burnout. Furthermore, nurses’ perceptions of work stress based

on nursing wards and nurses’ levels of work stress based on gender will also be

explored. In the following chapter, existing knowledge of the main work stress

determinants of burnout are reviewed.

 42

CHAPTER 3

A Review of the Literature on Burnout

The purpose of Chapter 3 is to take stock of the existing empirical knowledge

regarding burnout. More specifically, the chapter briefly examines how the burnout

construct has evolved to become the readily recognised three-component

psychological syndrome, as it is known today. In addition, the chapter clearly

distinguishes burnout from occupational stress and explores current developmental

models of burnout. Next, the prevalence of burnout among nurses is considered before

discussing the primary antecedents of burnout. Hypotheses regarding the

interrelationships among the burnout components and the specific determinants of

burnout are proposed.

Background

Burnout first emerged as a social problem, not as a scholarly construct. Thus

the initial conception of burnout was shaped by pragmatic rather than academic

concerns (Maslach & Schaufeli, 1993). In its pioneering phase of conceptual

development, the focus was on clinical descriptions of burnout. Later there was a

second empirical phase in which the emphasis shifted to systematic research on

burnout and in particular the assessment of the phenomenon.

The first few articles about burnout appeared in the mid-1970s in the United

States (Freudenberger, 1974, 1975; Maslach, 1976). The significance of these first

articles was that they provided the initial description of the burnout phenomenon,

gave it its name, and showed that this psychological syndrome was relatively common

among human service professionals. Freudenberger, a psychiatrist who was employed

in an alternative health care agency, was the first to note that many of the volunteers

with whom he was working experienced a gradual emotional depletion and loss of
 43

motivation and commitment. At about the same time, Maslach, a social psychology

researcher, was studying the ways in which people cope with emotional arousal on the

job. She was particularly interested in such cognitive strategies such as ‘detached

concern’ and ‘dehumanisation in self-defence.’ From interviews with human service

professionals emerged the realisation that the emotional stress inherent in these

occupations could be harmful and debilitating. When by chance she described the

results of her research to an attorney, she was told that poverty lawyers called this

particular phenomenon ‘burnout’. Maslach and her colleagues (Maslach, 1982;

Maslach & Jackson, 1981; Pines & Maslach, 1980) subsequently adopted this term,

and they discovered that it was immediately recognised by their interviewees; thus a

new colloquial expression was born.

Much of the work during this time consisted of personal experiences (e.g.,

Freudenberger, 1974, 1977a, 1977b) or narratives based on specific programs or case

studies (e.g., Maslach & Pines, 1977; Pines & Maslach, 1978, 1980). Perlman and

Hartman (1982) compiled a listing of the multiple conceptualisations used during this

period. Definitions of burnout included: (a) to fail, to wear out, become exhausted; (b)

a loss of creativity; (c) a loss of commitment for work; (d) an estrangement from

clients, coworkers, job and agency; (e) a syndrome of inappropriate attitudes toward

clients and toward self, often associated with uncomfortable physical and emotional

symptoms. Although these notions of burnout were similar, they lacked a common

and precise measure of burnout.

During the next phase of the 1980s, the work on burnout entered a more

focused, constructive, and empirical period. Many books and articles were written

about burnout, in which the authors outlined their working models of the

phenomenon, proposed various ideas and interventions, and presented various forms
 44

of corroborative evidence (questionnaire data, interview responses and clinical case

studies). Standardised measures of burnout were developed, thus providing

researchers with more precise definitions and methodological tools for studying the

concept. In particular, the development and widespread acceptance of the Maslach

Burnout Inventory (MBI; Maslach et al., 1996) fostered systematic research on

burnout, resulting in an increased number of articles published in scholarly journals.

In summary, burnout is not a new phenomenon – it has its roots in the past

(Schaufeli & Enzmann, 1998). The development of burnout as a psychological notion

took place along two lines. Initially, in the pioneer phase, a clinical approach

prevailed that was characterised by merely describing the symptoms of the burnout

syndrome. In the second empirical phase, social and organisational psychologists

studied burnout more systematically, using standardised measures. In the section

below, the definition and measurement of burnout are further elaborated.

Definition and Measurement of Burnout

Today, the most commonly accepted definition of burnout is the three-

component conceptualisation proposed by Maslach and her colleagues (Maslach,

1982; Maslach & Jackson, 1981; Pines & Maslach, 1980). Maslach et al. (1996, p.4)

define burnout as “a syndrome of emotional exhaustion, depersonalisation, and

reduced personal accomplishment that can occur among individuals who work with

people in some capacity.” According to Maslach (1993), emotional exhaustion

reflects the stress component of burnout, whereas depersonalisation captures a

dimension of interpersonal relations, and reduced personal accomplishment

incorporates a dimension of self-evaluation (Schaufeli & Enzmann, 1998).

The first dimension, emotional exhaustion, has been the most extensively

studied factor in the burnout literature. It is characterised by a lack of energy and a

 45

feeling that one’s emotional resources are used up (Cordes & Dougherty, 1993). This

‘compassion fatigue’ may coexist with feelings of frustration and tension as workers

realise they cannot continue to give of themselves or be as responsible for clients as

they had in the past (Cordes & Dougherty, 1993). A common symptom is dread at the

prospect of returning to work the next day.

The second component, depersonalisation, adds an interpersonal dimension.

Depersonalisation refers to the development of negative, callous, and cynical attitudes

towards the recipients of one’s services (Schaufeli & Enzmann, 1998). Workers may

display a detached and an emotional callousness, and they may be cynical to

coworkers, clients, and the organisation (Cordes & Dougherty, 1993). Although a

certain degree of psychological distance may be necessary and even beneficial when

dealing with stressful and highly arousing situations, too much detachment may result

in the individual developing negative attitudes towards his or her clients/patients

(Parker & Kulik, 1995). Visible symptoms include the use of derogatory or abstract

language (e.g., the ‘kidney’ in room 212), strict compartmentalisation of professional

lives, intellectualisation of the situation, withdrawal through longer breaks or

extended conversations with coworkers, and an extensive use of jargon (Maslach &

Pines, 1977). While the human service professional may still feel concern, they can no

longer give of themselves as they had formerly (Maslach & Jackson, 1986).

The final component of burnout involves feelings of reduced personal

accomplishment. It is characterised by a tendency to evaluate oneself negatively. In

other words, the individual experiences a decline in his or her feelings of job

competence and successful achievement in their work or interactions with people

(Cordes & Dougherty, 1993). Frequently there is the perception of a lack of progress

or even lost ground. Feelings of diminished personal accomplishment may result from
 46

factors suggesting one is unappreciated or that one’s efforts are ineffective (Jackson,

Turner, & Brief, 1987), or from factors that suggest one’s competence or performance

is low (Burke, Shearer, & Deszca, 1984).

The popularity of this definition and its almost universal acceptance at this

time has corresponded with the growth and acceptance of the MBI (Maslach et al.,

1996). Its popularity relates to its ability to assess the three dimensions of burnout as

proposed by Maslach (1982) and also evidence supporting its reliability and validity

in assessing these dimensions. Today, the MBI is in its third revision. The instrument

and consequently the definition it supports, has been strengthened by demonstrations

of factorial validity and stability across professions and cultures (Gil-Monte & Peiro,

1999; Schaufeli & Van Dierendonck, 1993). Due to its psychometric soundness, the

current research also uses the MBI to measure burnout among nurses. Its

psychometric properties are discussed further in Chapter 5.

Despite the growing consensus surrounding the concept of burnout, the

distinction between burnout and stress has not been clearly delineated in the literature.

Examination of the literature suggests that the concepts of stress and burnout are

closely related. Despite similarities in theoretical assumptions and models, burnout

may be distinguished from occupational stress. In the section below, the concept of

burnout will be distinguished from job stress.

Burnout – An Outcome of Job Stress

Stress and burnout are generally conceptualised as the product of a complex

transaction between individual needs and resources and the various demands within

an individual’s immediate environment (Handy, 1990). Furthermore, both are

generally conceptualised as highly subjective phenomena in which perceived stressors

are more important than actual environmental conditions (e.g., Caplan, 1983; Cox,
 47

1978; Golembiewski, Munzenrider, & Stevenson, 1986; Maslach, 1982). It would

appear that models of occupational stress and burnout are concerned with the stressed

individual within the workplace, however models of occupational stress

predominantly analyse causes of stress whereas models of burnout predominantly

investigate the individual’s response to occupational stress. As Ganster and

Schaubroeck (1991) argued, burnout is, in fact, a type of stress – specifically a

chronic affective response pattern to stressful work conditions that feature high levels

on interpersonal contact.

In addition, burnout can be distinguished from occupational stress with respect

to time. Burnout can be considered a response to prolonged job stress, that is, when

the demands at the workplace tax or exceed an individual’s resources (Maslach &

Schaufeli, 1993). This longer time perspective is also implied in its terminology:

burning out (depleting ones resources) is a long-term process (Schaufeli, Maslach, &

Marek, 1993). A remarkable parallel exists with the work of Seyle (1967). Seyle

suggested that after prolonged exposure to stress, the physiological resources are

depleted, and irreversible damage is caused to the organism (Maslach & Schaufeli,

1993). Referring to Seyle’s adaptation syndrome, Etzion (1987) argued that burnout is

a latent process of psychological erosion resulting from prolonged exposure to stress.

Typically, exhaustion is reached before the individual consciously has noticed both

the preliminary stages: alarm and resistance. Brill (1984) has also conceptualised

burnout as the result of long-term exposure to job stress. According to Brill, stress

refers to an adaptation process that is temporary and is accompanied by mental and

physical symptoms, whereas burnout refers to a breakdown in adaptation

accompanied by chronic malfunctioning. It would appear that burnout is response-

 48

based and concentrates on the mental health and work outcomes of extended job

stress.

Specifically, stress is a generic term that refers to the temporary adaptation

process that is accompanied by mental and physical symptoms (Schaufeli, 1999). In

contrast, burnout can be considered as a final stage in a breakdown of adaptation that

results from the long-term imbalance of demands and resources. Moreover, burnout

includes the development of dysfunctional attitudes and behaviour towards recipients

(depersonalisation), the job, and the organisation, whereas job stress is not necessarily

accompanied by such attitudes and behaviours. This assertion is empirically supported

by Schaufeli and Van Dierendonck (1993) who showed in a sample of nurses that

burnout can be distinguished from job-related distress, despite emotional exhaustion

sharing about 30% of its variance with distress. The fact that depersonalisation and

reduced personal accomplishment, however, are less substantively related to distress

implies that burnout is a unique, multidimensional, chronic stress reaction that goes

beyond the experience of mere exhaustion. Finally, it has been argued that everybody

can experience stress, while burnout can only be experienced by those who entered

their careers enthusiastically with high goals and expectations. Pines (1993) proposed

that individuals who expect to derive a sense of significance from their work are

susceptible to burnout, whereas those without such expectations would experience job

stress instead of burnout.

In summary, stress and burnout may not be distinguished on the basis of

symptoms, but only on the basis of process (Schaufeli et al., 1993). While job stress

and burnout are clearly linked, they are not identical constructs. Overall, researchers

agree that burnout is a pattern of responses that develops as a result of prolonged

exposure to stressors at work. The sequencing of the three components of burnout or

 49

the process of burnout has been conceptualised differently among researchers of

burnout. There are at least three competing models of burnout. In the section below,

the development of the three burnout components are considered.

Burnout as a Process

The original burnout sequencing was proposed by Maslach (1982). Maslach

assumes that burnout is a sequential process that starts with emotional exhaustion

resulting from the emotional demands of dealing with clients. Perhaps as a defensive

coping strategy, they then limit their involvement with others and distance themselves

psychologically from clients (Cordes & Dougherty, 1993). Depersonalisation is

thought to provide an emotional buffer between the individual and the imposing

coping demands. Finally, the individual recognises the discrepancy between their

current attitude and their original optimistic expectations about the potential

contributions to society and to the organisation. As a result, individuals experience a

sense of inadequacy in terms of their ability to relate to people and to perform their

jobs and gradually a sense of diminished personal accomplishment develops

(Schaufeli & Enzmann, 1998). This model is consistent with the view that strain leads

to avoidance behaviours and, in turn, reduced effectiveness (Gaines & Jermier, 1983).

Maslach’s (1982) model of burnout is depicted in Figure 3.1.

 50

This figure is not available online.Please consult the

hardcopy thesis available from the QUT Library.

Figure 3.1. Maslach’s (1982) model of burnout.

An alternative model of the burnout process is a sequence that was advanced

by Golembiewski and Munzenrider (1981, 1984, 1988) and Golembiewski (1989).

Although Golembiewski and Munzenrider agree with the three-dimensional nature of

burnout as proposed by Maslach, they provide a rather more strongly modified

process of burnout (Schaufeli & Enzmann, 1998). Golembiewski and Munzenrider

suggest that in human service settings, depersonalisation comes first, which leads to

reduced personal accomplishment and subsequent emotional exhaustion. They

maintain that depersonalisation is first experienced because a certain degree of

professional detachment is often functional in dealing with others in a more

‘objective’ manner and is often reinforced by one’s peers and superiors (Lee &

Ashforth, 1993). Beyond a certain degree, however, detachment becomes

depersonalisation; thus impairing the ability to form necessary relationships with

others and undermining performance (Lee & Ashforth, 1993). As depersonalisation

deepens and one’s sense of accomplishment diminishes, work stress may surpass
 51

one’s ability to cope, leading to emotional exhaustion. Golembiewski and

Munzenrider’s model of burnout is depicted in Figure 3.2.

This figure is not avaible online. Please consult the

hardcopy thesis available from the QUT Library.

Figure 3.2. Golembiewski and Munzenrider’s (1988) model of burnout.

Golembiewski and Munzenrider’s model, however, has been criticised by

researchers (e.g., Ashforth & Lee, 1997) for not articulating a detailed or compelling

theoretical rationale for their sequence model. According to Maslach’s model,

depersonalisation is argued to be a means (albeit futile) of stanching the flow of

emotional energy, of coping with growing exhaustion. Golembiewski and

Munzenrider do not make it clear how and why depersonalisation develops. Indeed

human service professionals are often socialised to assume a stance of detached

concern toward their clients, and bureaucratic structures and cultures may reinforce

detached concern. However, it is not apparent why this detachment would slide into

outright depersonalisation for service providers in the absence of at least some strain

(Ashforth & Lee, 1997). Second, Maslach (1982) argues that emotional exhaustion

often stems from unrealistic desires to solve the frequently intractable problems of

clients. Indeed, the more idealistic the service provider, the greater the risk of burnout.

In Golembiewski and Munzenrider’s model, the provider is already retreating from a

 52

commitment to his or her clients through depersonalisation and low personal

accomplishment. It is not clear why emotional exhaustion would develop at this point

rather than earlier. Furthermore, on the empirical front, research consistently indicates

that emotional exhaustion and depersonalisation are far more strongly correlated than

is personal accomplishment with emotional exhaustion or depersonalisation. For

instance, Lee and Ashforth (1996) conducted a meta-analysis of 61 studies and found

that emotional exhaustion was strongly positively related to depersonalisation (r =

.64), whereas both dimensions were moderately negatively related to personal

accomplishment (r = -.33 for emotional exhaustion and r = -.36 for

depersonalisation). Cordes, Dougherty, and Blum (1997) reported similar findings.

Emotional exhaustion was more strongly correlated with depersonalisation (r = .58),

than was personal accomplishment with either emotional exhaustion (r = -.26) or

depersonalisation (r = -.19). These associations cast doubt on Golembiewski and

Munzenrider’s argument that (low) personal accomplishment in effect mediates the

association between depersonalisation and emotional exhaustion. Recent evidence

suggests that the three-factor sequential model as conceived by Maslach is slightly

superior to Golembiewski and Munzenrider’s model of burnout (Cordes et al., 1997;

Lee & Ashforth, 1993).

These relatively weak associations between emotional exhaustion and

depersonalisation with personal accomplishment prompted Leiter (1993) to speculate

that personal accomplishment follows an independent but parallel process to

emotional exhaustion and depersonalisation. Leiter (1990, 1991, 1993) conducted a

series of studies in which he distinguished between quantitative job demands (e.g.,

workload), qualitative job demands (e.g., interpersonal conflict) and lack of resources

(e.g., lack of social support, lack of opportunities for skill enhancement, poor client
 53

cooperation, lack of autonomy and participative decision making). Demands were

expected to be related to emotional exhaustion, whereas resources were expected to

be related to depersonalisation and lack of personal accomplishment. Leiter (1990,

1991, 1993) proposed that demanding aspects of the work environment aggravate

exhaustion, which in turn contributes to increased depersonalisation, while the

presence of resources influences personal accomplishment. The results largely

confirmed this hypothesised model, although the model was based on cross-sectional

data. Contrary to Maslach’s model, however, personal accomplishment seemed to

develop rather independently from emotional exhaustion and depersonalisation (see

Figure 3.3).

Interpersonal Workload
Demands

Emotional
Exhaustion

Depersonalisation
Lack of
Resources
Reduced Personal
Accomplishment

Figure 3.3. Leiter’s (1993) model of burnout.

_______________________________________________________________

Note. From The burnout companion to study and practice: A critical analysis (p.118), by W.

Schaufeli and D. Enzmann, 1998, Philadelphia: Taylor & Francis.

Leiter (1993) described reduced personal accomplishment as developing in

parallel with the other two burnout dimensions, provided that resources were lacking.
 54

Lee and Ashforth’s (1996) concluded from their meta-analysis that their results were

largely consistent with Leiter’s (1993) mixed sequential and parallel development

model of burnout. They found that emotional exhaustion is particularly related to job

demands (e.g., role problems, workload, work pressure, stressful events, unmet

expectations), whereas poor personal accomplishment is related to lack of resources

(e.g., poor social support). Depersonalisation appears to be related to both job

demands as well as lack of resources.

In contrast to the original Maslach model, which proposed an entirely internal

model in which emotional exhaustion triggered the whole burnout syndrome in

response to environmental stressors, Leiter (1993) suggests the relationship between

emotional exhaustion and personal accomplishment exists external to individuals. It

exists within their social context, in the provision of organisational resources, in the

personal conflict among colleagues and service recipients, and in the pressure of

emotional demands. From this perspective, the mild but persistent correlation between

emotional exhaustion and personal accomplishment (Maslach & Jackson, 1986) arises

because of consistent relationships among aspects of work environments: work

settings characterised by excessive work loads and personal conflict are often

experienced by their participants as lacking in organisational supports necessary for

effective professional functioning (Cherniss, 1980; Hackman, 1986). According to

Leiter (1993) work settings that persistently deplete the emotional energy of its

personnel are likely to provide inadequate resources in terms of social support or

opportunities for professional development.

At this point in time, it cannot be confirmed or disconfirmed whether Leiter’s

revised sequencing model is a more accurate model of burnout. Further studies need

to be replicated with similar samples to ensure that the obtained effects are in fact
 55

significant and that they are not due to chance. While the revised model appears to

better explain the data than the original Maslach model, other possible models could

fit the data equally well. Any inference about the casual relations at this stage among

the burnout variables would be premature. It may be that there is no strict order in

which the three burnout dimensions affect each other (Janssen, Schaufeli, & Houkes,

1999). What is known, however, is that burnout is a process and the three components

of burnout influence one another over time. Furthermore, the three burnout

components have distinct relationships with environmental conditions (Leiter, 1991).

In the present thesis, the three burnout components will be explored independently in

relation to a range of work stressors. Taking into consideration the previous

discussion, the following hypothesis was formulated.

Hypothesis 2: Burnout consists of three related, but empirically distinct

components – emotional exhaustion, depersonalisation, and personal accomplishment.

Prevalence of Burnout Among Nurses

Several researchers (e.g., Adali & Priami, 2001; Cordes & Dougherty, 1993;

Duquette et al., 1994; Greenglass et al., 2001; Kilfedder et al., 2001; Kirkcaldy &

Martin, 2000) have concluded that nurses have a high risk of burning out. It could

therefore be assumed that the prevalence of burnout in the nursing profession is high.

For instance, two European epidemiological studies reported that burnout affects

approximately 25% of all nurses (Landau, 1992; Saint-Arnaud et al., 1992). However,

other empirical studies have suggested that the prevalence of burnout among nurses is

much smaller. For instance, Bourbonnais, Comeau, Vézina, and Dion (1998) using a

sample of 1,891 Canadian nurses, reported that the prevalence of burnout (i.e., high

levels of emotional exhaustion and depersonalisation and low levels of personal

accomplishment) was 8.4%. Kilfedder et al. (2001) using a randomly selected sample
 56

of 510 Scottish nurses, found that only 2% of their study sample could be categorised

as having high burnout.

In reality it is difficult to determine the prevalence of burnout among nurses

working within a hospital environment. Although the majority of studies investigating

burnout levels have utilised Maslach et al.’s (1996) MBI, some authors have preferred

to use an alternative such as Pines and Aronson’s (1981) Tedium Scale (e.g.,

McCranie, Lambert, & Lambert, 1987), Jones’ (1980) Staff Burnout Scale for Health

Professionals (e.g., Duquette et al., 1995) and Ebbinghaus’ (1996) Oldenburg Burnout

Inventory (e.g., Demerouti, Bakker, Nachreiner, & Schaufeli, 2000). In addition, the

purpose and aims of the studies have varied widely, and therefore heterogeneous

samples have been used. Most studies do not report the overall level of burnout

among nurses. Studies more commonly report the percentage of the sample

population scoring low, moderate and high levels of emotional exhaustion,

depersonalisation, and personal accomplishment (McGrath et al., 1989). Another

limitation of more recent studies on burnout in nursing is that researchers have

predominantly directed their attention to psychiatric or mental health nurses, at the

exclusion of general nurses (e.g., Edwards, Burnard, Coyle, Fothergill, & Hannigan,

2000a, 2000b; Fagin, Brown, Bartlett, Leary & Carson, 1996; Kilfedder et al., 2001;

Leiter & Harvie, 1996). To gauge the degree of burnout in the nursing industry,

descriptive studies that investigated general nurses working in a hospital environment,

that used the MBI, and reported the mean scores for the three burnout dimensions are

examined. The level of burnout was established using the norms provided in the MBI

manual (Maslach et al., 1996).

Earlier studies (Ceslowitz, 1989; Maslach & Jackson, 1981; Topf, 1989)

investigating general nurses working in a hospital setting found that nurses

 57

consistently reported mean scores for emotional exhaustion, depersonalisation and

personal accomplishment that fell in the moderate range (i.e., the middle third of the

MBI normative distribution). In a study of 62 medical nurses and 66 surgical nurses

however, Ogus (1992) found that medical and surgical nurses reported high levels of

emotional exhaustion and depersonalisation but still reported high levels of personal

accomplishment.

In a more recent study, Schmitz, Neumann, and Oppermann (2000) studied a

convenience sample of 366 nursing staff from five German hospitals across nine

wards and found that the mean scores for emotional exhaustion and personal

accomplishment were moderate, however depersonalisation was high. Iacovides,

Fountoukalis, Moysidou, and Ierodiakonou (1997) using 368 nurses from a general

hospital in Greece, found that nurses reported moderate levels of emotional

exhaustion, depersonalisation, and personal accomplishment. Butterworth, Carson,

Jeacock, White, and Clements (2000), in a study of British and Scottish nurses (n =

586), found that ward nurses reported moderate levels of emotional exhaustion,

depersonalisation, and high levels of personal accomplishment. Finally, Turnipseed

and Turnipseed (1997) compared the burnout level of 129 American nurses to 71

nurses from the Philippines. Their results indicated that overall, American nurses

reported moderate levels of emotional exhaustion, depersonalisation and personal

accomplishment. In contrast, nurses from the Philippines reported borderline

low/moderate emotional exhaustion, moderate depersonalisation, and personal

accomplishment at the margin of moderate to high.

A thorough examination has revealed that there are no recent Australian

studies that have investigated the degree of burnout among general nursing staff

working in public hospitals. In the present study, the mean MBI scores will be
 58

generated for each component of burnout. It is assumed that the level of burnout

among Australian nurses will be similar to that found in American, British, and

European studies. It is therefore expected that:

Hypothesis 3: Australian nurses will report moderate levels of emotional

exhaustion, depersonalisation, and personal accomplishment.

A further limitation of most nursing research is that studies have failed to

consider how nurses’ burnout levels compare to other human service professions. This

makes it difficult to comment on whether nurses are in fact at a greater risk of

experiencing burnout in comparison to other human service professions. Schaufeli

and Enzmann (1998) however, analysed a MBI data set based on 73 US studies

including over 17,000 respondents that were published between 1979 and 1998. They

summarized the mean MBI burnout levels for six occupational fields and various

professions. Table 3.1 presents the mean burnout levels for a range of human service

professions (i.e., teachers, social workers, nurses, physicians, psychologists and

counsellors).
 59

Table 3.1

Burnout Levels Among the Human Service Professions Based on Studies Published

Between 1979 and 1998.

Emotional Exhaustion Depersonalisation Personal Accomplishment

Field/Profession Samples n Mean SD Samples n Mean SD Samples n Mean SD
Teaching 6 5481 28.15 11.99 6 5481 8.68 6.46 7 6191 36.35 7.41
Social Services
Social
Worker 6 628 20.82 10.17 6 628 6.94 5.60 6 628 35.15 7.80
Medicine
Nurse 11 1542 23.80 11.80 11 1542 7.13 6.25 11 1541 34.47 8.15
Physician 3 479 24.03 10.77 3 479 10.59 6.46 3 475 39.36 5.93
Mental Health
Psychologist 6 1382 19.75 9.77 6 1382 6.14 4.45 6 1382 40.94 6.07
Counsellor 4 422 20.52 8.97 4 422 6.64 4.26 4 422 35.35 6.75
Total 36 36 37
Note. Adapted from The Burnout Companion to Study and Practice: A Critical Analysis (p.62), by W.

Schaufeli and D. Enzmann, 1998, London: Taylor & Francis.

Table 3.1 indicates nurses generally report moderate levels of emotional

exhaustion and depersonalisation but moderately low levels of personal

accomplishment (Schaufeli, 1999). In comparison with physicians, nurses experience

slightly less emotional exhaustion, much less depersonalisation, but much lower

levels of personal accomplishment. In comparison with teachers, nurses report

significantly lower levels of emotional exhaustion, lower levels of depersonalization

but lower levels of personal accomplishment. When compared with psychologists and

counsellors, nurses report higher levels of emotional exhaustion and depersonalisation

and much lower levels of personal accomplishment. Overall, it could be concluded

that nurses report comparable levels of emotional exhaustion and depersonalisation to

other human service professionals, but lower levels of personal accomplishment.

The present research contributes to our knowledge of whether nurses are

indeed at a higher risk of experiencing burnout than other human service professions

by generating mean MBI scores for a sample of Australian nurses and comparing
 60

them to mean MBI scores for a range of human service professions as detailed in the

MBI administration manual.

Antecedents of Burnout

Burnout in nurses has been primarily examined from the point of view of its

potential antecedents (e.g., Cordes & Dougherty, 1993; Eastburg, Williamson,

Gorsuch, & Ridley, 1994; Firth, McKeown, McIntee, & Britton, 1987; Kandolin,

1993; Leiter & Maslach, 1988; Maslach, 1982; Pick & Leiter, 1991; Robinson et al.,

1991). Burnout is believed to originate from nursing work itself as well as from

characteristics of the worker and the work environment (Koivula, Paunonen, &

Laippala, 2000). Among the factors which are considered to be the main antecedents

of burnout, those referring to the work environment are particularly significant

(Schaufeli & Janczur, 1994).

However, findings regarding the precise work-related determinants of burnout

in nursing have been somewhat inconsistent. This may be because empirical studies

have conceptualised burnout in different ways. Whilst some studies have investigated

burnout as a unidimensional concept and have analysed burnout using aggregate

burnout scores (e.g., Duquette et al., 1995), some studies have only examined what is

deemed to be the core component of burnout, that is emotional exhaustion (e.g.,

Janssen, De Jonge, et al., 1999). Other studies have examined burnout as a two

component syndrome - emotional exhaustion, and depersonalisation (e.g., Demerouti

et al., 2000), and have excluded personal accomplishment altogether. Furthermore,

studies have been conducted on different nursing groups working in a variety of

environmental contexts. Understandably, researchers are still unclear how various

work-related factors exert their influence across the three burnout dimensions. One of

the purposes of the present research is to enhance our understanding of the

 61

antecedents of nursing burnout by examining the predictive relationships between

work stressors and the three burnout components. Extensive studies have been

conducted in the nursing literature to examine work related stressors and burnout. In

particular, stressors specific to the nursing industry and generic stressors associated

with the role of nursing have been examined. There is some evidence to suggest that

stressors that are more job-specific have the greatest impact on individual strains

because they are most salient to employees in a particular job (Beehr, Jex, Stacy, &

Murray, 2000; Kilfedder et al., 2001). However, very few nursing studies (e.g., Gray-

Toft & Anderson, 1981b) have investigated both job-specific stressors and role

stressors in the same study. Thus, in the section below, job-specific nursing stressors

and role stressors will be reviewed separately.

The Relationship Between Job-Specific Stressors and Burnout

The nursing literature consistently reveals that certain job-specific stressors –

such as exposure to death and dying and client problems that do not have solutions,

coupled with work overload, have an adverse psychological impact on nurses. It is

difficult, however, to summarise the results of these burnout studies because they use

different measures of burnout and examine a wide array of job-specific stressors.

A multitude of job-specific stressors have been found to be associated with

burnout among nurses. For example, nurses’ burnout has been found to be related to

death and dying (Payne, 2001), conflict with staff (Payne, 2001), high workload

(Demerouti et al., 2000; Duquette et al., 1994; Janssen, Schaufeli, et al., 1999;

Tummers, Janssen, Landerweed, & Houkes, 2001), poor social support (Bourbonnnais

et al., 1998; Corrigan et al., 1994; Duquette et al., 1994; Turnipseed & Turnipseed,

1997), lack of control (Ellis & Miller, 1994), lack of participation in decision-making
 62

(Ellis & Miller, 1994), and a perceived imbalance between investments and outcomes

in relationships with patients (Schaufeli & Janczur, 1994; Van Yperen et al., 1992).

Duquette et al. (1995) examined the psychosocial determinants of burnout

among 1,545 geriatric nurses. Work stressors were measured by the NSS (Gray-Toft

& Anderson, 1981a) and burnout was assessed by the Staff Burnout Scale for Health

Professionals (Jones, 1980). The results of a hierarchical multiple regression analysis

revealed that work stressors explained an important part of the variance (21%) in

burnout. The most significant predictors of burnout included conflict with physicians

and a heavy workload.

Demerouti et al. (2000) investigated the influence of ten job demands and

eleven job resources (e.g., social support) on burnout using a sample of 109 hospital

nurses. Burnout was measured using the Oldenburg Burnout Inventory (Ebbinghaus,

1996). The results of a series of structural equation modeling analyses suggested that

demanding patients, high workload, time pressures, unfavourable environmental

conditions and problems with shift work schedule was significantly related to

emotional exhaustion. In contrast, an attitude of disengagement was primarily

associated with a work environment lacking resources (performance feedback, job

control, task variety, supervisor support, rewards and participation in decision

making). These findings are consistent with Leiter (1993) and Lee and Ashforth’s

(1996) findings in which emotional exhaustion was most strongly correlated to work

demands and depersonalisation and personal accomplishment were most strongly

related to resources.

Kilfedder et al. (2001) examined the relationship between job-specific nursing

stressors (as measured by Gray-Toft and Anderson’s NSS) and generic role stressors

(i.e., role conflict and role ambiguity) using a sample of 510 psychiatric nurses.
 63

Hierarchical regression analysis revealed that higher levels of emotional exhaustion

were associated with job-specific nursing stressors. Neither job-specific nor generic

work stressors predicted depersonalisation and personal accomplishment. Lack of

social support (assessed using House and Wells’ (1978) Social Support Measure), was

found to be a significant predictor of emotional exhaustion and depersonalisation.

Lower levels of personal accomplishment were associated with less control over job-

related events and less use of coping strategies.

Payne (2001) examined nursing stressors as determinants of burnout in 89

female hospice nurses. Work stressors were measured by the NSS and burnout was

assessed by the MBI. Multiple regression analyses found that death and dying and

conflict with staff contributed to emotional exhaustion. Conflict with staff and

inadequate preparation contributed to depersonalisation. Inadequate preparation also

contributed to lower levels of personal accomplishment.

Some of the more consistent findings, however, suggest that workload and

lack of support are the strongest predictors of burnout. Van Servallen and Leake

(1993) found that burnout is not associated with particular patient groups and their

treatment, but especially with haste at work. Miller, Reesor, McCarrey, and Leiken

(1995) also found that long work hours related to burnout. In a study of 156 Dutch

nurses employed at a general hospital, Janssen, De Jonge, et al. (1999), reported that

emotional exhaustion was primarily predicted by a lack of support from colleagues,

and by demanding aspects of work, such as working under time pressure and

strenuous work. Lee and Ashforth’s (1996) meta-analysis of the effects of various

work characteristics on burnout also identified a high perceived workload as the

single most important source of emotional exhaustion. Furthermore, Janssen,

Schaufeli, et al. (1999), using a sample of 156 Dutch hospital nurses, examined the
 64

relationships between a range of job demands and work resources in relation to the

three burnout components. They found that emotional exhaustion is primarily and

strongly correlated (r = .45, p ≤ .01) with work overload. Depersonalisation was

associated with lack of social support (r = -.22, p ≤ .05). Personal accomplishment

was not associated with work-related factors. Instead, personal accomplishment was

primarily associated with the dispositional attribute - self-esteem (r = -.26, p ≤ .01).

Based on the above findings, it could be argued that different stressors are related to

different aspects of burnout.

The Relationship Between Role Stressors and Burnout

According to Baba et al. (1999), as role pressures mount, nurses experience

stress and tend to burn out. Role conflict and role ambiguity are frequently studied as

sources of stress in the nursing literature (e.g., Gray-Toft & Anderson, 1985; Jones,

1987; Peiro & Zurriaga, 1985) however, few studies (e.g., Cash, 1991; Gil-Monte et

al., 1995) have investigated the relationship between role stressors and burnout in the

nursing profession. Despite this, numerous investigations have demonstrated a

consistent link between role stress and burnout among other human service

professionals. In particular, role conflict and role ambiguity have been repeatedly

examined in relation to burnout. For example, Holloway and Wallinga (1990) in a

study of child care specialists found that although role ambiguity and role conflict

were both significantly correlated with burnout, role ambiguity was a stronger

predictor of burnout. Similarly, Fimian (1984) found that while both role ambiguity

and role conflict were correlated with burnout, role conflict was not a significant

predictor of burnout among special education teachers. The major limitation of these

empirical studies, however, is their use of a single score to represent burnout, rather
 65

than examining the differential effects of role stressors on the three burnout subscales

of emotional exhaustion, depersonalisation and reduced personal accomplishment.

There is evidence to suggest that role conflict and role ambiguity contribute

independently to burnout dimensions. When burnout relationships are reported on the

basis of the three subscales, however, the relationship between the three role stressors

and burnout is still not clear. For instance, Boyd and Pasley (1989) designed a study

to determine whether role ambiguity or role conflict best explained the variance in the

level of burnout experienced by 85 full-time, childcare professionals. The results of

the regression analyses revealed that role ambiguity was the strongest predictor of

emotional exhaustion (R2 = 18%, β = .43, p = .000) and depersonalisation (R2 = 14%,

β = .37, p = .001). However, role conflict was the strongest predictor of personal

accomplishment (R2 = 9%, β = -.30, p = .005). Boyd and Pasley suggested that in their

particular sample, the ability of role ambiguity to explain variance in emotional

exhaustion and depersonalisation may be explained by factors such as unclear

organisational structures and difficulty in defining success. Because the needs of

individual children and groups of children differ from day to day, it may become

difficult for a caregiver to measure the efficacy of their approach. These results

suggest that working in an ambiguous environment may result in increased role

ambiguity on the part of the childcare professionals. According to Boyd and Pasley,

the ability of role conflict to explain variance in personal accomplishment may stem

from holding overly high expectations which may not be met in the work setting, or

having to choose between providing quality care and performing other responsibilities

such as food preparation. However, since so little of the variance in personal

accomplishment was accounted for by role conflict, it is likely that other more salient
 66

factors are responsible for child care professionals’ sense of personal

accomplishment.

Boyd and Pasley’s results can be contrasted with the findings of Crane and

Iwanicki (1986). In a sample of 443 special education teachers, Crane and Iwanicki

found that both role ambiguity and role conflict were significant predictors of

emotional exhaustion and depersonalisation. Although role ambiguity was a

significant predictor of emotional exhaustion and depersonalisation, it only accounted

for 1% and 2% of the variance in the burnout components, respectively. In contrast,

role conflict accounted for 14% and 19% of emotional exhaustion and

depersonalisation, respectively. It is therefore suggested that the significance of role

ambiguity may be attributed to the large sample size used in this study. Role

ambiguity, however, was the only role stressor that accounted for variance in personal

accomplishment (R2 = 6%). Similar findings were reported by Schwab and Iwanicki

(1982) in a sample of elementary and high school teachers. Crane and Iwanicki

concluded that working in a job that is loosely defined and in an environment that is

open to role conflicts (such as trying to address the needs of a diverse group of special

education students while complying with limited teaching resources provided by the

education institution) is certainly stressful and likely to lead to burnout.

More recently, in a study of 102 nursing professionals, Gil-Monte et al. (1995)

found that role conflict and role ambiguity contributed independently to different

aspects of burnout. They found a significant positive relationship between role

conflict and emotional exhaustion (β = .46, p < .001) and depersonalisation (β = .43, p

< .001). In addition, role ambiguity was a significant negative predictor of personal

accomplishment (β = -.27, p < .01). The relationship between emotional exhaustion

and role conflict has been found to be significant in other studies (Boyd & Pasley,
 67

1989; Capel, Sisley, & Desertain, 1987; Crane & Iwanicki, 1986; Jackson, Schwab, &

Schuler, 1986). Gil-Monte et al. suggested that a possible explanation for the

significant positive relationship between emotional exhaustion and role conflict is that

nursing work requires warmth and sympathy, together with objectivity and

assertiveness, which could result in intra-role conflict. The consequent effort for

dealing with this conflict could be related to the depletion of one’s emotional

resources (Gil-Monte et al.). Furthermore, the significant positive relationship

between depersonalisation and role conflict has been replicated in other studies (Boyd

& Pasley, 1989; Kottkamp & Mansfield, 1985; Linquist & Whitehead, 1986). Gil-

Monte et al. proposed that depersonalisation is a defense mechanism developed by

nurses to cope with conflicting demands; the result of which is to dehumanise care to

patients. Finally, other researchers (Cash, 1991; Holgate & Clegg, 1991) have also

found a significant negative relationship between personal accomplishment and role

ambiguity. Cash (1991) suggested that although the focus of nursing is on care and

not cure, many nurses have come to equate health care with cure. When patients are

not cured nurses may see themselves as failures and experience lower levels of

personal accomplishment. Role ambiguity seems a rather salient stressor within

nursing (Buunk, Van Yperen, Taylor & Collins, 1991; Molleman, Pruyn, & Van

Knippenberg, 1986). Although the role of a nurse may seem clear, there may be

considerable uncertainty in how to carry out this role (McGrath et al., 1989). For

example nurses may wonder if they are too involved with patients or not involved

enough, may feel uncertain about how to deal with various problems of patients

(including appeals for help and expressions of anxiety), and may experience

insecurities about whether they are doing their job correctly (Buunk & Schaufeli,

1993).
 68

It is evident from the literature reviewed that both role ambiguity and role

conflict can function as intervening variables between environmental conditions in the

workplace and a variety of behavioural and affective responses (Cash, 1991; Pearce,

1981) and hence act as predictors of job-related outcomes such as burnout in

individuals. Boyd and Pasley (1989) proposed that the conflicting empirical findings

between studies suggest that role stressors vary in importance and degree contingent

upon the specific occupation held and the job-related outcome measured. However, as

Jackson and Schuler (1985) noted, more attention is needed to explain the

mechanisms by which role stressors are linked to negative consequences.

While the effect of role overload on burnout is not as well documented as role

ambiguity and role conflict, it has been found to be a significant predictor of stress-

related outcomes such as job dissatisfaction, depression, and burnout (Kahn &

Byosiere, 1992). Originally, role overload was “treated as a variant of role conflict, in

which conflict was experienced as a necessity to compromise either quantity, time

schedule, or quality” (Kahn & Byosiere, 1992, p. 601). However, in recent years, role

overload has been treated as a construct in its own right and earlier arguments of

interdependency between role conflict and role overload have been opposed (Kahn &

Byosiere, 1992). Although literature examining the relationship between role overload

and burnout is limited, the following empirical studies have examined the dimension

of role overload independently of other role stressors. These studies provide empirical

evidence that role overload is a significant predictor of strains, independent of other

role stressors, and therefore should be recognised as a separate construct from role

conflict in the stress-burnout literature.

Support for the independence of the three role stressors is found in a study by

Iwata and Suzuki (1997) involving 256 Japanese bank workers. The results of this
 69

research demonstrated that role overload (r = .42, p < .01) rather than role conflict (r

= .25, p < .01) or role ambiguity (r = .16, p < .05), had the largest association with

overall mental health. Iwata and Suzuki (1997) proposed that when workers have to

meet strict deadlines as well as completing other regular daily routines (e.g.,

administrative work), role overload may play a more significant part in predicting

overall negative mental health than other role stressors such as role conflict and role

ambiguity.

In a study of active trade union members, Nandram and Klandermans (1993)

examined the extent to which role stressors (i.e., role conflict, role ambiguity, and role

overload) correlated with the core component of burnout, emotional exhaustion.

These researchers reported that emotional exhaustion was most strongly associated

with role conflict (r = .33, p < .01) and role overload (r = .20, p < .01). Only role

ambiguity failed to significantly correlate with emotional exhaustion (r = .10, ns).

Similarly, Kottkamp and Travlos (1986) found in a study of high school principals

that emotional exhaustion was most strongly associated with role conflict (β = .22, p

< .01) and role overload (β = .33, p < .01). A significant limitation of these studies

however, is that only one component of burnout, that of emotional exhaustion, was

examined.

In contrast, Koniarek and Dudek (1996) using a sample of 1,023 hospital

nurses, investigated the relationship between role stress and the three burnout

components. They reported that role overload, role conflict and role ambiguity were

significantly and positively correlated with emotional exhaustion, with correlation

coefficients of .47, .41, and .35 (p < .01) respectively. In addition, role overload (r =

.25, p < .01), role conflict (r = .29, p < .01), and role ambiguity (r = .25, p < .01)

positively correlated with depersonalisation, although the correlation coefficients

 70

were less strong. Finally, role conflict (r = -.13, p < .01) and role ambiguity (r = -.14,

p < .01) were significantly but weakly associated with personal accomplishment.

Baba et al. (1999) using a sample 119 hospital nurses, also reported that role conflict

and role overload predicted burnout.

A study by Cordes et al. (1997) involving 354 human resource professionals

(managers and professionals), found role overload to be significantly associated with

emotional exhaustion (r = .52, p < .001) and depersonalisation (r = .25, p < .001).

Although role overload was negatively associated with personal accomplishment, it

was not significant (r = -.10, p > .05). Cordes et al. suggested that individuals with

insufficient time and human resources to accomplish the job at work, may expend an

excessive amount of emotional energy and use their time inefficiently to maintain

performance standards, thus leading to emotional exhaustion and depersonalisation.

In summary, based on the literature reviewed, it could be argued that where

working conditions are characterised by high role overload, employees are likely to

experience higher levels of burnout. More specifically, conditions of high role

overload are associated with high levels of emotional exhaustion and

depersonalisation, and low levels of personal accomplishment. Excessive workloads

are frequently used to describe employment conditions in the human service

industries (Kahn & Byosiere, 1992) where high levels of burnout have been recorded

(Maslach et al., 1996). These findings provide evidence that role stressors should be

taken into consideration when examining the stress-burnout relationship among

nurses. Recent literature confirms that role stressors are significantly associated to

burnout. Based on the findings above, the following hypotheses were generated:

Hypothesis 4: Both job-specific stressors and role stressors will be related to

burnout in nurses.
 71

Hypothesis 5: Job-specific stressors will be significantly and positively related

to emotional exhaustion and depersonalisation and negatively related to personal

accomplishment.

Hypothesis 6: Role stressors will be significantly and positively related to

emotional exhaustion and depersonalisation and significantly and negatively related to

personal accomplishment.

The Relationship Between Sociodemographic Factors and Burnout

Research demonstrates that sociodemographic variables have little influence on

burnout among nurses (Dallender, Nolan, Soares, Thomsen, & Arnetz, 1999). In most

studies, no relationship has been found between burnout and marital status (Ackerley,

Burnell, Holder, & Kurdek, 1988; Dallender et al., 1999; Kandolin, 1993; LeCroy &

Rank, 1987; Raquepaw & Miller, 1989), ethnic background (LeCroy & Rank, 1987;

Naisberg-Fennig, Keinan, & Elizur, 1991; Raquepaw & Miller, 1989) or level of

education (Leiter & Harvie, 1996). Duquette et al.’s (1994) meta-analytic review of the

nursing burnout literature suggested that only age and length of nursing experience

appear to contribute to burnout among nurses.

Robinson et al. (1991) and Van Servallen and Leake (1993) found that younger,

less experienced nurses reported more emotional exhaustion and depersonalisation, but

that in older staff members, personal accomplishment was low. This finding suggests

that although practical experience may guard against emotional exhaustion and

depersonalisation, it does not enhance one’s sense of personal accomplishment.

Bennett, Michie, and Kippax (1991) also reported that the age of nurses influenced

levels of burnout, with older nurses experiencing less burnout than younger nurses.

Kilfedder et al. (2001) found higher levels of depersonalisation were associated with

younger, more recently qualified nurses. Burnout has been shown to be more common
 72

among younger employees perhaps because of the initial shock of the job in reality, a

lack of adaptation to or insecurity in working life, or a perception of more role

ambiguity. Younger nurses seem to experience the discrepancy between their nursing

education and the reality of working as difficult, which reduces their enthusiasm in the

early stages of their career (Ketola & Kevätsalo, 1994). Those that remain in nursing

are perhaps those that did not experience burnout early on in their careers. Koivula,

Pauonen, and Laippala (2000) suggest that burnout in older nursing staff may be

explained by the fact that the last stages of their career lack the possibility to learn new

things and develop themselves, since possibilities for professional education have been

exhausted. There is also some evidence to suggest that length of nursing experience

correlates negatively with burnout. The length of nursing experience may lead to an

increasing ability to cope with the job, over and above the increase in coping

attributable to age (Cherniss, 1980). There is evidence to suggest that full-time nursing

staff report greater levels of emotional exhaustion (Burke & Greenglass, 2001,

Duquette et al., 1995). Furthermore, higher level staff (e.g., nurse managers, nurse

practitioners, nurse educators) report higher levels of personal accomplishment from

their jobs (Butterworth et al., 1999).

Gender differences in burnout are typically not reported or not found across

nursing studies. This may be because the sample size of male nurses is generally

small and therefore significant differences may not be detected from the data.

However, in their meta-analytic review of stress and stress interventions of mental

health nurses, Edwards and Burnard (2003) found males scored higher than females

on all three burnout variables. Similarly, Kilfedder et al. (2001) reported that male

nurses reported more depersonalisation than female nurses. In the wider burnout

literature, males have been consistently found to report higher levels of

 73

depersonalisation than females (Ogus, Greenglass, & Burke, 1990). Zani and

Pietrantoni (2001) in a sample of 314 health professionals also found that men scored

higher than women on depersonalisation and reduced personal accomplishment.

Greenglass, Burke, and Ondrack (1990) propose that women may be better at dealing

with problems emanating from interpersonal work situations because of norms

associated with the feminine gender role. The role prescribed for women emphasises

interdependence, the ability to relate meaningfully to others in interpersonal

relationships (Greenglass, 1982) as well as caring and nurturing others. Thus, women

should be less likely to respond to people and their problems in a detached manner

(Maslach & Jackson, 1985).

Overall, sociodemographic factors typically explain a very small part of the

explained variance in burnout. For instance, Duquette et al. (1995) found that

sociodemographic factors explained only 2% of the explained variance in burnout. It

is therefore plausible to conclude that sociodemographic factors are not the main

determinants in producing burnout. In the present thesis, sociodemographic factors

will be taken into consideration when examining the antecedents of burnout in Study

3.

Based on the previous discussion, the following hypothesis was formulated.

Hypothesis 7: Younger nurses will report higher levels of burnout than older

nurses.

Hypothesis 8: Less experienced nurses will report higher levels of burnout

than more experienced nurses.

 74

Summary

In summary, nurses have been identified as a high risk occupation for

developing burnout due to their close involvement with emotionally demanding

patients. Empirical studies have found, however, that overall nurses experience

moderate levels of burnout and the percentage of nurses reporting high burnout is

relatively small. Furthermore, in relation to other human service professionals, nurses

report comparable levels of burnout.

A review of the literature indicates that work stressors (both job-specific

nursing stressors and role stressors) are stronger predictors of burnout in nurses than

sociodemographic factors. Only age and length of nursing experience are strongly

associated with burnout, with younger, less experienced nurses reporting higher levels

of burnout than older more experienced staff. Many authors, however, have taken the

simplistic view that excess pressure will result in burnout without accounting for the

fact, that when exposed to the same conditions, some individuals burn out while

others do not. This suggests that there are individual differences in coping that may

reduce or prevent the potentially negative impact of nursing work (Payne, 2001).

Among health care professionals, the coping resource, social support, has been

frequently studied (De Jonge, Janssen, & Van Breukelen, 1996; De Jonge &

Schaufeli, 1998; Janssen, De Jonge, et al., 1999; Janssen, Schaufeli, et al., 1999). This

is not surprising because particularly in performing ‘people work’, which often

implies intensive emotional experiences, social support from others may help an

employee to cope effectively with these experiences (Peeters & LeBlanc, 2001). In

the following chapter, pertinent social support literature is reviewed.

 75

CHAPTER 4

Review of the Social Support Literature

Background

At the workplace, the provision and receipt of social support has been highly

recommended as a means of protecting employees from the deleterious effects of

exposure to unavoidable or unmodifiable worksite stressors (House, 1981). Social

support has been conceptualised as a resource that people call upon when coping with

stress (Holahan, Moos, & Bonin, 1997). Social support, like other coping resources,

can modify the impact of chronically stressful conditions on health status (Sarason,

Sarason, & Pierce, 1992). Although there is a rapidly growing body of research

evidence regarding the relationships between social support, stress and health, there

are still many conceptual issues that remain unclear.

Nurse researchers have contributed to the conceptualisation, measurement and

empirical study of social support (Stewart, 1993), enhancing the social support

literature primarily through research with clinical populations encountered in nursing

practice (Norbeck, 1988). However, only a small number of studies (e.g., Duquette et

al., 1995; Eastburg et al., 1994; Ellis & Miller, 1994; Kilfedder et al., 2001) have

examined the influence of occupational support on hospital nurses’ burnout.

Therefore, in reviewing the social support construct and its influence on the stress-

burnout relationship, the wider social support literature has been consulted. Chapter 4

reviews the literature on social support and examines the way in which the social

support construct has been defined and measured. In addition, the role social support

plays in the stressor-strain relationship is investigated.

 76

Definition of Social Support

Social support has been broadly defined as “social transactions that are

perceived by the recipient or intended by the provider to facilitate coping in everyday

life, especially in response to stressful situations” (Pierce, Sarason, & Sarason, 1990,

p.173). More simply, social support is “the assistance one receives through his or her

interpersonal relationships” (Quick et al., 1997, p.196). Despite research advances in

this field, there is little consensus on how social support should be defined. Social

support, like stress, represents a meta-concept (Beehr & McGrath, 1992; Vaux,

Phillips, Holly, Thompson, Williams, & Steward, 1986). It has been suggested that

the definitions of social support are so vague or broad that the concept is in danger of

losing its distinctiveness (Barrera, 1986; Cohen & McKay, 1984). This has resulted in

the term encompassing an increasing amount of psychosocial phenomena. Gottlieb

(1983) noted the proliferation of social support concepts by observing that “with each

new study a new definition of social support surfaces” (p.50). Turner (1992)

described this notional explosion as a kind of ‘conceptual imperialism’ with social

support coming to refer to “an ever widening domain of content” (p.217). It is

acknowledged by most contemporary researchers that social support is a

multidimensional construct. Veiel and Baumann (1992) argue that in order to prevent

permanent fragmentation of the field, a widely shared notion of what is meant by

‘social support’ is necessary. Upon closer examination, it becomes apparent that

social support is defined and measured from two broad perspectives - structural

support and functional support.

Structural Support

Structural definitions of social support focus on an individual’s social

embeddedness. From this perspective, social support refers to “the connections that
 77

individuals have to significant others in their social environment” (Barrera, 1986,

p.415). More specifically, it refers to the organisation of people’s ties to one another –

in particular, to the number of relationships or social roles a person has, and to the

frequency of his or her contact with various network members (Thoits, 1995).

From this perspective, the composition and organisation of an individual’s

social network is of significant interest and it does not involve any evaluation on the

part of the individual (McNally & Newman, 1999) and is therefore not influenced by

subjective biases. It has its roots in social network analysis and may involve a number

of related concepts. Tolsdorf (1976) identified the main concepts to include the size of

the network, operationalised in terms of the number of people listed by the respondent

as being network members (i.e., according to the criteria of knowing each other by

name and having some contact at least once a year). Another concept is that of

adjacency density, that is, an approximation of the extent to which the respondent’s

network is interconnected based on the number of dyadic relationships that exist

within it.

Gottlieb (1983) noted that community psychologists and epidemiologists have

traditionally focused on structural features of social networks that could be changed to

enhance the health and social functioning of entire populations (Wethington &

Kessler, 1986). More specifically, these researchers have been interested in studying

the health promoting effects of (1) interactions among friends, neighbours, and

relatives, (2) participation in church and social groups, and (3) employment outside

the home. Although structural measures of social support have been found to be

predictive of mortality in large scale epidemiological studies such as those of

Berkman and Syme (1979) and House, Robbins, and Metzner (1982), the use of such

measures pose problems in smaller studies. Sarason, Sarason, and Pierce (1990)
 78

warned that these scales are not very reliable due to their small number of items and

lack of internal consistency. Furthermore, the mere existence of relationships does not

mean they are supportive. Antonucci (1990) suggested that “support structures and

social networks describe those people who are available to (but do not necessarily)

provide support” (p.208). Wellman (1981) correctly pointed out that when social

network analysis identifies important social relationships, it is erroneous to assume

that all such linkages involve the provision of social support. Even when social

interactions and potentially supportive behaviours occur, social support is not simply

an objective property of these social behaviours. It is not possible to determine

whether a specific social interaction constitutes social support without reference to the

cognitive appraisals of that interaction (Sarason et al., 1992). In more recent years,

researchers have relied less on objective structural approaches and more on subjective

functional approaches when conceptualising social support. Functional approaches are

not just concerned with the existence of relationships, but also to what purpose these

relationships serve an individual.

Functional Support

Functional support refers to definitions of social support requiring an

indication that the socially supportive person or people serve a function for the focal

person. For instance, Cobb (1976) defined social support as the individual perception

that one is cared for and loved, is esteemed and valued, and belongs to a network of

communication and mutual obligation. Gottlieb (1983) defined social support as

“verbal and/or non-verbal information or advice, tangible aid, or action that is

proffered by social intimates or inferred by their presence and has emotional and

behavioural effects on the recipient” (pp.28-29). From this perspective, social support

is increasingly seen as a multidimensional construct, encompassing diverse and

 79

distinct components. Examples of such functions include the provision of esteem,

information or companionship, or of being instrumental to the focal person’s

accomplishment of tasks (Cohen & Wills, 1985). Virtually all known studies of social

support in relation to work-related stress have used functional measures of social

support. In contrast to structural measures that reflect the individual’s embeddedness

in an interpersonal network, regardless of whether these relationships are likely to

prove helpful in times of need, functional measures assess either the perceived

availability of particular support functions if needed, or the actual support functions

received within a recent time period.

Perceived social support views support as a cognitive appraisal of one’s

connections to others. From this perspective, it is recognised that not all linkages

between individuals and the environment result in social support. Even if the potential

exists for a particular relationship to generate expressions of social support, it is not

likely to do so unless it is perceived to be available and adequate to meet the need. In

some cases, efforts to provide support are inappropriate, poorly timed, or against the

wishes of the person being helped (Rook, 1984; Rook & Dooley, 1985). Measures of

perceived support focus on the individual’s cognitive appraisal of his or her

environment and the level of confidence he or she has that when support is needed it

will be available, sufficient to meet the need, and offered in a way that is perceived to

be beneficial (Tracy, 1990).

Received social support reflects the amount of support individuals receive or

report to receive during a specified period. Received support, otherwise known as

enacted support, refers to the specific behaviours or actions performed by others as

they exhibit expressions of support and assistance. Support behaviours can include

activities such as listening, expressing concern, lending money, helping with a task,
 80

offering suggestions, giving advice, and showing affection. Measures of enacted

support emphasise what people actually do when they provide support. In a sense,

received support is a behavioural assessment of social support (Tardy, 1985). Most

measures of received support are self-report measures that depend on recall of past

experiences rather than actual observations of supportive behaviours. It could be

argued that self-report scales of enacted support are actually assessing perceived-

received support because these measures rely on retrospective evaluations (Barrera,

1986). A number of studies (e.g., Carveth & Gottlieb, 1979; Lefcourt, Martin, &

Saleh, 1984; Sandler & Barerra, 1984) have focused on received support to assess the

responsiveness of one’s environment to requests for assistance and the behavioural

response that request receives.

Types of Support

Several different classification schemes have been developed for

distinguishing between different functions or types of support. It is proposed that

there are different types of supportive functions provided through social relationships,

and it is posited that these functions may be differentially useful for various types of

problems or stressors (Cohen & McKay, 1984; Cohen, Underwood, & Gottlieb, 2000;

Cutrona & Russell, 1990; Sandler, Miller, Short, & Wolchick, 1989). Several

dimensions of functional support have been delineated (Argyle, 1992; Cobb, 1979;

House, 1981; Kahn, 1979; Schaefer, Coyne, & Lazarus, 1981; Cohen, Mermelstein,

Karmarck, & Hoberman, 1985; Weiss, 1974). These include emotional or esteem

support, the availability of one or more persons who can listen sympathetically when

an individual is having problems and can provide indications of caring and acceptance

(Cohen et al., 2000). Instrumental support involves practical help when necessary,

such as assisting with chores and providing tangible aid such as money and other
 81

physical objects (tools and equipment). Informational support is defined as providing

knowledge that is useful for solving problems such as providing information, advice

and guidance. Companionship support involves the availability of persons with whom

one can participate in social and leisure activities such as trips, parties, or recreational

activities. A dimension variously termed appraisal, feedback, validation, or social

comparison support is based on the concept that social relationships can provide

information about the appropriateness or normativeness of behaviour.

Within the occupational stress literature, the two functions most often

examined are emotional and instrumental support (House, Kahn, McLeod &

Williams, 1985). This may be because theoretically, these dimensions are most easily

distinguished. Of the two, emotional support is the one that most people think of when

the term social support is mentioned. Many early thoughts about the subject were that

emotional support would be the one that would help alleviate the effects of

occupational stress better, but this has not been confirmed by more recent empirical

research (Himle, Jayaratne, & Thyness, 1991). While emotional support has been

found to be an effective coping resource in handling the emotions associated with

stressful problems, instrumental support appears to be more effective in handling the

source of the stress (Nätterlund & Ahlström, 1999). This may be because instrumental

support is able to directly help an employee by reducing their workload, or by solving

a work-related problem.

A number of measures have been created to assess the dimensions of social

support. The theoretical conceptualisation of social support proposed by Weiss (1974)

has inspired the greatest number of measurement efforts (Brandt & Weinert, 1981;

Cutrona & Russell, 1987; Henderson, Duncan-Jones, Byrne, & Scott, 1980; Russell &

Cutrona, 1984; Turner, Frankel, & Levin, 1983). Other scales have been based on the
 82

dimensions of support described by Caplan (1974) and Cobb (1976) (Caplan et al.,

1975; Turner et al., 1983) and Kahn (1979) (Kahn & Antonucci, 1980; Kahn,

Wethington, & Ingersoll-Dayton, 1987; Norbeck, Lindsey, & Carrieri, 1981, 1983).

While several types of support have been defined theoretically, in empirical studies

(Brandt & Weinert, 1981; Cohen et al., 1985; Norbeck et al., 1981; Schaefer et al.,

1981; Tetzloff & Barrera, 1987; Turner et al., 1983) it is common to find significant

correlations among subscales. This may occur because persons who seek support need

and obtain several functions and because persons who provide support are able to do

so in several ways. The magnitude of the correlations can be influenced to some

extent by the research - for example, one might anticipate higher intercorrelations

among functions when studying support from primary relationships than when

studying support in medical settings - however, it may also a property of the

measurement instruments (Cohen et al., 2000). The high intercorrelations between

support components pose significant problems for researchers. Although less common

in current research, some researchers combine the types of support, whilst keeping the

sources of support separate (e.g., Caplan et al., 1975; Kaufmann & Beehr, 1986, 1989;

Mattimore, 1990). Other researchers (e.g., Brookings & Bolton, 1988; Cohen & Wills,

1985; Cutrona & Russell, 1987; 1990; Vaux, Riedel, & Stewart, 1987) argue,

however, that the effects of social support on strains will only be discovered if

multidimensional measures of social support are used.

An issue that has received little explicit attention in the social support

literature is that of sources of support. In the section below, the main sources of social

support measured in the stress-strain literature are discussed.

 83

Sources of Support

Empirical studies in the occupational stress literature recognise that work

support (e.g., supervisors and coworkers) and non-work support (e.g., family and

friends) may be differentially effective in reducing work-related stressors and strains.

Most occupational stress studies consider the supervisor and/or colleagues to be the

two major sources of support for employees. It is proposed that in dealing with

stressors at the workplace, organisational sources will provide more support than

family and friends outside of the workplace (Ellis & Miller, 1994). As Ray (1987)

noted, “the overt and subtle stresses in the workplace are known to members and are

unclear to non-members. As a result, supervisors and coworkers are probably in a

better position to provide support than non-organisational members” (p.174). The

argument that work support is more effective than non-work support in protecting

employees from strains has been supported in the literature (Beehr, 1985; House,

1981; LaRocco, House, & French, 1980).

Although it is recognised that family and friends (i.e., non-work support) may

be an important source of support in assisting individuals to cope with stress, the

present research confines its study to examining the influence of work support on the

stressor-strain relationship. Most work support measures assess the degree and/or

satisfaction of support provided to employees by their supervisor and their colleagues.

To date, however, there has been no research investigating nurses’ main sources of

support. Hospital nurses work closely with a variety of health professionals (e.g.,

nurse administrators, physicians, physiotherapists, social workers, counsellors) and

non-professionals (e.g., wards persons) yet it is not known whether the staff fulfilling

these occupational roles are perceived to be a major source of support to nurses.

 84

In an attempt to refine the definition of social support in the nursing literature,

the concept of social support will be examined in more detail in Study 1. One of the

principal research questions addressed in Study 1 is:

To what extent is the way in which nurses’ perceive work support consistent

with how social support is conceptualised and operationalised in recent empirical

literature?

Prior to measuring social support among nurses in Study 2, the researcher will

collect qualitative data in Study 1 to examine nurses’ perceptions of the types and

sources of support they receive at work. This will enable the researcher to determine

whether nurses’ perceptions of social support are consistent with the contemporary

view that support from different sources serve various functions and therefore

multidimensional measures of social support should be used in empirical studies. The

major advantage of using a functional measure of support is that the differential

effects of socially supportive behaviours on the stress-strain relationship can be

evaluated.

Support from Supervisors and Coworkers

There is some research to suggest that supervisor support is the most potent in

reducing job-related stressors and strains (Beehr et al., 1990; Ganster et al., 1986;

Jackson et al., 1986; Miller, Ellis, Zook, & Lyles, 1990). For instance, O’Driscoll and

Beehr (1994) found that supervisors influence the degree of role stress and uncertainty

that their subordinates experience, which in turn may affect levels of satisfaction,

strain and turnover intentions. More specifically, when supervisors were perceived to

initiate structure, set goals, assist with problem solving, provide social and material

support, and give feedback on job performance, their subordinates experienced lower

ambiguity and uncertainty, and hence greater satisfaction with their job. Munn,
 85

Barber, and Fritz (1996) also found that only supervisor support was related to

turnover intentions among child care specialists. Their results indicated that both

emotional and instrumental support from supervisors positively affected an

employee’s professional well-being.

In contrast, other studies have found that coworkers are the most effective

person in terms of relieving employees’ strains with social support. For instance,

Greenglass, Burke, and Konarski (1997) examined the influence of supervisor and

coworker support on teachers’ burnout. Whilst supervisor support did not contribute

to the prediction of burnout, it was found that greater coworker support contributed to

the prediction of burnout, particularly decreased depersonalisation and increased

feelings of personal accomplishment. The support provided by the teachers’

coworkers comprised informational, practical, and emotional aspects. Greenglass et

al. (1997) proposed that it may be the possession of knowledge gained from

informational coworker support and significant help to complete work related tasks

gained from practical coworker support that may result in teachers more positive

feelings about their accomplishments on the job. Furthermore, the perception of

emotional support from coworkers may serve as a reminder to teachers of the

importance of relating to students as individuals.

In the present research, it is expected that nursing colleagues will be a

significant source of support to nurses. It is reasoned that because a nurse’s colleagues

are exposed to the same stressful working conditions, they will be in a better position

to empathise and understand the nature of the stressful event. Furthermore, since

nurses work in close proximity to their colleagues, it is likely that they will seek

assistance from their nursing colleagues before confronting other sources of support.

It is therefore hypothesised that:

 86

Hypothesis 9: Nurses will report higher levels of coworker support than

supervisor support.

Some studies (Winnubust, Marcelissen, & Kleber, 1982) seem to show that

various sources of support have different effects on different variables. For instance,

Gil-Monte et al. (1995) used a sample of 102 nurses to examine the relationship

between social support and role stress. Only supervisor support explained significant

variance in role ambiguity (β = -.37, p < .001). However, only coworker support

explained significant variance in role conflict (β = -.31, p < .01). In other words, for

nurses, supervisor support is negatively related to role ambiguity and coworker

support is negatively related to role conflict. More recently, in a study of 816 Dutch

oncology nurses, Peeters and LeBlanc (2001) found that coworker support moderated

the relationships between job demands and depersonalisation.

Finally, other studies have found that supervisor and coworker support act in a

similar manner. For instance, in a study of 1,498 social workers, El-Bassel, Guterman,

Bargal, and Su (1998) reported a high level of emotional support from coworkers and

supervisors was associated with less stress, depression, anxiety, irritability and few

somatic complaints among social workers. It is evident that there is a lack of

consistency regarding the effects of supervisor and coworker support on strain-related

outcomes. It would seem, however, that supervisor support and coworker support are

differentially effective in reducing stressors and strains.

In summary, researchers have conceptualised the term social support from two

distinct aspects. The functional nature of social support is commonly examined in the

occupational stress literature, where as the structure of a person’s social relationships

is most often examined in the life stress literature. Each approach focuses attention on

different, yet important, aspects of the social support phenomenon. In the following
 87

section, three models of social support are briefly explained before examining the

methodological issues that may have attributed to the ambiguous and inconclusive

findings regarding the effects of social support in the job stress research.

Models of Social Support

One of the principal aims of Study 3 is to investigate the differential effects of

social support on burnout. Although different frameworks have been used to explain

the effects of social support on a wide variety of outcomes, three models have

emerged in the social support literature. A number of authors (e.g., Beehr, 1985;

House, 1981; Kong et al., 1994) have proposed that social support reduces burnout in

one of three ways: (a) by acting directly on strains (i.e., main/direct effect), (b) by

acting directly on stressors (i.e., an indirect effect), and (c) by interacting with

stressors so that “the relation between stressors and strains is stronger for persons with

low levels of support than for those with high levels of support” (i.e., an interaction or

buffering effect as noted by Ganster et al., 1986, p.102). Of these, the buffering effect

model has received the most attention and caused the most debate in the stress-coping

research. A synthesis of these three models of social support is represented in Figure

4.1 in the context of a triangle relationship.

 88

This figure is not available online. Please consult the

hardcopy thesis available from the QUT Library.

Figure 4.1. Potential effects of social support on work stress and health.

_____________________________________________________________________

Note. From Work stress and social support (p.31), by J. S. House, 1981, Reading, MA: Addison –

Wesley.

The three models of social support are briefly discussed below.

Main effect model. The ‘main’ or ‘direct’ effect model assumes that social

support and work stressors act independently of one another on strains (Viswesvaran,

Sanchez, & Fisher, 1999). The main effect model of social support postulates that

social support reduces the level of strain regardless of the intensity of work stressors

experienced (Beehr, 1985; Cohen & Wills, 1985; Sullivan & Bhagat, 1992).

The main effect model of social support is most commonly investigated in

terms of the correlation between social support and strains. In operational terms, a

negative correlation between a measure of support and a measure of strain can be

construed as supporting evidence for this process mechanism. A different form of the

main effect model would be reflected by a positive correlation between social support

and strain (Viswesvaran et al., 1999), for example, in the case where support is

mobilised when strains are encountered (or conversely, the support providers extend

support primarily when the individual is afflicted with strain). Alternatively,

hierarchical regressions may be conducted to simultaneously test for both main and
 89

buffering effects on strain. In hierarchical regression analyses, the scores of the

predictor variables (e.g. stressors, social support) are entered separately into the

regression equation in Step 1, while the interaction terms (cross-product of stressors

and social support) are entered in Step 2. A significant standardised partial regression

coefficient reflects evidence of a main effect.

The idea that social support can directly reduce strains is consistent with most

empirical literature on job stress (El-Bassel et al., 1998; Ellis & Miller, 1994). Kahn

and Byosiere’s (1992) meta-analysis of the social support literature found that the

majority of empirical studies (20 out of 22) provide evidence of main effects of social

support (from supervisors, coworkers, and non-work supports) on levels of well-

being. Viswesvaren et al.’s (1999) meta-analysis of 68 social support studies also

found supporting evidence for a main effect of support on strain. The sample size-

weighted average corrected correlation across measures of support and measures of

strain was -.21. The upper 90% credibility interval was also negative (-.08 to -.34),

suggesting that social support and strain are negatively correlated across most

situations.

Nursing literature that examines the influence of social support on burnout

also provides reliable evidence for main effects. For instance, Eastburg et al. (1994)

found in a sample of 76 nurses a strong negative correlation (r = -.58, p < .001)

between work-related support and emotional exhaustion. Ellis and Miller (1994)

found that emotional coworker support was negatively related to all three aspects of

burnout, whereas instrumental coworker support was negatively related to emotional

exhaustion and depersonalisation among nurses. Duquette et al. (1995) found in their

study of 1,545 geriatric nurses that supervisor support (β = -.10) and peer support (β =

-.08) had a significant negative effect on burnout (p < .001). More recently, Kilfedder
 90

et al. (2001) examined the influence of social support on burnout among 510

psychiatric nurses and reported that lower levels of emotional exhaustion (r = -.30, p

< .001) and depersonalisation (r = .20, p < .001) are associated with greater

availability of social support.

The results of these studies imply that nurses with lower levels of social

support have a higher risk of experiencing burnout. Cohen and Wills (1985) propose

that irrespective of the level of stress, social support is likely to have a beneficial

effect on well-being as it provides a person with a sense of social integration. Despite

relatively strong empirical support for the main effect model, some researchers

(Cohen & Wills, 1985; Himle et al., 1991; Kahn & Byosiere, 1992; Terry, Nielsen, &

Perchard, 1993) argue that this model only explains a proportion of the variance in the

complex relationship between social support and strains, and thus other models of

social support (indirect effect, buffering effect) have been incorporated into empirical

investigations (LaRocco et al., 1980). Based on the sound supporting evidence for the

main effect model, the following hypothesis was generated:

Hypothesis 10: Work support will be significantly negatively correlated to

emotional exhaustion and depersonalisation and significantly positively associated

with personal accomplishment.

Indirect effect model. The indirect effect model postulates that social support

exerts an indirect effect on strains by directly reducing the impact of work stressors.

Cohen and Wills (1985) proposed that if social support affects the appraisal of the

situation, the situation may be perceived as less threatening, and therefore perceived

stressors will be fewer and weaker. This model has also been referred to as the

insulating hypothesis (Antonucci & Depner, 1982) or the stress-prevention model of

social support (Barrera, 1986).

 91

In operational terms, social support is negatively correlated to stress. From this

perspective, social support prevents the occurrence of stressful conditions or reduces

the likelihood that events would be perceived as highly stressful. The important

feature of this ‘stress-prevention’ model is that social support has no direct influence

on strains.

The stress-prevention model was demonstrated in an earlier study by Dignam,

Barrera, and West (1986). Correctional officers in county prisons were the

participants in this cross-sectional study. Supportive transactions between coworkers

and supervisors were hypothesised to prevent job stress by decreasing role ambiguity

and thereby influencing burnout levels. Using path analysis, there was no support for

the buffering effect or direct effect between social support and burnout. However, the

data fit a model that linked support with reductions in role ambiguity, which led to

reductions in burnout symptoms. In particular, the zero-order correlation between

social support and perceived stress was negative and significant, but social support

was not related to burnout. However, the cross-sectional design of the study and the

use of only single indicators of each construct precluded a strong test of the causal

relationships implied in the indirect model.

Dignam and West (1988) in their study of state correctional officers, however,

found no evidence to support the indirect effect model using path analysis. Indeed, the

direction of the structural coefficient between workplace support and job stress was

positive, directly contrary to the prediction of the indirect effect model. Two

differences between the two studies were attributed to the discrepant results. First, in

Dignam et al.’s study, the indirect effect of social support on burnout operated

through its direct effect on role ambiguity, whereas role ambiguity was removed from

the analyses in Dignam and West’s (1988) study because of problems with its
 92

convergent and discriminant validity. Second, Dignam et al. used a composite

measure derived from measures of satisfaction with support from supervisors and

coworkers as well as perceptions of support frequency, whereas Dignam and West

(1988) used measures from coworkers and supervisors as indicators of social support.

In a review of the social support literature, Barrera (1986) observed that studies that

utilise measures of perceived availability and adequacy of social support or mixed

social support scales (e.g., social embeddedness and perceived social support) were

more likely to report results consistent with an indirect model.

In a study of 500 scientists working in U.S. pharmaceutical firms, Kong et al.

(1994) found some mixed evidence for an indirect effect. They investigated the effect

of social support (family and work support) on work stress and organisational

commitment. Using bivariate correlations and multiple regression, family support

was negatively correlated to role ambiguity and future ambiguity, and both role

ambiguity and future ambiguity had a negative effect on organisational commitment.

Family support was positively correlated to organisational commitment, but the effect

became non-significant when role ambiguity and future ambiguity was added to the

equation. Their findings suggested that family support had a significant indirect effect

on organisational commitment via reduced role ambiguity and future ambiguity.

There was no evidence, however, that workplace support had an indirect effect on

organisational commitment via role stress. Workplace support, however, did have a

significant direct effect on organisational commitment.

In a more recent study, Viswesvaran et al. (1999) meta-analytically cumulated

the correlations between support and stressors to illuminate the validity of the indirect

model. Viswesvaran et al. reported a sample size-weighted correlation between

measures of stressors and measures of support of -.12 (with a 90% credibility interval
 93

of -.00 to -.24). This estimated true score correlation (after taking into account

sampling error and unreliability in the support and stress measures) suggests that,

although evidence of an indirect effect is weak, social support does mitigate perceived

stressors.

Since evidence for an indirect effect is weak, the present study will not

specifically test the indirect effect model. Instead, the present study will focus

primarily on the two models that have gained the most empirical attention in the

social support literature: the main effect and buffering effect. However, based on the

above findings, it is expected that work support will influence nurses’ perceptions of

work stress. Thus, the following hypothesis was generated:

Hypothesis 11: Work support will be significantly negatively correlated to

work stress.

Buffering effect model. The ‘buffering effect’ hypothesis postulates that an

individual experiencing significant stress, but with strong social support, will be

protected from developing stress-related outcomes as compared to an individual with

weak support (Cohen & Wills, 1985). The buffering hypothesis states that the

relationship between stress and strain is strongest at low levels of social support, and

the effects of support only become apparent in high stress environments (Cohen,

1988; Cohen & Wills, 1985; House 1981). According to the buffering model, the

relationship between work stress and burnout is contingent upon social support. In

this sense, social support may have no effect on burnout when employees experience

a low level of stress, or no stress, as support is not actualised or mobilised, even

though potential support is available. However, when employees experience high

stress, social support may be mobilised either by requests from people in need or by

offers from others who recognise someone in need.

 94

The stress-buffering model postulates that social support buffers or protects

individuals against the deleterious consequences of work stress. Some confusion and

imprecision has accompanied the use of the term ‘buffering.’ Most authors, however,

have used the term in a way consistent with the explicit definition of LaRocco et al.

(1980). Within this definition, social support is hypothesised to interact with stressors

such that the correlation between stressors and strains is stronger for persons with low

levels of social support than those with high levels of support. In other words, social

support moderates the stress-strain relationship. It is important to distinguish between

the terms moderate and mediate as both have been used in reference to the buffering

effect of social support. A mediation effect occurs to the extent that a mediator

variable such as social support, accounts for the relationship between a given

predictor (e.g., work stress) and an outcome of interest (e.g., burnout), such that the

mediator specifies how or why the relationship occurs (Parker & Kulik, 1995). In

contrast, a moderating effect can be described as an interaction between a predictor

variable (e.g., work stress) and a moderator variable (e.g., social support), such that

the relationship between the predictor and an outcome variable differs depending on

the level of the moderator. In other words, the effect of work stress on burnout is

conditional upon social support. In the current study, the buffering role of social

support refers to a moderating effect; thus the terms buffering and moderating will be

used interchangeably.

To test for moderator effects, the most widely used data analytic strategy

involves conducting hierarchical multiple regressions and examining the increase in

R2 when the interaction term (the cross-product of stressors and support) is added to

the regression equation of strain on the main effects of stressors and support. An

alternative to the regression approach employs analysis of variance (ANOVA). In

 95

ANOVA, strain is conceptualised as a continuous dependent measure, but both

stressors and support are treated as discrete independent factors. A significant F value

corresponding to the interaction of the two independent factors is taken as evidence of

moderator effects. Because most stressors and social support measures are

conceptually closer to continuous rather than categorical variables, the use of

ANOVA is questionable. Not surprisingly, fewer studies have used ANOVAs to test

for moderator effects.

Support for the buffering effect has been controversial over the past decade of

job stress research and has been the subject of numerous investigations (Beehr &

McGrath, 1992; Himle et al., 1991; Koeske & Koeske, 1993; Terry et al., 1993;

Parasuraman, Greenhaus, & Granose, 1992; Manlove, 1994; Munn, Barber, & Fritz,

1996). Findings regarding the buffering effect have been mixed. Whilst several

investigators have found at least some evidence that social support buffers burnout

(Burke & Greenglass, 1995; Carr, Roseingrave, & Fitzgerald, 1996; Capner &

Caltabiano, 1993; Corrigan et al., 1994; Greenglass, Fiksenbaum, & Burke, 1996;

Koeske & Kelly, 1995; Peeters & LeBlanc, 2001; Schmieder & Smith, 1996), others

have found weak, or no evidence for the buffering effect of social support

(Bourbonnais et al., 1998; El-Bassel et al., 1998; Gil-Monte, et al., 1995; Koniarek &

Dudek, 1996).

Furthermore, some studies (e.g., Fenlason & Beehr, 1994; Kaufman & Beehr,

1986; Patterson, 2003) have found that not all buffering effects occur in the expected

direction, that is, to reduce strains. Occasionally a reverse buffering effect has been

reported such that, as the level of social support increases, the relationship between

stressors and strains becomes stronger. For example, in a study of 102 female nurses,

Kaufmann and Beehr (1986) found that both supervisor and coworker support
 96

interacted with work stressors (i.e., workload and future ambiguity) to predict strains.

Contrary to predictions, however, these interactions were in the opposite direction.

That is, the slope of the stressors-strain relationship was steeper at higher levels of

support than at lower levels of support (Beehr, 1995). Two possible explanations were

offered by the researcher. It was proposed that the supervisor may be the source of the

stress and therefore interaction between the subordinate and the supervisor may be

perceived as stressful by the employee (Blau, 1981). Alternatively, it could be that

supportive communications between employees may convince the stressed worker

that things are actually worse than the stressed individual thought (LaRocco et al.,

1980). Similarly, Fenlason and Beehr (1994) also found evidence to suggest that the

availability of social support may exacerbate, rather than buffer, the negative effects

of occupational stressors on well-being. For instance, Fenlason and Beehr reported

that instrumental coworker support interacted with role conflict to predict strains (F =

5.89, p < .05) such that relationship between role conflict and strains was more

positive under conditions of high instrumental coworker support. While

counterintuitive buffering effects are rare, these studies point to a need for further

research into the quality and nature of the support employees receive at work. The

current research proposes that:

Hypothesis 12: Work support will significantly buffer the relationship between

work stress and burnout.

Methodological Issues

Examination of the literature reveals that the debate over the stress-buffering

effect of social support is far from settled. Recent studies (e.g., El-Bassel et al., 1998;

Peeters & LeBlanc, 2001) investigating social support examine the buffering model in

addition to the main and indirect effect model. According to Cohen and Wills (1985),
 97

each model seems to be correct in certain situations but each represents a different

process through which social support affects well-being. Some researchers (Beehr,

1995; Fenlason & Beehr, 1994; Ganster et al., 1986; Terry et al., 1993) suggest that

the conflicting findings regarding the moderating influence of social support may be a

consequence of the ambiguity regarding the specific meaning of the construct and of

the ways in which social support has been measured.

As discussed earlier, the definition and operationalisation of the social support

construct has differed in objectivity, dimensionality, and meaning. Early researchers

tended to treat social support as a unidimensional construct and often used aggregate

support scores to examine the effect of social support on stress-related strains. Most

contemporary authors agree, however, that social support is a multidimensional

construct. Closer examination of the social support literature suggests that there is

some degree of overlap among the various definitions and component analyses of

support (Himle et al., 1991). However, due to the various definitions and component

analyses of social support, it is sometimes difficult to identify which type of social

support a researcher is measuring.

For the most part, it seems that researchers have not properly examined the

effects of different types of social support on the stressor-strain relationship. In fact, it

could be argued that the primary type of social support measured is actually emotional

support (e.g., El-Bassel et al., 1998) whilst other methods of support have been

neglected in much of the empirical literature on occupational stress. Occupational

stress studies examining the stress-buffering hypothesis have primarily used Caplan et

al.’s (1975) Caplan Social Support Instrument. This social support scale assesses the

perceived availability of support from three different sources: a) supervisor, b)

coworkers, and c) an extra-organisational source, such as friends or family. The scale

 98

comprises four items. The items assess the availability of what could be classified as

emotional support (e.g., “How much is each of the following people willing to listen

to your personal problems?” “How easy is it to talk to each of the following types of

people?”) and instrumental support (e.g., “How much can each of these people be

relied on when things get tough at work?” “How much does each of these people go

out of their way to make your life easier?”). Some researchers (e.g., El-Bassel et al.,

1998) however, assume that Caplan et al.’s (1975) Social Support Instrument assesses

only emotional support. The wording of the items is problematic. Since they are very

general in nature, it is difficult to clearly distinguish between the support components.

In addition, the belief that one can easily talk to one’s friends, for instance, does not

necessarily mean that they will be a useful source of support for work-related

problems (Greenglass et al., 1990; Terry et al., 1993). Furthermore, Caplan et al.’s

(1975) Social Support Instrument is limited by a small number of items. Despite the

obvious limitations associated with this measure, it continues to be the most

commonly used social support scale in the occupational stress literature.

Another possible reason why empirical studies may have failed to find

supporting evidence for the buffering effect of social support could be due to small

sample sizes and homogenous samples (Ganster et al., 1986). To obtain sufficient

statistical power to detect interactions, a relatively large sample is needed. Large

samples have not been the norm in social support literature and therefore much of the

apparent variability in reported findings may be a function of sampling error. The

majority of studies (e.g., Eastburg et al., 1994; Munn et al., 1996; Patterson, 2003;

Peeters & LeBlanc, 2001) employ a homogenous sample from a particular setting that

was selected purposively by the researcher (Edwards & Burnard, 2003). A better

sample would be one that represents a range of different work roles or groups, thereby
 99

increasing the size of the sample and allowing the testing of higher order interactive

effects.

It is evident from the literature that the buffering effect of social support is

more likely to be found when components of social support are examined, rather than

more global measures of social support. Several studies have been criticised for

failing to match the requirements of the stressful situation with a specific type of

support. Although stressful events may elicit needs for multiple resources, Cohen and

Wills (1985) argue that it is reasonable to assume that specific events elicit particular

salient coping requirements. This hypothesis predicts that buffering effects will be

observed when the support functions measured are those most relevant to the stressors

faced by the person. In other words, social support is more likely to buffer the

negative effects of stress if available support is able to address the needs of the

situation (Cohen & Wills, 1985). This line of thinking is referred to as the stressor-

support matching or specificity hypothesis.

Stressor-Support Matching Hypothesis

Some empirical evidence has been found for the stressor-support matching

hypothesis. For instance, Ellis and Miller (1994) in a study of 1,356 hospital nurses

examined the stress-buffering effects of emotional and instrumental coworker support

on burnout. The only equation in which the interaction term was significant was the

prediction of reduced personal accomplishment. In this equation, informational

coworker support buffered the effects of role ambiguity on reduced personal

accomplishment (β = -.11, F(2, 433) = 13.03, p <.000). The interaction term however,

only added 1% to the variance in reduced personal accomplishment, which suggests

that the stressor-support matching hypothesis although significant is relatively weak.

 100

In contrast, Greenglass et al. (1996) in a study of 833 school teachers found

that informational support from coworkers (β = -.29, p <.001) and supervisors (β = -

.38, p <.001) buffered the effects of stress on emotional exhaustion. Emotional (β = -

.52, p < .001), informational (β = -.51, p < .001) and practical (β = -.52, p < .001)

coworker support buffered the effects of stress on depersonalisation. Practical

supervisor support (β = -.37, p < .001) also buffered the effects of stress on

depersonalisation. No significant buffering effects were found when lack of personal

accomplishment was the criterion. Although sources of stress explained most of the

variance of burnout, the interaction between stress and support also contributed to the

explanation in burnout variance. Greenglass et al. (1996) did not report, however, how

much additional variance was explained by the interaction terms.

Terry et al. (1993) also examined the stress-buffering role of social support on

psychological well-being in a sample of 153 employees from a large public sector

organisation. They applied Cohen and Will’s (1985) stressor-support matching

hypothesis to examine different sources of support rather than different types of

support. Although Cohen and Will’s matching hypothesis does not allow for clear

predictions concerning the utility of different sources of support, Terry et al. proposed

that because of the proximity of work-based sources of support (supervisor and other

people at work) to the source of stress, these supports would be more likely than non-

work supports (partner, family and friends) to buffer the negative effects of work

stress. A similar argument for the buffering effects of sources of support was made in

earlier research by Beehr (1985) and House (1981). Their study found some evidence

to support Cohen and Wills’ (1985) stressor-support matching hypothesis. Their

results indicated that the perceived availability of supervisor support (β = .17, p < .05)

buffered the negative effects of work overload on both psychological well-being and
 101

job satisfaction, accounting for 3% additional variance in psychological well-being

(F(3,142) = 2.18, p < .05). There was also some evidence to suggest that high levels

of supervisor support (β = .23, p < .01) helped to buffer the negative effects of role

conflict on psychological well-being. Inclusion of the interaction terms involving role

conflict accounted for an additional 5% of variance in the psychological well-being

scores (F(3,142) = 2.86, p < .05). Contrary to expectations, however, no buffering

effects were found for coworker support. This pattern of results may suggest that

unlike supervisors, coworkers are not in a position of power to influence the sources

of stress.

Similarly, Schmieder and Smith (1996) examined different sources of support

(supervisor, coworker, spouse and friends/relatives) in shift working and non-shift

working nurses (N = 191). Since strongest buffering effects have been found in high

stress environments, they hypothesised that shift working nurses would demonstrate

stronger moderating effects of work support on the job stress-strain relationship than

non-shift workers because of the stressful nature of shift work and the social

integration difficulties associated with shift work. For shift workers, supervisor

support buffered the effect of role ambiguity in predicting global job satisfaction and

intent-to-quit. Coworker support also buffered the effects of role ambiguity in

predicting global job satisfaction. However, none of these relationships held for non-

shift workers. Schmieder and Smith suggested that it might be that shift work nurses,

in their more solitary non-work lives, look to their supervisors as a source of support

when they experience stress at work. Nurses who perceived their supervisors to be

supportive were less likely to allow workplace stress to lead to negative attitudes

about their job. The relationship between worker and supervisor may not be

psychologically important and effective in moderating stressful job conditions for day
 102

workers because they have access to more non-work sources of social support. Based

on these findings, it could be inferred that job context (e.g., shift work/non-shift work)

may influence the buffering effect of social support.

Peeters and LeBlanc (2001) in a study of hospital healthcare staff (i.e., nurses,

physicians, oncology nurses, and radiotherapy assistants) also examined the potential

moderating effect of three different sources of support (colleagues, supervisor and

family) on the relationship between job demands and burnout (i.e., emotional

exhaustion and depersonalisation) thereby predicting a specific match between the

source of social support and the type of job demand on burnout. No significant

buffering effects were found for emotional exhaustion. However, being confronted

with death and dying of patients led to less depersonalisation among employees with

high colleague support (β = -.08, p < .001) than among employees with low colleague

support. Although the interaction terms contributed little additional variance in

explaining burnout, the findings demonstrate that it is valuable to investigate the

source of support as well as the type of support when examining the stressor-support

matching hypothesis. Furthermore, this finding provides further evidence that the

effects of receiving social support are strongly affected by the nature of the

relationship between the support-giver (colleagues, supervisor or family) and the

support-receiver (employee) (Sarason, Sarason, & Pierce, 1994).

Although literature reviews generally conclude that existing evidence is

consistent with matching notions (Cohen & Wills, 1985; Cutrona & Russell, 1990),

there are few studies designed to test specific hypothetical predictions. It requires the

definition and measurement of distinct categories of stressors and of social support,

relatively orthogonal measures of subtypes within each category, and a conceptual

link between stress and support categorisations (Cohen & McKay, 1984; Tetzloff &
 103

Barrera, 1987; Viswesvaran et al., 1999). An elaboration of the matching model

proposed by Cutrona (1990; Cutrona & Russell, 1990) provides some of this

specificity.

Optimal Matching Theory

Cutrona’s (1990) optimal matching theory of support suggests that the

controllability of stressors is the primary dimension in terms of an appropriate match

between stressors and supporting requirements. Perception of controllability or

‘perceived situation control’ refers to the extent to which a person has control over

specific aspects of their work or work environment (Sullivan & Bhagat, 1992; Warr,

1996). It is important to note that perceived situation control is clearly distinguished

from locus of control. Locus of control is an enduring individual characteristic which

differentiates people who assume that what happens to them is mostly under their

personal control from people who believe that life events are mostly determined by

factors beyond their control (Kahn & Byosiere, 1992; O’Driscoll & Cooper, 1996).

According to Cutrona’s optimal matching support model, stressful events

perceived to be controllable are presumed to elicit needs for problem-focused coping

(instrumental support) to aid in preventing the event occurrence or consequences.

Stressful events perceived to be uncontrollable are presumed to elicit needs for

emotion-focused coping (emotional support) to help persons recover from the

negative emotions elicited by the event. It is generally thought stressors that are

perceived to be uncontrollable are more distressing than those that are more likely to

be under an individual’s control (Aldwin, 1994; Reich & Zautra, 1981). It is therefore

not surprising that emotional support is the most effective in managing uncontrollable

stressful events.
 104

Previous studies in the occupational stress literature have not attempted to

categorise work stressors into controllable and uncontrollable stressful events. This

gap in the occupational stress literature will be addressed in Study 3 in the present

research program. Two independent raters are asked to classify nurses’ stressful work

events based on situation controllability. This was a necessary step before analysing

the buffering effect of social support on the stress-burnout relationship.

Like Cohen (Cohen & McKay, 1984; Cohen & Wills, 1985), Cutrona (1990)

also emphasises matching supportive functions to the needs elicited by specific

stressor domains. However, in the current thesis, Cutrona and Russell’s (1990)

optimal matching model has been extended to the match between types of stressors

and types and sources of support. The present study assumes that if you get what you

want (type of support) from whom you want (source of support), social support will

have the most beneficial effect. In most studies, no explicit hypotheses are formulated

regarding the differential effects of sources of social support on stressors. Thus, in the

present research, specific hypotheses were only formulated about the match between

the type of stressors and type of support. The following hypotheses were generated:

Hypothesis 13: For stressful events deemed to be uncontrollable, emotional

support will have a significant buffering effect on burnout.

Hypothesis 14: For stressful events deemed to be controllable, instrumental

support will have a significant buffering effect on burnout.

Finally, it is important to note that success at indicating buffering effects is

primarily linked to research (Hutchinson, 1999; Schmieder & Smith, 1996; Turner,

1992) that has employed measures of perceived available support as opposed to

measures of received support. The literature on work stress has rarely examined the

relationship between perceived support and received support. It would seem,

 105

however, that the results of two studies (Cohen, McGowan, Fooskas, & Rose, 1984;

Wethington & Kessler, 1986) have been primarily responsible for the view that

perceptions of available support moderate the relationship between stress and

psychological outcomes and assessments of received support do not.

Cohen, McGowan, Fooskas, and Rose (1984) conducted a prospective study

with a sample of college students. They were interested in investigating the roles of

available and received support in moderating the relationship between negative life

events and psychological disorder (Dunkel-Schetter & Bennett, 1990). To assess

available support they used the college student version of the Interpersonal Support

Evaluation List (Cohen & Hoberman, 1983), and to determine received support they

used the Inventory of Socially Supportive Behaviours (Barrera, Sandler, & Ramsay,

1981). Outcome measures included the Beck Depression Inventory (Beck, 1967) and

the 22-item Langner Sympton Checklist (Langner, 1962). Stressful life events were

assessed by means of the College Student Life Events Schedule (Sandler & Lakey,

1982). Data was gathered on life events and outcome measures at two points with a

two-month duration between data collections. Data on social support was collected

only at the second point. Using regression analyses (both with and without initial

symptomatology controlled) they found both main and buffering effects for available

support but neither for received support. Cohen et al. concluded that it is the

perception that others will provide resources when they are needed that is key to

stress-buffering.

A second study conducted by Wethington and Kessler (1986) found similar

results. Wethington and Kessler analysed cross-sectional data from a large-scale

national survey. This study was different from Cohen et al.’s (1984) study in at least

two respects. First, the sample was a community sample and the respondents were
 106

married adults (aged 21-65 years). Second, it did not use standardised measures.

Stressful life events were assessed by asking respondents to describe "the last time

something really bad happened to you" (Wethington & Kessler, 1986, p.80).

Psychological distress was measured by a 20-item scale consisting of statements

about bodily feelings associated with depression and anxiety. Available support was

assessed by a single item -“These days I really don’t know who I can count on for

help,” and was rated on a 4-point scale ranging from 0 (very true) to 3 (not at all true).

Finally, for those individuals who reported a stressful event, the researchers

determined the received support for that event by asking them to describe who helped

and how they helped. Using these measures, Wethington and Kessler presented results

that document the primacy of cognition (perception of available support) over

received support in buffering the effects of stressful life events. They argue that

personal coping competence may be bolstered by the perception that support is

available if needed, whereas the actual mobilisation of support (i.e., received support)

may be needed only if personal-coping efforts fail.

Although the studies differed on several dimensions, they both indicated that

whether or not one actually receives support is less important for health and

adjustment than one’s beliefs about its availability (Cohen et al., 2000). Yet many

researchers contend that received support will have a greater influence on health

following a stressful event than will available support. For example, Gottlieb (1985)

stated that the behavioural manifestation of support (or its materialization in

interpersonal transactions) has the greatest significance for the course and outcome of

stressful experiences. Similarly, Gore (1985) observed that “the question of a stress-

buffering effect of social support, strictly speaking, is contingent upon evidence that

support is mobilized, not that it exists as a potential” (p.269). Thoits (1985) also
 107

proposed that support “may consist of words and deeds intended to alter the self-

perceptions of distressed individuals, and altered self-perceptions are the mechanisms

through which support operates to buffer, or reduce symptoms” (p.61). These authors

all imply that received support will be more powerful than perceived support in

creating a buffering effect. To date, it is impossible to make any conclusions

regarding the buffering effect of received support because received social support has

been largely ignored in the stress-support literature and previous measures of received

support have also lacked specificity. Dunkel-Schetter and Bennett (1990) argue that

received support, in terms of specific stressful situations and the support activation

process, should be considered in testing for buffering effects. Assessing the multiple

aspects of received support may be especially important when the sample used

experience different types of stressful events. It therefore follows that the levels and

types of support that would be beneficial to nurses would vary depending on the

stressful situation.

Summary

In summary, the present research proposes that it is important to document the

beneficial effects of perceived support as well as received support. It is evident that

the processes tapped by perceived and received support measures are not identical.

Conceptions of perceived available support are dispositional and cognitive, whereas

perceptions of received support are situational and behavioural, involving interactions

between individuals. In addition, Dunkel-Schetter and Bennett (1990) in a review of

empirical studies found that correlations between perceived and received support

ranged from .01 to .46. None of the studies reported more than 21% shared variance

between measures of the two aspects of support. It would therefore seem that both

perceived and received support require further study in the social support literature.
 108

In addition, the current literature review highlights the complexities

researchers confront when wanting to examine the social support construct. Social

support is an ambiguous concept and its role in reducing or preventing strains remains

unclear. In the present thesis, the social support concept will be first examined from a

qualitative perspective in Study 1. Focus group discussions will be conducted with

hospital nurses to examine nurses’ perceptions of work support. More specifically,

nurses’ views on their main sources of support and the types of support provided to

them whilst at work will be investigated. In Study 2, a measure of support will be

developed to assess the types of support nurses receive at work from their key support

providers. In Study 3, quantitative data will be examined for both main and buffering

effects of social support using hierarchical regressions. Study 3 will provide insight

into whether the types of support Australian nurses’ receive from work reduces

burnout. The thesis will provide further insight into the perceived functions of social

support and the role received support plays in the stress-burnout relationship. In the

following chapter, the major methodological considerations pertinent to this research

program are detailed.

 109

CHAPTER 5

Methodological Considerations

Chapter 5 presents important methodological issues pertinent to this research

program. The researcher details the steps taken to ensure that the current research

program adequately meets principles of scientific rigour. First, matters related to

choosing a representative sample are outlined. Next, the strategies used to enhance the

reliability and validity of the qualitative findings are discussed. Following this, the

reliability and validity of scales used to measure the main variables of interest to this

program are described. This chapter also discusses the rationale for employing a

qualitative research methodology in Study 1 and a quantitative methodology for the

final two studies of the research program. In addition, the principal research questions

and hypotheses that comprise these studies are delineated. Finally, an overview of the

data analysis techniques performed for each study are briefly summarised.

Sampling Issues

An important methodological objective for all empirical research is the ability

to draw conclusions about a larger population. This largely depends on the

representativeness of the sample. A representative sample is one in which the

sample’s characteristics mirror the population from which it comes. Associated with

the issue of representativeness is how the level of representativeness is demonstrated.

Of particular importance to the achievement and demonstration of representativeness

is the method by which respondents are selected to participate. Other issues pertinent

to this matter include sample size and the representativeness of the sample’s work

environment to that of the population from which it was drawn. Addressing this

matter goes to the centre of being able to draw valid conclusions about a population

from sample data. Issues pertaining to sample size are first discussed.
 110

Sample Size

The current program employs both qualitative and quantitative research

methods. Deciding on a sample size for qualitative inquiry can be difficult as there are

no definite rules to follow. The focus of qualitative research is on depth and richness

of information, rather than on sheer numbers of participants (Jones, 2002). According

to Patton (1990), determining an appropriate sample size in qualitative research

depends on “what you want to know, the purpose of the inquiry, what’s at stake, what

will be useful, what will have credibility, and what can be done in the available time”

(p.184).

In Study 1, a focus group methodology was used in an attempt to draw

conclusions about (1) the types of stressors nurses’ experience at work, and (2) the

main sources and types of support nurses receive at work. In order to obtain

reasonable coverage of the possible responses to the research questions, the researcher

purposefully recruited participants from a variety of hospital wards. The researcher

focused on wards that are typically viewed as stressful in the nursing literature (e.g.,

Intensive Care Unit, Emergency Department), as well as more general wards (e.g.,

Maternity, Operating Theatre). The researcher used a sampling method recommended

by Lincoln and Guba (1985) and Jones (2002) in which sampling continued to the

point of redundancy, stopping at the point at which no new information was coming

forth from participants. A total of 10 focus groups in 8 wards were conducted in order

to gain adequate answers to the research questions. A total of 68 nurses participated in

Study 1.

Studies 2 and 3 comprised the quantitative components of this research in

which a survey methodology was used to collect data on the main variables of interest

in this program. The statistical technique used in Study 2 was factor analysis. In
 111

determining the sample size for factor analysis, Comrey and Lee (1992) offer the

following guidelines: sample sizes of 50 are very poor, 100 are poor, 200 are fair, 300

is good and 500 is very good. As a general rule of thumb, a sample size of 300 is

comforting when performing factor analysis, however solutions that have several high

loading variables (>.80) do not require such large sample sizes and 150 cases are

considered sufficient.

The statistical analysis performed in Study 3 involved correlations and

regressions. In determining an appropriate sample size, Tabachnick and Fidell (1996)

recommend the following: N ≥ 50 + 8m for testing multiple correlations or N ≥ 104 +

m (m is the number of IVs) for testing individual predictors in a regression equation.

These rules of thumb assume a medium sized relationship between the independent

variables (IVs) and the dependent variable (DV), and that α = .05 and β = .20. N was

calculated both ways to determine what procedure produced the highest number of

cases.

Using the first formula, a sample size of 178 was generated (i.e. N ≥ 50 +

(8)(16) = 178). Using the later formula, a minimum sample size of 120 subjects (i.e.,

104 + 16 = 120) was required. Furthermore, when taking into account effect size,

Green (1991) offers the following, more complex rule: N ≥ (8/f2) + (m-1), where f2 =

.05 for a relatively small effect size. Using this formula, a sample size of 175 was

computed (i.e., N ≥ (8/.05) + (16-1) = 175). Based on the above calculations for Study

2 and Study 3, it was deemed that a sample size of 200 or greater would be

appropriate. In the present study, a sample size of 273 was obtained. In the following

section, the representativeness of the sample is discussed.

 112

Sample Representativeness

The focus of this research program is Australian nurses working across public

hospital settings. Ideally, a representative sample for the present study would require

randomly attained nurses from each ward in public hospitals across Australia. This

sampling approach however, was not practical and beyond the scope of the present

study. Due to limited financial resources, timeline constraints, and the inability to

adequately access hospital nursing staff in all areas of Australia, purposeful sampling

was undertaken. The researcher recruited nursing staff from three public hospitals in

Queensland. These hospitals were located in close proximity to one another and were

easily accessible. Furthermore, permission was granted from the Executive Director

of Nursing Services and ethics committee from each hospital. Although permission

was sought from other public hospitals in Queensland, the lengthy ethical clearance

process adopted by these hospitals ultimately excluded them from the research

program. For instance, some hospital ethics committees met quarterly or in one case,

only once per year. Hence, the present research was confined to public hospitals

within the Brisbane metropolitan district. In the section below, the sample

characteristics are detailed.

Sample Selection for Study 1

In Study 1, ward nurses from two public hospitals within the Brisbane

metropolitan district were invited to participate in focus group discussions. The

number of patients these hospitals attend to per year differ substantially. Hospital A is

an 850-bed hospital that serves the southern suburbs of Brisbane, Queensland. It

covers all major adult specialties, and has nationally recognised expertise in spinal

injury management and solid organ transplants. The hospital provides the focus for

the Major Trauma Service covering all of southern Brisbane; a road-based retrieval
 113

service, and a statewide aero-medical service. A comprehensive cardio-thoracic

facility has been a recent addition to the hospital’s service. The hospital treats

approximately 60,000 inpatients per year and over 340,000 outpatients. The

Emergency Department and Regional Trauma Service is one of the busiest in

Queensland. The full-time equivalent staff is about 3,500. During the time in which

the study was being conducted Hospital A was undergoing a major hospital

redevelopment to establish a tertiary teaching, research, and clinical facility.

In contrast, Hospital B is a relatively new, small, general acute public hospital

that opened in January 1999. There are 115 acute beds consisting of: 39 Medical

beds; 19 Surgical beds; 20 Obstetric beds; 7 Paediatric beds; 6 Neonatal ‘cots’ and 24

Mental Health beds. The hospital has four operating theatres and an Emergency

Department that is the seventh busiest in Queensland. The hospital’s focus is directed

at the management of common conditions and it has close ties with Queensland’s two

largest public hospitals. The hospital treats approximately 37,000 inpatients per year

and 115,000 outpatients. The full-time equivalent staff is about 321.

Following consent from the hospital’s ethics committee to conduct the

research, the researcher approached the Executive Director of Nursing Services from

each hospital to discuss the research project and to seek their advice on the selection

of potential hospital wards that were sufficiently diverse in patient illnesses and

medical conditions, as well as nursing care. This is known as ‘judgement sampling.’

Judgement sampling refers to “a sample that is obtained according to the discretion of

someone who is familiar with the relevant characteristics of the population” (Mugo,

2003, p.5).
 114

Five wards at Hospital A participated in the study. These included the

Orthopaedics ward (6 females, 4 males), Rehabilitation ward (8 females, 1 male),

Renal Unit (5 females, 1 male), Emergency Department (5 females, 3 males) and

Intensive Care Unit (10 females, 2 males). Three nursing wards from Hospital B also

participated. These included the Operating Theatre (7 females), Maternity ward (9

females) and the Emergency Department (6 females, 1 male). Overall, 56 female

nurses and 12 male nurses agreed to partake in the discussions.

It was necessary to facilitate two focus groups in the Intensive Care Unit

(Hospital A) and the Emergency Department (Hospital B) to ensure that these wards

were adequately staffed during the discussions. Therefore, two focus groups of six

nurses were conducted at the Intensive Care Unit over two consecutive days.

Similarly, one focus group of three and one focus group of four nurses were

conducted at the Emergency Department over two consecutive days.

Altogether, ten focus group discussions involving 68 nurses, in groups ranging

between three and ten individuals were conducted. Forty-five nurses (n = 45) were

employed at Hospital A and twenty-three nurses (n = 23) were employed at Hospital

B. Each discussion consisted of a homogenous group of nurses from a specific ward.

The nurses’ supervisors (e.g., Central Nurse Coordinator, Nursing Practice

Coordinator) were not invited to attend the focus group discussions in case their

presence may have influenced the extent to which nurses were prepared to discuss

issues openly and comfortably.

 115

Sample Selection for Studies 2 and 3

In Study 2, nurses completed a self-report survey measuring job-specific

stress, role stress, work support (i.e., supervisor and coworker support) and burnout

(see Appendix A). The construction of the survey is discussed in detail later in this

chapter. In addition, the sociodemographic profile of the participants was collated by

asking a series of questions that defined their demographic characteristics and

employment details. The data collected in Study 2 was also subject to further analyses

in Study 3. Consequently, the sample characteristics for Study 3 are the same as those

for Study 2.

The Executive Director of Nursing Services from each hospital gave the

researcher permission to survey nurses from the same wards that participated in Study

1. However, to increase the potential sample size, the researcher gained permission

from the Executive Director of Nursing Services at a third hospital to recruit

participants for Study 2. Hospital C is a 790-bed general hospital with a number of

specialties including medicine, surgery, orthopaedics, psychiatry, oncology and

trauma services. In contrast to the Hospital A, which primarily services the southern

suburbs of Brisbane, Hospital C predominantly services the northern suburbs of

Brisbane. Hospital C is the largest referral hospital in Queensland and services

patients throughout the state, northern New South Wales, the Northern Territory and

from neighbouring countries in the South West Pacific. The hospital treats

approximately 303,000 inpatients per year and 720,000 outpatients. The full-time

equivalent staff is greater than 4000.

Consequently, participants were drawn from two large public hospitals and

one smaller public hospital. A total of 273 nursing professionals (235 females, 38

males) completed the survey in Study 2. Five incomplete surveys were discarded.
 116

Participants represented a variety of nursing wards including: Orthopaedics,

Emergency, Renal, Adult, Operating Theatre, Children’s, Intensive Care,

Rehabilitation, Maternity, Surgical, Burns, Oncology, and Outpatients. Following

discussions with nursing practitioners, the different wards were categorised into seven

divisions: Medical, Surgical, Oncology, Critical Care, Maternity, Administration and

Other. The desired minimum sample size was obtained and the overall response rate

for this questionnaire was 67.74 percent. The response rate for each hospital varied.

For Hospital A, a total of 200 surveys were distributed to five wards and the response

rate was 60.5%. For Hospital B, a total of 75 surveys were distributed to three wards

and the response rate was 54.7%. For Hospital C, a total of 150 surveys were

distributed to twelve wards and the response rate was 74%. A possible reason for

nurses deciding not to participate in this research may be due to the growing number

of surveys being targeted at nurses in a bid to improve efficiency as stipulated in

recent Queensland health reforms. Some nurses indicated to the researcher that they

were becoming increasingly disinclined to complete yet another questionnaire

because they felt that imparting information about themselves or their workplace did

not result in constructive and/or effective organisational change.

Descriptive statistics for the sample population for Studies 2 and 3 are

presented in Table 5.1. Subject numbers and percentages are reported for hospital,

age, gender, employment status, nursing qualifications, number of years nursing, and

division of nursing.
 117

Table 5.1

Summary Statistics for the Demographic Variables

Variable Categories n %

Hospital A 121 44.3

B 41 15.0
C 111 40.7

Gender Male 38 13.9

Female 235 86.1

Age < 25 yrs 37 13.8

25-29yrs 71 26.4
30-39yrs 76 28.3
40-49yrs 57 21.2
> 50 yrs 28 10.4

Employment Status Full-time (> 35 hours) 173 64.6

Part-Time (< 35 hours) 87 32.5
Casual 8 2.9

Qualifications Enrolled* 6 2.5

Registered 266 97.5

Nursing Experience < 1 yr 12 4.4

1-2 yrs 17 6.3
2-5 yrs 36 13.3
5-10 yrs 68 25.1
>10yrs 138 50.9

Division of Nursing Medical 29 10.6

Surgical 121 44.3
Oncology 3 1.1
Critical Care 87 31.9
Maternity 20 7.3
Administration 8 2.9
Other 5 1.8
Note. * Enrolled nurses work under the direction and supervision of a registered nurse to provide
patients with basic nursing care. Registered nurses undertake all aspects of patient care, including
complex care procedures.

Despite a lower response rate than desired, it can be seen from Table 5.1 that

the sample population did resemble the sociodemographic composition of Australia’s

nursing population during 2000-2001. Nursing remains a predominantly female

profession and this was evident in the sample population in which 86% of nurses were

female. Approximately 14% of the sample population were male. Queensland labour

force statistics in 1999 indicated that male nurses represented approximately 10% of
 118

the nursing workforce (Queensland Health, 2001). It would appear therefore that male

nurses are slightly over-represented in the present study. However, since the

proportion of males commencing basic nursing courses has been maintained at around

13-15% since the 1990s, there has been an increasing trend of males joining the

nursing workforce. Hence, it is proposed that the percentage of males in the present

study may be a better estimation of the number of male nurses that were working in

public hospitals during 2001.

The majority of nurses were employed in the surgical division (n = 121,

44.3%), followed by critical care (n = 87, 31.9%), medical (n = 29, 10.6%), maternity

(n = 20, 7.3%), administration (n = 8, 2.9%), other (n = 5, 1.8%) and oncology (n = 3,

1.1%). The majority of participants were registered nurses (97.5%) working full-time

(64.6%) and aged between 30-39 years (28.3%). Over a third of the sample population

were aged 40 years or more, supporting recent statistics by the AIHW (2002)

suggesting that the nursing profession has an ageing population. Approximately a

third of the sample population were working on a part-time basis. In Australia,

increasing numbers of nurses are choosing to work on a part-time or casual basis.

Finally, the majority (76%) of the sample population had five years or more nursing

experience.

In summary, the researcher endeavoured to take appropriate steps to choose a

representative nursing sample despite the study’s limitation of having a relatively

small catchment area for recruiting nursing respondents. The researcher purposefully

sampled nurses from a variety of nursing wards across three public hospitals to ensure

that different types of nursing care were represented in this study. Although the

response rate was lower than desired, it was demonstrated that the sample’s

characteristics are similar to the Australian nursing population. Although empirically

 119

it is not possible to make generalisations about the larger nursing population, there is

good reason to believe that the results obtained from this research were typical of the

Australian nursing population during the period of 2000-2001.

Measurement Issues – Reliability and Validity

The quality of conclusions that can be drawn from the study’s results also

depends on two important requirements - the reliability and validity of the

measurement instruments and data. Reliability means “dependability or consistency,

that is, results are not erratic, inconsistent or unstable” (Neuman, 2000, p.164).

Validity suggests that “results are truthful and that there is a match between ‘reality’

and the theories, concepts, ideas and descriptions the researcher has used to analyse

the world” (Neuman, 2000, p.164). The reliability and validity of research findings is

an important consideration in both qualitative and quantitative research. The steps

taken to ensure the findings from this research are reliable and valid are discussed

below.

Qualitative Research – Study 1

There were various steps taken to ensure the findings from the qualitative data

were reliable and valid. In Study 1, the following actions were taken to make certain

scientific rigour was maintained. As with other approaches to studying social

phenomena, designing the focus group study required careful thought and reflection.

The first stage of the research process involved writing a clear statement of the

research problems to be addressed in Study 1. Next, the researcher delineated the

principal research questions and sub-problems. A plan detailing how the information

was to be obtained, from whom the information would be obtained, and the resources

required to conduct the focus groups, was established. Specific focus group questions

were then generated.

 120

Next, a pilot focus group was conducted with nursing staff from an aged care

facility. This enabled the researcher to determine whether the wording of the focus

group questions was easily understood by participants and more importantly, to

determine whether the focus group questions produced adequate data for answering

Study 1’s principal research questions and associated sub-problems. The pilot focus

group was audio-taped with the participants’ consent and later transcribed verbatim.

The data was then analysed according to the principles of content analysis. Further

details regarding the content analysis process is presented later in this chapter.

Analysis of the data confirmed that the focus group questions were suitably worded

and that they provided sufficient data to address the principal research questions and

sub-problems examined in Study 1.

The researcher was aware that in order to obtain honest information from

nurses, a relaxed and open environment in which nurses felt comfortable sharing

information was desirable. To do this, the researcher conducted separate focus groups

with small, homogenous groups. In addition, focus groups were conducted in a

familiar environment (i.e., staff common rooms or staff meeting rooms). Furthermore,

the nurses’ supervisors were not invited to attend the discussions. To help establish a

rapport with the nurses, the researcher visited each ward prior to facilitating the focus

group discussions. The researcher spent some time with nurses discussing the purpose

of the research, how the data would be used, issues of confidentiality and anonymity,

and addressed the nurses’ queries about the research.

When conducting the focus groups, the researcher was careful not to influence

the direction of discussions. The researcher utilised the strategies recommended by

Kreuger (1993) to minimise researcher bias such as listening without interjecting,

using neutral verbal responses such as ‘uh huh,’ ‘mm,’ ‘okay,’ and ‘thank you,’ and
 121

by not offering personal opinions. At the same time, the researcher tried to keep

participants from deviating too far from the topic by asking appropriate follow-up

questions and re-directing participants to the task at hand. The researcher was also

aware that the more dominant members of the group might influence what other

members were willing to say. The researcher was careful to maintain a balance in

participation by encouraging quieter participants to respond and by diplomatically

shifting the conversation away from the most active talkers.

The researcher also took careful notes during the focus groups and recorded

observations about the nursing group (e.g., group dynamics, non-verbal

communication between members) and additional information that was relevant to

understanding the nurses’ views and perceptions (i.e., opinions that were commonly

shared by the group, minority opinions, and when shifts in opinions occurred). As the

focus groups came to an end, the researcher summarised the main points raised by the

nurses in response to each question. Participants were asked if the summary was

accurate and complete. This was an opportunity for nurses to provide further

information if important points had been missed. It also enabled the researcher to

check the reliability of her notes.

The focus group discussions with the hospital nurses were audio-taped to

ensure the information was accurately recorded and to prevent any loss of data. The

researcher listened to each tape a number of times and transcribed the interviews. To

aid the researcher’s contextual understanding of the discussions, observational data

recorded during the focus groups was added to the transcripts.

The data was then analysed using a systematic technique known as ‘content

analysis.’ Krippendorf (1980) describes content analysis as “a research technique for

making replicable and valid inferences from data to their context” (p.21). Content
 122

analysis entails classifying data at multiple levels that must be theoretically justifiable

(Berg, 1995; Dey, 1993). Classifying the data is an integral part of the analysis as it

lays the conceptual foundations upon which interpretation and explanation are based

(Dey, 1993).

The process of developing content categories was one of continuous

refinement and the criteria for assigning responses to specific categories evolved over

time. To understand the deep structural meaning conveyed by the text, the researcher

recorded all the issues that were discussed in response to each question before

creating broad overarching categories that depicted the main themes/issues identified

by the nurses. The researcher consulted relevant literature and theories to ensure that

broad content categories were conceptually and empirically grounded.

Finally, comprehensive instructions for assigning data to categories were

gradually developed. This involved assigning units of data (key words, sentences

and/or phrases) to categories. Categories were then further refined and subcategories

were identified. To assist in the validation of the content categories, at least three

independent examples were supplied for each subcategory. Heeding the advice of Dey

(1993), the researcher also kept notes during the classification process on how

decisions about categorising data were made, why some issues were considered

important and why others were discarded. These steps ensured that the reliability and

validity of the qualitative data was maintained.

Quantitative Research – Study 2 and 3

From a quantitative perspective, the reliability and validity of the findings

predominantly depends on the instruments chosen to measure the variables under

investigation. Self-report measures must demonstrate adequate levels of reliability

 123

and validity in order to be confident that research findings are accurate and not the

result of measurement error.

Self-Report Data

The use of self-report data has been very common in investigations of

psychological variables. Some of the advantages associated with the use of self-report

data to draw conclusions about the psychological function of humans are simply that

it is often the most effective method for determining the experiences of subjects under

investigation. This approach entails a standardised procedure for asking subjects how

they feel, what they think, what they believe, or what attitudes they have toward a

particular stimulus. Asking is often the best method of obtaining the information

required. It also has the advantage of being quick and relatively inexpensive in

comparison to alternative methods of obtaining information on a topic. Alternative

data collections usually require the experimenter to impute meaning to more

‘objective’ observations and this requirement often leads to significantly longer and

more invasive data collection processes to ensure valid interpretations are made.

Just as the objective measurement of a subject’s response to a stimulus is open

to the influence of experimenter bias in recording and interpretation of such data, the

self-report methodology is also open to inaccuracies. Failure to recall the information

being asked about is one source of inaccuracy (Mayo, 1983). This problem, however,

is unlikely to be significant in this program of research as the instruments chosen ask

questions of nurses that are not historically focused. Instead, they focus on the present

and/or immediate past behaviours, feelings and observations.

The most likely source of error that could influence this program of study

would be common method bias which results from the fact that the predictor and

criterion variables are obtained from the same person (Podsakoff, MacKenzie, &
 124

Podsakoff, 2003). There is a substantial body of theory (Heider, 1958; Osgood &

Tannenbaum, 1955) and research (McGuire, 1966) suggesting that people try to

maintain consistency between their cognitions and attitudes. According to Podsakoff

et al. (2003), people responding to questions posed by the researcher may have a

desire to appear consistent and rational in their responses and may search for

similarities in the questions asked of them - thereby producing relations that would

not otherwise exist at the same level in real-life settings. This ‘consistency effect’ is

likely to occur when respondents are required to provide retrospective accounts of

their attitudes, perceptions and/or behaviours. To help reduce the effects of response

pattern biases, measures were used that had different scale response formats (e.g.,

never/rarely to very often, strongly disagree to strongly agree) and reverse-coded

items were included.

There is also the potential for responses to be prejudiced by social desirability,

with participant’s overestimating or underestimating their level of work stress, social

support or burnout, depending on how they wish to represent themselves. Answers on

questionnaires may not only be a function of respondents’ true perceptions, but also of

their own agenda regarding their participation in the research and how they wish to

appear (McNally & Newman, 1999). It is difficult to determine whether research

findings have been influenced by social desirability bias. Some measures incorporate

a social desirability scale, but most do not. It is reasoned in the present research that if

similar findings to other nursing studies investigating levels of work stress, social

support and burnout are found, then it is unlikely that survey results have been

significantly influenced by social desirability.

 125

Choice of Instruments

The essential requirement for reliable and valid measurements has been

addressed by selecting survey instruments which are reported to have adequate

psychometric properties, and which could be expected to give reliable and valid

results for use with nurses. In particular, the choice of a reliable and widely accepted

instrument to measure the study’s criterion variable – burnout, was seen as vital for

the present research, especially since one of the research objectives was to establish

the degree of burnout among a sample of Australian nurses. It is hoped that the

study’s results will provide a benchmark for future investigations into nursing

burnout. The second requirement pertinent to the choice of instruments was a

straightforward administration process, simply worded instructions, and easily

understood survey items. This issue was addressed by choosing instruments that have

been specifically designed to meet these requirements.

Description of Variables Investigated

In Study 2, the researcher used a multi-measure survey to assess nurses’ levels

of job-specific stress, role stress, work support, and burnout. The findings from Study

1 (to be discussed in detail in Chapter 6) influenced the researcher’s decision to select

Wolfgang’s (1988a) HPSI to measure nurses’ job-specific stress. Role stress was

measured by a recognised measure in the occupational stress literature – the

Occupational Roles Questionnaire (Osipow & Spokane, 1987). In addition, the

findings from Study 1 also directed the development of a contextually relevant

measure of work support for hospital nurses. Taking into consideration the sample’s

perceptions of key supportive behaviours received at work, the researcher modified

items from established social support scales in the literature (Shinn, Wong, Simko,

and Ortiz-Torres’ (1989) Supervisor Support Scale, Ray and Miller’s (1994)
 126

Supervisor/Coworker Support Scale, and King, Mattimore, King, and Adams’ (1995)

Family Support Inventory for Workers). Hence, a measure of work support which

assessed emotional and instrumental support from nurses’ supervisors and coworkers,

was developed. Finally, the researcher used the most commonly used measure of

burnout – the MBI (Maslach et al., 1996) to assess emotional exhaustion,

depersonalisation and reduced personal accomplishment in the research program. The

rationale for choosing the following measures and developing an appropriate measure

of work support is detailed below.

Work Stress

The findings from Study 1 (described in detail in the next chapter) identified

work stressors that are common to most Australian nurses. The results also indicated

that nurses are exposed to work stressors that are unique to the health profession

industry, as well as role stressors that are common to most occupations. Therefore, in

the present research program, job-specific stressors as well as generic role stressors

are assessed.

Job-specific stress. In Study 1, the work stressors commonly experienced by

Australian nurses were classified under four broad job stress factors: poor job

conditions, job uncertainty, interpersonal conflict and lack of professional recognition

and support. The most highly reported job stressors included: work overload, lack of

respect and recognition from doctors, dealing with difficult and demanding patients,

dealing with difficult and demanding relatives, and lack of support.

When these major sources of stress were compared to the seven work stress

factors measured by the NSS, both similarities and differences were found. Similar to

Gray-Toft and Anderson (1981a), Study 1 found that work overload, conflict with

physicians and lack of support were major sources of stress for Australian nurses.
 127

However, Australian nurses did not indicate that dealing with dying patients,

inadequate preparation, conflict with other nurses, and uncertainty concerning

treatment, as major sources of stress. These findings led the researcher to conclude

that Gray-Toft and Anderson’s NSS may not be the most appropriate measure of work

stress for Australian nurses.

In an attempt to find a nursing stress scale that more closely aligned to nurses’

work stress, the researcher consulted recent nursing stress literature. Upon closer

examination of the most commonly used nursing stress scales, such as Gray-Toft and

Anderson’s (1981a) NSS, Numerof and Abrams’ (1984) Nursing Stress Inventory,

and Wolfgang’s (1988a) HPSI, it became apparent that there is considerable overlap

among the survey items across the instruments. For instance, all measures include

items concerning workload, patient care responsibilities, interpersonal conflict, and

professional uncertainty. The scale, however, that most closely fit Australian nurses’

perceptions of work stress was the HPSI. This measure was therefore chosen to

measure nurses’ job-specific stress.

The HPSI comprises 30 items that reflect stressful situations frequently

encountered by professionals working in the health care industry. The HPSI provides

a measure of the amount and source of stress experienced by nurses. Respondents

answer how often they find each situation to be stressful in their work setting. A five-

point likert scale is used, ranging from 0 (never/rarely) to 4 (very often). Wolfgang

(1988) suggests that the 30 items be totalled to establish a global measure of job

stress. Total scores therefore range between 0 and 120. Higher scores indicate higher

levels of job stress. The aim of the present research, however, was to examine how

different types of job stress predict different components of burnout. Thus in the

current research, a total score for each subscale was generated. It is believed that
 128

using the HPSI subscales to assess the significant determinants of nurses’ burnout

would provide more meaningful findings than using a global measure of job-specific

stress.

The HPSI appears to possess adequate reliability and validity. It is important

to note, however, that the HPSI has not been commonly used in the nursing stress

literature and few empirical studies have comprehensively explored its psychometric

properties to date. Using data collected from 356 nurses, Wolfgang (1988a) reported

high internal consistency and concurrent validity for the scale. Specifically,

Cronbach’s alpha coefficient for the HPSI was .89, with a split-half reliability of .70.

Pearson’s product-moment correlations between scores on the HPSI and Lyons’

(1971) Tension Index was .78 (p < .001). Gupchup and Wolfgang (1994) also found

that the HPSI was significantly positively correlated to job dissatisfaction (r = .51, p <

.0001) and significantly negatively related to coworker social support (r = -.43, p <

.0001). Factor analysis of the HPSI also found that the 30 items load onto four broad

stress factors: Job Conflicts, Professional Uncertainty, Professional Recognition and

Patient Care Responsibilities (Gupchup & Wolfgang, 1994). Using a sample of 2,267

nurses from across 43 public hospitals in Hong Kong, Akhtar and Lee (2002)

performed confirmatory factor analysis to assess the factor structure of the HPSI.

Their results confirmed the generalisability of Gupchup and Wolfgang’s four factor

structure. The four factors demonstrated adequate coefficient alphas of .78 for

Professional Recognition, .70 for Job Conflicts, .80 for Professional Uncertainty, and

.61 for Patient Care Responsibilities. Professional Recognition, Job Conflicts and

Professional Uncertainty had moderate positive correlations with Emotional

Exhaustion and Depersonalisation, and weak but statistically significant correlations

with Personal Accomplishment.

 129

One of the limitations of the HPSI, however, is that it does not assess stress

associated with working closely with patient’s relatives. The results from Study 1

indicated that dealing with relatives was perceived by nurses to be as stressful as

dealing with the patients themselves. One additional item was therefore added to the

inventory. It was assumed that item 31 – “dealing with difficult and demanding

relatives” would load onto the stress factor – Patient Care Responsibilities, and

therefore would not alter the factor structure of the survey.

Another limitation of the HPSI is that it does not tap into doctors’ lack of

professional recognition and support for nurses. One reason for this is that the HPSI

has been designed for all staff working in the health profession industry, including

doctors. To accommodate this, item 4 – “Not receiving the respect or recognition that

you deserve from the general public” was slightly modified and the words ‘the

general public’ were replaced by the word ‘physicians’. It was expected that this item

would still load onto the stress factor – Professional Recognition and therefore would

not alter the factor structure of the HPSI.

To confirm the overall factor structure of the HPSI had not been altered by the

addition of one item and the modification of another item, factor analysis was

performed in Study 2. Factor analysis of the HPSI also provides further evidence of

the scale’s construct validity and its generalisability to Australian nurses.

Role stress. In Study 1, nurses identified three role stressors (i.e., role over

load, role conflict and role ambiguity) as important sources of stress. As such, for

Study 2, role stressors were measured using Osipow and Spokane’s (1987)

Occupational Stress Inventory (OSI). The OSI, or as it was titled in its first edition -

Measures of Occupational Stress, Strain, and Coping (Osipow and Spokane, 1981),

was designed for use in employed adults working across a diverse range of
 130

occupations. The OSI consists of three separate questionnaires. The Personal Strain

Questionnaire, the Occupational Roles Questionnaire and the Personal Resources

Questionnaire. Only the Occupational Roles Questionnaire (ORQ) was used in the

present study. The ORQ comprises six subscales, three of which were applicable to

this study. The three subscales measure: 1) role overload, 2) role ambiguity, and 3)

role boundary or role conflict as it is defined in the present study. Each subscale

contains 10 items. Item 14 – “I feel I have a stake in the success of my employer” was

slightly modified. To make the item more contextually relevant to nurses, the words

‘my employer’ was replaced with the words ‘this hospital.’ Responses were made on

a 5-point likert scale ranging from never or rarely (1) to most of the time (5). Each

subscale scores in a positive direction, with higher sores indicating higher levels of

stress.

The ORQ demonstrates acceptable reliability. Osipow and Spokane (1981)

using a sample of 31 employed adults reported a test-retest reliability coefficient of

.90 over a two-week interval for the ORQ (sum of scores across all six scales). In

addition, Osipow and Spokane (1983) in a study of 549 working adults of various

occupations, reported a Cronbach’s alpha coefficient of .78 for role ambiguity; .83 for

role overload; and .82 for role conflict. Using a sample of 155 physicians, Alexander

(1983) reported comparable internal consistency coefficients of .81 for role

ambiguity; .76 for role overload; and .77 for role conflict. More recently, Decker and

Borgen (1993) reported alpha coefficients of .79 for role ambiguity, .80 for role

overload, and .75 for role conflict using a sample of 249 adults representing 75

occupations, including hospital staff.

Several studies have provided evidence supporting the validity of the ORQ.

For example, Decker and Borgen (1993) found evidence of the ORQ’s convergent
 131

validity. They reported significant positive correlations between vocational strain and

role overload (r = .33), role ambiguity (r = .48) and role conflict (r = .48). In contrast,

Golec (1983) found evidence of the ORQ’s discriminant validity. Golec (1983)

reported significant negative correlations between job satisfaction and role overload (r

= -.07), role ambiguity (r = -.12), and role conflict (r = -.10). Confirmatory factor

analysis using varimax rotation supported the independence of the three role stress

factors (Osipow & Spokane, 1987), providing further evidence of its construct

validity. Finally, the OSI administration manual – research version (Osipow &

Spokane, 1987) provides normative data based on a sample of 909 adults employed in

130 different occupations. This normative data bank is unique among measures of

role stress.

Work Support

In Study 1, some of the conceptual ambiguity surrounding the social support

construct was overcome by exploring nurses’ perceptions of work support. In Study

1, work support is defined as the emotional and instrumental assistance one receives

through his or her interpersonal relationships at work. Nurses indicated that the two

most salient sources of support at work were their nursing colleagues and their

supervisor. Nurses identified a number of supportive behaviours demonstrated by

their supervisors and their colleagues that could be categorised under two

conceptually distinct categories: emotional support and instrumental support.

In the present study, emotional support refers to the emotional comfort an

individual receives during a stressful situation that leads the person to believe that

they are cared for and valued by others. The most frequently reported emotionally

supportive behaviours identified in Study 1 included: being listened to, being

appreciated, identifying with the problem, and demonstrating respect. These are the
 132

characteristics agreed upon by most authors as defining emotional support (Beehr,

1985; Caplan et al., 1975; Fenlason & Beehr, 1995).

Instrumental support refers to the instrumental assistance that a person

receives as a result of being given the necessary resources (e.g., physical help with a

task) to cope with the stressful situation, or guidance or advice to help solve a

problem. In Study 1, the most commonly reported instrumentally supportive

behaviours included: helping or offering assistance, sharing advice, and problem

solving. These characteristics are also commonly agreed upon by most authors as

defining instrumental support (Beehr, 1995; Caplan et al., 1975; Greenglass et al.,

1990).

Upon close examination it became clear that no one scale presently exists in

the nursing literature which adequately assesses emotional and instrumental support

from work. Nurse researchers have predominantly designed measures of social

support for general and clinical populations. Few of the social support measures have

ever been tested for reliability and validity (Norbeck, 1988). According to Stewart

(1989, 1993), authors of only one-third of the approximately 26 reported social

support measures in the nursing literature describe any psychometric testing. There

are three social support measures however, developed by nurses (Norbeck, 1984;

Tilden, Nelson, & May, 1990; Weinert, 1987; Weinert & Tilden, 1990) that have

undergone extensive psychometric evaluation. These include: the Norbeck Social

Support Questionnaire (NSSQ; Norbeck, Lindsey, & Carrieri, 1981), the Personal

Resources Questionnaire (PRQ; Brandt & Weinert, 1981; Weinert & Brandt, 1987),

and the Tilden Interpersonal Relationship Inventory (IPRI; Tilden et al., 1990).

The NSSQ and the PRQ are designed to measure aspects of relational and

structural network dimensions. The NSSQ is based on the concept of the social
 133

network as a convoy of relationships that travel with an individual throughout the life

span. The NSSQ asks respondents to list 20 social network members and then to

answer nine questions about each. Six of the questions are concerned with functional

properties of social support (e.g., affect, affirmation, and aid). There is one question

about the duration of each relationship, one question on frequency of contact, and a

final question on recent losses of social support. The PRQ consists of two parts. The

first part presents 10 life situations in which a person might need assistance (e.g.,

during sickness) and asks who from a list of support sources (e.g., parent, spouse,

neighbour) would be available to help. The second part is a global measure of

received support based on Weiss’s (1974) conceptualisation of social support:

intimacy, social integration, nurturance, worth, and assistance.

The IPRI differs from the NSSQ and PRQ. It measures not only network

structure and perceived social support, but also reciprocity and interpersonal conflict.

It is based conceptually on social exchange theory. This theory proposes that

interpersonal relationships within social networks depend on reciprocal exchanges of

emotional and tangible assistance. The IPRI consists of several social network items

(size of network, size of household, proximity of relatives) and three separate

subscales to measure types of support, reciprocity and conflict.

It is evident, however, that the measures above do not conceptualise social

support as it has been defined in Study 1. It seems that the above measures are

primarily designed to assess the structural dimensions of social relations (e.g., size,

duration, loss of relationships, frequency of contact of network members), rather than

the functional dimensions of support. The present study required a contextually

relevant measure of social support. Due to the methodological insufficiencies of the

scales frequently used to measure social support in the nursing literature, the
 134

researcher developed a measure of work support specifically for nurses. Items that

closely resembled the nurses’ perceptions of work support (as described in Chapter 6)

were taken from established social support scales although slight modifications were

needed to make them applicable to the nursing context.

In order to assess supervisor support, items were selected from Shinn, Wong,

Simko, and Ortiz-Torres’ (1989) Supervisor Support Scale, Ray and Miller’s (1994)

Supervisor/Coworker Support Scale, and King, Mattimore, King, and Adams’ (1995)

Family Support Inventory for Workers as shown in Table 5.2. For King et al.’s

support scale, the word ‘family’ was replaced by the word ‘supervisor’ and the word

‘home’ replaced by the word ‘work.’ Items 1 to 7 were designed to assess emotional

supervisor support. Items 8 to 12 were designed to assess instrumental supervisor

support.

Table 5.2

Supervisor Support Scale for Nurses

Items Authors

1. My supervisor listens to my problems. Shinn et al., (1989)

2. My supervisor is understanding and sympathetic. Shinn et al., (1989)
3. My supervisor respects me. Ray & Miller (1994)
4. My supervisor appreciates the work I do. Ray & Miller (1994)
5. My supervisor seems to make time for me if I need to discuss my work. King et al. (1995)
6. I feel comfortable asking my supervisor for help if I have a problem. King et al., (1995)
7. When I’m frustrated by some aspect of my work, my supervisor tries to King et al., (1995)
understand.
8. My supervisor will help me figure out a work problem. Shinn et al., (1989)
9. My supervisor cooperates with me to get things done at work. King et al., (1995)
10. If my job duties become very demanding, my supervisor will take on extra King et al., (1995)
work responsibilities.
11. My supervisor can be relied on to help when things get tough at work. Ray & Miller (1994)
12. My supervisor shares ideas or advice with me. Shinn et al., (1989)
Note. Items 1-7 = emotional supervisor support. Items 8-12 = instrumental supervisor support.

The Supervisor Support Scale is designed to measure the level of emotional

and instrumental support nurses’ receive at work. Participants respond on a 5-point

likert scale the degree to which they receive support from their supervisor (1 =

strongly disagree to 5 = strongly agree). Items may be totalled to give a global

 135

measure of supervisor support. The total scores range between 12 to 60 with higher

scores representing higher levels of supervisor support. Alternatively, the items for

the emotional and instrumental support subscales may be aggregated to give a

separate measure of emotional supervisor support and instrumental supervisor

support. As there are seven items comprising the emotional support subscale, total

scores range between 7 and 35. As there are five items on the instrumental support

subscale, total scores range between 5 and 25. Similarly, a higher score represents

higher levels of support.

To assess coworker support, the same items from the supervisor support scale

were used, however the word ‘supervisor’ was replaced by the word ‘coworkers’.

Using the same items for supervisor and coworker support enabled comparisons to be

made between the two support scales. The Coworker Support Scale is also scored in

the same way as the Supervisor Support Scale. Since the psychometric properties of

supervisor and coworker support scale were not known, further analysis to determine

each scale’s reliability and validity was required. In Study 2, separate reliability tests

will be conducted to examine the internal consistency of the supervisor and coworker

support scale (see Chapter 7). The validity of the work support scales will be derived

from two principal sources: (a) factor analysis and (b) correlations with related

measures.

Burnout

Several self-report inventories for the measurement of burnout have been

developed. Possibly the most notable include the MBI (Maslach et al., 1996), the

Burnout Measure (Pines and Aronson, 1988) and the Staff Burnout Scale for Health

Professionals (Jones, 1980). While all three measures are reported to possess

acceptable psychometric properties (Schaufeli et al., 1993), it has been the MBI that
 136

has emerged over the past decade and a half as the most widely used instrument for

measuring burnout (Lee & Ashforth, 1996). Its factorial validity and reliability as a

burnout measure has been demonstrated repeatedly across a range of occupations and

cultures, including Australian teachers (Pierce & Molloy, 1989, Goddard & O’Brien,

2003) and case managers (Goddard, Patton, & Creed, 2000; Patton & Goddard, 2003).

More recently, the MBI was described as “the standard research measure in the

burnout field” (Maslach & Leiter, 1999, p.50).

The MBI differs from other burnout measures as it has three dimensions, and

does not collapse to a unitary score. Although there has been debate about the validity

of the three dimensions in non-human service samples (Garden, 1989; Shirom, 1989),

there is strong empirical support for the validity of the three dimensions in samples of

human service workers. Over the past two decades, three different versions of the

MBI have been developed for different occupational groups. Probably the most well

researched version of the MBI is the Human Services Survey (MBI-HSS) which has

been designed specifically for use with personnel who provide direct services to

human clients. In view of the support for the MBI-HSS in the literature, the current

research program also uses the MBI-HSS to measure burnout. The MBI-HSS is a 22-

item self-report instrument which yields three separate sub-scores reflecting

Emotional Exhaustion (nine items), Depersonalisation (five items), and feelings of

reduced Personal Accomplishment (eight items). Participants rate on a 6-point

response format how often they feel a particular way about their job (Maslach &

Jackson, 1981), with the range being 0 (never) to 6 (every day). The MBI manual

defines high levels of burnout by high scores on Emotional Exhaustion and

Depersonalisation and by low scores on the Personal Accomplishment subscale. Low

levels of burnout are reflected by low scores on the Emotional Exhaustion and
 137

Depersonalisation subscales and by high scores on the Personal Accomplishment

subscale.

Maslach and Jackson (1986) demonstrated adequate internal consistency alpha

coefficients of .90 for Emotional Exhaustion; .79 for Depersonalisation; and .71 for

Personal Accomplishment, as well as test-retest reliabilities over a four-week interval

of .82 for Emotional Exhaustion, .60 for Depersonalisation, and .80 for Personal

Accomplishment. The MBI has also been shown to be valid and discriminate

successfully from other psychological measures in the human services (e.g., Jackson,

et al., 1987; Maslach & Jackson, 1981, 1986). Evidence of convergent validity

included significant correlations between employees’ MBI scores and a) coworkers’

descriptions of employees’ reactions to clients, b) spouses’ descriptions of employees’

behaviours at home, c) case load sizes, and d) amount of time spent in direct contact

with clients/patients. Discriminant validity evidence included low correlations

between MBI scores and job satisfaction, and non-significant correlations with social

desirability scores.

Despite sufficient empirical evidence supporting the psychometric soundness

of the MBI, some authors have reported high correlations between Emotional

Exhaustion and Depersonalisation (e.g. Koeske & Koeske, 1989; Lee & Ashforth,

1990). Results of confirmatory analyses, however, seem to generally support the

existence of three distinct components of burnout (Golembiewski & Munzenrider,

1981; Green & Walkey, 1988; Iwanicki & Schwab, 1981; King & Beehr, 1983;

Maslach & Jackson, 1981, 1986). The original MBI measured both the frequency and

intensity of burnout. Subsequent research using the MBI has revealed that the

intensity and frequency dimensions are highly correlated. Thus, the revised MBI

version includes only the frequency dimension.

 138

Most comparative studies investigating the construct validity of the MBI have

demonstrated that the MBI scales consistently measure the same construct that is

assessed by other burnout measures (Schaufeli et al., 1993; Schaufeli & Van

Dierendonck, 1993). In particular, the emotional exhaustion dimension has been

reported to have the strongest correlation with other self-report measures. The

relationship of the MBI scores to ratings by peers and experts have been described as

less strong (Schaufeli et al., 1993) however, they have been described as satisfactory

in the technical manual (Maslach et al., 1996) and elsewhere (Byrne, 1993, 1994;

Pruessner, Hellhammer, & Kirschbaum, 1999; Taris, Schreurs, & Schaufeli, 1999).

Rationale for Research Method

Study 1

Study 1 incorporated a focus group methodology to probe nurses’ opinions

and views on work stress and work support. Focus groups are defined as “in-depth,

open-ended group discussions that explore a specific set of issues on a predefined and

limited topic” (Robinson, 1999, p.905). They are particularly useful for exploratory

research where rather little is known about the phenomenon of interest (Stewart &

Shamdasani, 1990). In Study 1, two principal research questions and two associated

sub-problems are explored. The principal research questions include:

(1) What are the chronic sources of work stress commonly experienced among

Australian nurses?

(2) To what extent is the way in which nurses’ perceive work support

consistent with how social support is conceptualised and operationalised in

recent empirical literature?

The sub-problems to be examined in this study are associated with principal research

question 1.
 139

(1a) To what extent are the sources of stress experienced by Australian nurses

consistent with the seven stress factors measured by Gray-Toft and

Anderson’s (1981a) NSS?

(1b) Do Australian nurses from various wards share similar perceptions of

work stress?

Focus groups typically consist of between five to eight participants and are

convened under the guidance of a facilitator. The duration of focus group discussions

vary, however one to two hours is not uncommon. In the present study, the researcher

facilitated one hour focus groups in sizes ranging from three to ten nurses. Although

small groups with an unlimited amount of time to discuss issues would have been

preferable, it was not feasible to take nurses off-duty for much longer than an hour.

Focus groups provide a direct method for obtaining rich information within a

social context (Robinson, 1999). Participants are usually a homogenous group of

people who are asked to reflect on a series of questions posed by the facilitator.

Participants have the opportunity to hear other peoples’ responses and are able to

make additional comments as they see fit (Robinson, 1999). It is not necessary for the

group to reach any kind of consensus or to disagree. The primary objective of focus

groups is to obtain accurate data on a limited range of specific issues and within a

social context where people consider their own views in relation to others (Robinson,

1999).

In the present study, the researcher conducted a series of focus groups with

homogenous nursing groups from a range of public hospital wards. Due to the

restricted time available with nurses, the researcher was limited to asking only three

questions. These included:

 140

1) Which particular events within your daily working life cause you the most

stress?

2) Which people within your work environment serve as a support to you

when you are experiencing a stressful day at work?

3) What do these people do that makes you feel supported?

Group interaction is an integral part of this method. Participants are strongly

encouraged to talk to one another, ask questions, exchange anecdotes and comment on

others’ experiences and views (Kitzinger, 1994). This enables complex dimensions to

be revealed that are not accessed by more traditional methods, such as surveys, where

responses are limited. An advantage of focus groups is that they can be used to probe

the underlying assumptions that give rise to particular views and opinions. Not only

are people’s knowledge and experiences explored but also what people think, how

they think, and why they think that way (Kitzinger, 1996). New information, or a

question rephrased within a group, allows the researcher to observe when opinions

shift and under what circumstances. For each focus group, the researcher documented

in a notepad the opinions that were commonly shared by most nurses, and the

opinions that were not shared by the majority of the group.

Another advantage of this method is the natural setting in which the

discussions take place. Unlike quantitative research methods such as surveys,

information is expressed in the nurse’s own words and context without being

constrained by categories. It is also possible for the researcher to observe nonverbal

responses (e.g., gestures, smiles, frowns) which may carry information that

supplements the verbal responses (Stewart & Shamdasani, 1990). The success of this

research approach depends on the quality rather than the quantity of the responses,
 141

which means that the process can be used in various contexts and for a range of

research topics.

Analysis of the focus group data enabled the researcher to identify the

occupational stressors commonly experienced by Australian nurses. The findings

from Study 1 assisted the researcher in choosing an appropriate measure of job-

specific stress and role stress. Furthermore, exploring work support from a nurse’s

perspective provided the researcher with the opportunity to offer a more clearly

defined operationalisation of the social support concept. The findings were used to

develop a contextually relevant measure of work support for nurses. This study is

reported in Chapter 6.

Studies 2 and 3

The present research program is designed to address significant gaps in the

Australian nursing literature in the areas of work stress, burnout, and social support.

The research program’s three main objectives are: 1) to identify key sources and

levels of occupational stress among Australian nurses; 2) to establish the average level

of burnout for a sample of Australian nurses and to compare the current sample’s

level of burnout to other nurses and health professionals and; 3) to identify the main

antecedents of burnout and the effects of social support on the stress-burnout

relationship. The collection of quantitative data was deemed the most appropriate

method to address these objectives.

Study 2 involved the development of a multi-measure questionnaire to assess

the main variables of interest to this research program – job-specific stress, role stress,

work support and burnout. The findings from Study 1 assisted the researcher in

choosing appropriate tools to measure work stress and work support. Since the

validity of the modified HPSI and the work support scales had not been previously
 142

explored, the psychometric properties of these measures were examined. In particular,

the factor structure of the measures was assessed using factor analysis. The HPSI and

work support scales were correlated with other relevant measures (e.g., burnout) using

Pearson’s product-moment correlations to assess their convergent validity. The

internal consistency of the measures was assessed using Cronbach’s coefficient alpha.

Examining the psychometric properties of these scales was a necessary step to ensure

that the findings generated in Study 3 were empirically reliable and valid. Study 2 is

reported in Chapter 7.

The data collected in Study 2 was subjected to further statistical analysis in

Study 3. Specifically, the key sources of stress associated with nursing and the overall

level of work stress for the sample of nurses was established. The participants’ overall

level of burnout across the three dimensions was generated and compared to other

burnout studies in the nursing literature. In addition, the present sample’s level of

burnout was examined in relation to other human service professionals. Furthermore,

the researcher took the opportunity in Study 3 to examine gender differences in work

stress, work support and burnout levels. Finally, the main antecedents of burnout were

explored before examining the influence of social support on burnout.

In Study 3, the main research hypotheses for the program are examined. The

research hypotheses are depicted below.

Hypothesis 1: Nurses’ levels of work stress will not be significantly different

based on gender.

Hypothesis 2: Burnout consists of three related, but empirically distinct

components – emotional exhaustion, depersonalisation, and personal accomplishment.

Hypothesis 3: Australian nurses will report moderate levels of emotional

exhaustion, depersonalisation and reduced personal accomplishment.

 143

Hypothesis 4: Both job-specific stressors and role stressors will be related to

burnout in nurses.

Hypothesis 5: Job-specific stressors will be significantly and positively related

to emotional exhaustion and depersonalisation and significantly negatively related to

personal accomplishment.

Hypothesis 6: Role stressors will be significantly positively related to

emotional exhaustion and depersonalisation and significantly negatively related to

personal accomplishment.

Hypothesis 7: Younger nurses will report higher levels of burnout than older

nurses.

Hypothesis 8: Less experienced nurses will report higher levels of burnout

than more experienced nurses.

Hypothesis 9: Nurses will report higher levels of coworker support than

supervisor support.

Hypothesis 10: Work support will be significantly negatively correlated to the

emotional exhaustion and depersonalisation and significantly positively related to

personal accomplishment.

Hypothesis 11: Work support will be significantly negatively correlated to

work stressors.

Hypothesis 12: Work support will significantly buffer the relationship between

work stress and burnout.

Hypothesis 13: For stressful events deemed to be uncontrollable, emotional

support will have a significant buffering effect on burnout.

Hypothesis 14: For stressful events deemed to be controllable, instrumental

support will have a significant buffering effect on burnout.

 144

The techniques used to analyse the qualitative and quantitative data are present

in the following section.

Overview of Data Analyses

Study 1

Qualitative data analysis involves “describing phenomena, classifying them

and seeing how concepts are interrelated” (Dey, 1993, p.30). In Study 1, the

qualitative data obtained from the focus groups discussions were analysed according

to the principles of content analysis as detailed by Babbie (1995). The core

assumption of content analysis is that words are a window into important features of a

person’s ‘world view.’ Content analysis involves multi-levels of classification

whereby many words or text were classified into fewer content categories (Weber,

1990). The steps taken to analyse the qualitative data will now be outlined.

The first step of the data analysis process involved listening to the taped

discussions and generating an accurate transcript for each focus group. The researcher

also supplemented the transcripts with observational data that was documented during

the discussions. Next, the nurses’ responses to each focus group question were

reduced to fewer content categories. To do this, the researcher read each transcript

several times and recorded the issues that were raised in response to each question.

The researcher then examined the issues for broad overarching themes or

content categories. Conceptual and empirical challenges exist when creating

categories from qualitative data because they must be related to an appropriate

analytic context and supported with empirical evidence (Dey, 1993). In the present

study, the researcher consulted relevant literature and existing work stress and social

support measures to help develop the broad content factors.

 145

Preliminary definitions for the broad categories were then created. A set of

criteria for assigning data to a specific category were gradually established by

generating a list of the key words, sentences, and phrases that underpinned the broad

categories. Within the broad work stress categories, subcategories also emerged.

According to Tesch (1990), categories can be viewed as having fuzzy boundaries

because of the potential overlap involved in classifying comments. Words are

assumed to have multiple meanings and interpretations (Tierney, 1996). The different

levels within these factor structures enable several classification layers to be used

which, according to Dey (1993), may be required if questions are general in nature.

Over time, the preliminary definitions for the broad content categories and the criteria

for classifying the content of the discussion into broad content categories and

subcategories were further refined. To assist in the validation of the research

categories, at least three independent examples for each content category were

produced, as recommend by Berg (1995).

The next step of the content analysis procedure involved what Janis (1965)

referred to as ‘attribution analysis’ in which the frequency with which certain

characteristics (key words, sentences, phrases) were mentioned was tabulated. This

simple counting exercise enabled the researcher to identify the most common sources

of work stress for nurses, as well as the primary sources and types of support nurses

receive at work. To verify that the qualitative data was appropriately categorised by

the researcher, a second researcher also independently coded the transcripts according

to the criteria provided (see Appendix B). Sources of disagreement were identified

and discussed between the researcher and the independent coder before finally

assigning the unit of data to a category. An overall inter-rater reliability coefficient

was established, as well as inter-rater reliability coefficients for the work stress and
 146

work support categories. Calculation of the inter-rater reliability coefficients is

discussed briefly below.

Inter-Rater Reliability

According to Tesch (1990, p.304), “qualitative research is to a large degree an

art” and the critical reader could query whether the researcher has selected only those

fragments of data that support his/her argument (Silverman, 1985). In the present

study, the consistency of the categorisation of data was therefore checked by a second

researcher. The second researcher independently coded the comments according to

the guidelines provided by the researcher. The inter-rater reliability coefficient was

calculated by dividing the number of coding agreements by the number of coding

agreements plus the number of coding disagreements (Goodwin & Goodwin, 1985).

The overall inter-rater reliability estimate for the coding of comments was .88. The

inter-rater reliability coefficient for the work stress categories was .91. The inter-rater

reliability coefficient for sources and types of support categories was .87 and .86

respectively. These results were adequate given the suggestion by Miles and

Huberman (1984) that 70% inter-coder reliability is satisfactory.

Studies 2 and 3

Studies 2 and 3 comprise the quantitative aspect of the research program. All

quantitative data was analysed using the Statistical Package for the Social Sciences

(SPSS) computer program Version 11.0. In Study 2, the psychometric properties of

Wolfgang’s (1988a) HPSI and the work support scales were explored. Specifically,

factor analysis was performed to examine the factorial composition of each scale.

Cronbach’s alpha reliability coefficient was generated to assess the internal

consistency of the scales. In addition, inter-factor and inter-scale correlations were

 147

examined using Pearson’s product-moment correlations to provide further evidence of

the scales’ construct validity.

In Study 3, descriptive statistics for the independent variables and the

dependent variable were generated. The main variables of interest in the current

research program were examined independently, before examining the relationships

between the variables.

To assess nurses’ levels of work stress, the mean scores for the total sample

population, and for males and females, were generated. To identify the work stressors

that cause nurses the most concern, the total mean scores for each item on the HPSI

and ORQ were examined. To assess the level of burnout among Australian nurses, the

total subscale means were generated for the current sample and were then compared

to: 1) the total subscale means for the ‘medicine’ normative sample provided in the

MBI administration manual (Maslach et al., 1996); 2) the total subscale means

provided in international nursing studies; and 3) the total subscale means summarised

for other human service professionals (Schaufeli & Enzmann, 1998). To examine the

prevalence of burnout in the current sample, the percentage of respondents that fell

within Maslach et al.’s definition of high, moderate, and low categories of burnout

were established. Finally, to assess nurses’ levels of coworker and supervisor support,

total mean scores for the sample, as well as for males and females, were generated.

In order to explore the stressor-support matching theory, the controllability of

nurses’ sources of work stress was first established. Since participants who completed

the survey had been informed that one of the main research objectives of Study 2 was

to explore Cutrona and Russell’s optimal matching theory, it was deemed appropriate

to use independent raters to assess the controllability of stressful events. This ensured

that the ratings of controllability were not influenced by familiarity with the research
 148

program’s objectives. Furthermore, independent raters were less likely to be

influenced by the wording of the survey items as they were not privy to the entire

questionnaire. Thus, two independent raters with several years nursing experience

were asked to classify the stress factors as controllable or uncontrollable. The

researcher provided the raters with a definition for each source of job-specific stress

and each source of role stress (see Appendix C). The survey items that loaded onto the

stress factor were also given. A stressor was categorised as uncontrollable if a nurse

could not do anything to prevent the event or lessen its consequences. In contrast, a

stressor was categorised as controllable if a nurse could prevent its occurrence or

consequences.

The proposed buffering effect of social support on the stress-burnout

relationship was examined using hierarchical multiple regressions (Cohen & Cohen,

1983). One regression equation was conducted for each dimension of burnout.

Variables were entered into the hierarchical regression equation based on logical

assumptions. First, it was important to enter the control variables that might have a

prior or distorting effect on the substantive relationships. Sociodemographic factors

were viewed as having temporal precedence, so they were entered first. By entering

the sociodemographic variables as covariates in the first step of the hierarchical

analyses, systematic variance attributable to these peripheral factors could be

removed. To test for main effects, the workplace stress variables were entered next,

followed by the work support variables. At the fourth step, the interaction terms

(stressor x support) were entered. Buffering (i.e., moderating) effects were indicated

by a significant incremental change in the R2 term for the interaction term (Schmieder

& Smith, 1996).

 149

To reduce problems of multicollinearity, the independent variables and the

dependent variables were centred on their means before creating the interaction terms

(Aiken & West, 1991; Aldwin, 1994; Bryk & Raudenbush, 1992). Furthermore, to

reduce the possibility of Type II error, only the sociodemographic factors that were

significantly correlated to the specific burnout dimension were entered at the first step

of the hierarchical regression equation. Similarly, to also reduce the possibility of

Type II errors, only the work stress factors that were found to be significant predictors

of the burnout components were entered at the second step of the hierarchical

regression equation.

Summary

This chapter has presented important methodological issues considered in this

research program. In particular, issues relating to sample selection and the strategies

observed to enhance the reliability and validity of the qualitative and quantitative

findings of the research program were described. This chapter presented the rationale

for the research methods employed for the three studies, as well as the research

questions and hypotheses that comprise the present research program. Finally, an

overview of the data analyses performed for each study are briefly summarised. In the

following chapter the first study that comprises this research program is detailed.
 150

CHAPTER 6

Study 1

A Qualitative Study of Nurses’ Perceptions of Work Stress and Social Support

Overview

The purpose of collecting qualitative data in Study 1 is two-fold. First, the

study seeks to address the lack of published research concerning the chronic sources

of work stress among Australian nurses working in public hospitals. Chronic work

stressors may be defined as “aspects of the work environment that are demanding on

an ongoing and relatively unchanging basis” (Eckenrode, 1984, p.911). Before

investigating the relationship between work stressors and burnout, it was necessary to

clarify what nurses perceive to be the main determinants of stress at work. The thesis

aims to identify work stressors that are commonly experienced by nurses by collating

responses from nurses from a range of public hospital wards. Furthermore, recent

empirical studies have generally used Gray-Toft and Anderson’s (1981a) NSS to

measure nurses’ levels and sources of work stress. However, it is unclear whether the

job stressors experienced by Australian nurses align to the seven occupational stress

factors measured by this scale. In Study 1, a focus group methodology was used to

establish the main sources of work stress for a sample of Australian nurses. Upon

discovering these work stressors, appropriate measures of work stress for Study 2 will

be identified.

Second, the study aims to clarify some of the confusion surrounding the

conceptualisation and operationalisation of the social support construct. Focus groups

will be used to explore nurses’ sources of work support and the types of support they

receive at work. Nurses’ perceptions of work support will be examined in relation to

 151

the way social support is currently conceptualised and measured in the stress-coping

literature.

In the section below, the principal research questions and the sub-problems to

be addressed in Study 1 are outlined. Next, the rationale for utilising focus groups will

be established. The procedure for analysing the content of the data will also be

explained before examining the results of Study 1 and discussing the implications of

the findings.

Principal Research Questions and Associated Sub-Problems

Study 1 incorporates a focus group methodology in order to examine two

principal research questions and two associated sub-problems. The principal research

questions addressed in the present study are:

(1) What are the chronic sources of work stress commonly experienced among

Australian nurses?

(2) To what extent is the way in which nurses’ perceive work support

consistent with how social support is conceptualised and operationalised in

recent empirical literature?

The sub-problems to be examined in this study are associated with principal research

question 1. The researcher took the opportunity to determine:

(1a) To what extent are the sources of stress experienced by Australian nurses

consistent with the seven stress factors measured by Gray-Toft and

Anderson’s (1981a) NSS?

(1b) Do Australian nurses from various wards share similar perceptions of

work stress?
 152

Rationale for Conducting Qualitative Research

Notable gaps in the nursing and social support literature prompted the

researcher’s decision to employ a focus group methodology to collect data on nurses’

perceptions of work-related stressors and support. Stewart and Shamdasani (1990)

suggest that focus groups are particularly useful for exploratory research, particularly

when there is relatively little known about a topic. As a result, focus groups are

generally used very early in a research project and are often followed by other types

of research that provide more quantifiable data from larger groups of respondents.

Rationale for Investigating Nurses’ Work Stressors

There is a lack of empirical studies investigating the chronic sources of stress

among Australian nurses in the nursing literature. Lazarus’ (1993) cognitive-

phenomenological theory of stress suggests that it is the chronic minor stressors (daily

hassles) experienced day after day that are likely to affect health and well-being

(Kanner, Coyne, Schaefer, & Lazarus, 1981; Lazarus & Cohen, 1977; Monat &

Lazarus, 1977). According to the cognitive-phenomenological perspective, if an

individual is exposed to a work stressor for a short period of time, then the stressor

may not cause stress-related outcomes. However, when the individual is exposed to

the same work stressor constantly, day after day, they may experience detrimental

consequences, such as burnout (Carayon, 1995).

As discussed previously in Chapter 1, the Australian public hospital system

has undergone significant changes over the last decade primarily as a result of fiscal

tightening of the health budget (Lumby, 1996). Nurses are being confronted with

increasing patient loads and more acute illnesses. Significant numbers of nursing staff

are leaving the profession or are opting to work part-time as a result of the high level

of stress associated with nursing. Full-time staff are therefore required to work even
 153

harder to compensate for staff shortages. Recent changes to the health care sector

may suggest that nurses have never felt so pressured, and despite this, few empirical

studies have investigated how these changes have impacted on Australian nurses’

perceptions of work-related stress and how these stressors may lead to burnout.

The present study will serve to establish greater clarity with respect to the

chronic sources of stress experienced by Australian nurses working in public

hospitals. In particular, the findings from Study 1 will also provide insight into which

job-specific stressors and which role stressors are considered major sources of stress

for nurses. This was considered a necessary preliminary step before choosing

appropriate scales to measure Australian nurses’ occupational stressors in Study 2.

The findings will also extend current nursing literature by confirming whether

Australian nurses are exposed to the same sources of work stress that have been

identified among nursing studies conducted in other countries.

Rationale for Investigating Social Support

Despite the number of studies that promote the concept of social support as a

buffer to occupational strains such as burnout (Burke & Greenglass, 1995; Capner &

Caltabiano, 1993; Carr et al., 1996; Corrigan, Homes, & Luchins, 1995; Etsion &

Westman, 1994; Greenglass et al., 1996; Koeske & Kelly, 1995; Plante & Bouchard,

1995), other researchers have not found any confirming evidence of the buffering

effect (Burke & Greenglass, 1993; Chay, 1993; Cheuk & Wong, 1995; Dolan &

Renaud, 1992; El-Bassel et al., 1998; Koniarek & Dudek, 1996; Maslanka, 1996).

One explanation for the inconsistent findings regarding the buffering effect of social

support is that authors are still not entirely sure how social support should be defined.

Operationalising social support is difficult because it remains conceptually ambiguous

and therefore multiple meanings have been attributed to it (Ogus, 1990). Researchers
 154

suggest that the various ways used to conceptualise and operationalise social support

have resulted in a range of different measures used to examine the link between social

support and well-being (Beehr et al., 1990). The qualitative data from Study 1 will

enable the researcher to clearly define the social support construct before measuring

social support in Study 2. The findings will also provide important information

concerning nurses’ primary sources of support from work, as well as insight into the

specific types of support that should be considered when examining the influence of

social support on the stress-burnout relationship among nurses.

Method

Subjects

A total of 68 nurses in homogenous groups ranging in size between three and

ten individuals participated in the focus groups. Forty-five nurses (11 males and 34

females) were employed at Hospital A and twenty-three nurses (1 male, 22 females)

were employed at Hospital B. Nurses were recruited from a variety of wards to ensure

that adequate coverage across a range of nursing care roles and patient illness was

obtained. These included: Orthopaedics, Rehabilitation, Renal, Emergency, Intensive

Care, Operating Theatre, and Maternity. A total of 10 focus groups were conducted.

Further details on the selection of the sample were presented in Chapter 5.

Procedure

Ethical approval to facilitate the focus groups was obtained from each

participating hospital and the university. Next, the researcher approached the

Executive Director of Nursing Services from each hospital to gain their consent to

conduct the research and to recruit nurses from specific wards. The researcher then

gained consent from the head nurse from each ward for his/her nursing staff to

participate in the study.

 155

The researcher liaised with the head nurse from each ward to arrange a time to

discuss the research project with the nurses. The researcher personally attended each

ward two weeks before conducting the focus groups in order to inform nurses about

the study and to address their questions and/or concerns related to the study. The

researcher gave information sheets to the head nurse from each ward to circulate to

staff that were not present at the meeting (see Appendix B). The information sheet

detailed the purpose and nature of the study, the time and venue of the discussions,

the conditions of the focus groups, and how the information was going to be used and

stored. The information sheet also clearly stated that participation in the study was

voluntary and that they could remove themselves from the study if at any time they

felt uncomfortable.

Focus groups were conducted in the nurses’ common rooms or staff meeting

rooms immediately after ‘hand-over’ meetings. During hand-over, staff that work

during the day brief night duty staff about each patient’s medical condition,

medication instructions, and treatment plan. The focus groups consisted of both day

and night shift staff. The nurses’ supervisors (e.g., the Central Nurse

Coordinator/Nursing Practice Coordinator) were not invited to attend the focus group

discussions in order to promote an atmosphere of open enquiry and to avoid inhibition

between the group members. The researcher was permitted one hour to conduct the

focus groups. Before commencing the focus groups, the researcher briefly reiterated

the nature and aims of the study, issues concerning confidentiality, and how the focus

group data was to be used and stored. The researcher informed participants that it was

necessary to tape record the session to ensure accurate information was documented

and to aid the data analysis process. It was also explained to participants that brief

notes would be taken throughout the session to assist the researcher in providing
 156

feedback to staff at the end of the session on the main points that were raised in

response to each question. Finally, participants signed a consent form stating that they

understood the nature and conditions of the study (see Appendix B).

Each focus group was asked the following set of questions:

1) Which particular events within your daily working life cause you

the most stress?

2) Which people within your work environment serve as a support to

you when you are experiencing a stressful day at work?

3) What do these people do that makes you feel supported?

The researcher observed that most participants appeared comfortable and

willing to discuss the questions posed by the researcher. Groups comprising six or

more nurses provided the most fruitful discussions. In the larger groups, nurses tended

to generate more ideas and have deeper discussions about the research questions

raised by the researcher.

Results

The results are presented into two sections. First, participants’ comments

regarding nurses’ common sources of work stress are classified. Next, the major

sources and types of support nurses receive at work are classified.

Classification of Comments into Work Stress Categories

To establish the chronic sources of stress common to Australian nurses,

content analysis of the qualitative data was undertaken. This multi-step process is

briefly described. Step 1 identified the range of stressors experienced by nurses (see

Table 6.1). Step 1 involved ‘decontextualising’ the data or more simply, taking

comments out of their original context. It was considered important to examine the

work stressors in their entirety to gain a thorough understanding of the different

 157

demands nurses are exposed to when working in a public hospital. At Step 1, the

researcher also examined the occupational stressors in relation to each ward with the

intention of determining whether (a) nurses from different types of wards have

exposure to similar types of stress, and (b) whether nurses have similar perceptions on

major sources of work stress.

At Step 2, broad overarching themes or factors that most aptly described

nurses’ sources of work stress were depicted. Step 2 involved ‘recontextualising’ the

data by aligning specific work stressors to particular work stress categories. (see

Table 6.2).

Step 3 involved identifying the major work stress subcategories that

underpinned the broad work stress categories (see Table 6.3). Specific work stressors

were assigned to particular work stress subcategories. During Step 3, the major

stressors experienced by nurses at work were determined. This was achieved by

tabulating the frequency with which work sub-categories were mentioned. In cases

where the work stress subcategory was reported infrequently (i.e., less than five

times), the work stress subcategory was removed. It was considered that stressors that

were rarely mentioned were more likely to be a source of stress for a smaller number

of participants or for a specific ward. In the present study, however, the researcher

was interested in exploring work stressors that were commonly experienced by most

nurses. Work stressors that were mentioned by the nurses more than five times were

deemed by the researcher to be a reasonable source of work stress.

At Step 4, the main sources of stress for Australian nurses were compared to

the seven stress factors assessed using Gray-Toft and Anderson’s (1981a) NSS (see

Table 6.4). Both similarities and differences in the work stress factors were noted.
 158

At Step 5, typical phrases or statements that best capture the major sources of

work stress experienced by the sample of Australian nurses are provided. In the

following section, each step will now be presented.

Step 1: Range of Work Stressors Identified by Australian Nurses

The work stressors identified by the sample of public hospital nurses are

depicted in Table 6.1.

Table 6.1

Sources of Stress for a Sample of Australian Nurses

• An excessive workload • Lack of support

• Disrespectful doctors • Ethical dilemmas (e.g., prolonging a patient’s life
• Demanding patients under all circumstances)
• Abusive/violent patients • Inability of doctors to make decisions
• Difficult/demanding relatives • Doctors not being available for consultation
• Demanding doctors • Too much administration/documentation
• Unfamiliar medical problems • Lack of communication between departments
• Constant interruptions • Dealing with multiple pressures at the one time
• Lack of experienced staff • Shortage of medical equipment
• Inadequate staffing • Insufficient wages
• Shortage of beds • No meal/toilet breaks
• Shortage of space • Noise (e.g., alarms)
• Fear of litigation • Being responsible for running the department as
• Heavy patient load well as the patients’ well-being
• Poor rostering schedule • Computers breaking down
• Patient crises • Being expected to do things you haven’t done
• Trying to meet the needs of different before
people (patients, relatives, doctors,
health professionals)

When the work stressors were examined in relation to each ward, it appeared

that some job stressors were limited to specific wards. For instance, nurses from the

Maternity Ward expressed concern about the general public’s increasing propensity to

threaten legal action. Nurses from the Intensive Care Unit (ICU) indicated that they

found the constant noise from alarms on machines stressful. ICU nurses also

confronted a serious ethical dilemma when caring for critically ill patients –‘quality of

life’ versus ‘quantity of life.’ ICU nurses indicated that it was sometimes stressful
 159

trying to prolong a patient’s life if the patient’s quality of life would be severely

reduced. Nurses from the Renal Unit indicated that there were not enough dialysis

machines to cater for the increasing number of patients and that it was stressful having

to keep very sick patients waiting for treatment. The Emergency Department reported

that the shortage of beds and the lack of space to move was a constant source of

frustration.

Overall, however, there did appear to be a range of stressors that were

common to most nurses across all wards. Most nurses reported that they were exposed

to excessive workloads, constant interruptions, staff shortages, especially lack of

experienced staff, lack of support from, and conflict with, doctors as well as dealing

with difficult and demanding patients and their families. This suggests that there are

some work stressors that are experienced by most nurses. Finally, the data also

suggested that nurses from the Emergency Department recorded more sources of

work stress than the other wards. The Emergency Department services the public 24

hours per day. Nurses have to contend with unexpected rises in patient loads and may

be faced with medical problems they have not been exposed to previously. During

peak periods, cubicles become full and nurses have to care for patients in corridors.

Delays in admitting patients to other hospital wards increase the demands on nurses

by generating activities outside their normal nursing duties. They also contend with

abusive relatives who are frustrated by delays in admission.

In summary, the focus groups revealed that the demands placed upon nurses in

public hospitals are numerous and varied. Nurses have to contend with a high

workload with limited available resources. They are expected to perform their jobs

whilst managing conflict with doctors, patients and their families. They are
 160

responsible for the smooth running of the ward and yet they have little support or

guidance from other health professionals.

Overall, the findings indicate that nurses are exposed to similar work stressors.

The focus group discussions revealed that although the sample nurses worked across a

variety of wards, there were several work stressors that were common to all nurses.

This may suggest that there are more similarities among nurses working on different

wards than differences.

Step 2: Broad Work Stress Categories and Associated Stressors

Analysis of the data suggested that the majority of work stressors could be

classified under four overarching categories (see Table 6.2). The four broad work

stress categories identified were: 1) job conditions, 2) job uncertainty, 3) interpersonal

conflict, and 4) lack of professional recognition and support. Job conditions may be

described as the demands associated with being in a nursing role (e.g., work overload,

role conflict, role ambiguity) as well as the conditions associated with working as a

nurse in a public hospital setting (e.g., limited resources). Job uncertainty takes into

consideration unexpected stressful events that are often beyond a nurse’s control (e.g.,

patient load, doctors not being available to make decisions, medical crises,

malfunctioning medical equipment). Interpersonal conflict involves conflicts that

arise from working so closely with patients and their relatives, as well as doctors, all

in times of acute stress conditions. Lack of professional recognition and support

refers to the undervaluing of nurses’ skills, experience, and qualifications by other

health professionals, especially doctors, and the insufficient provision of support.

 161

Table 6.2

Classification of Australian Nurses’ Sources of Work Stress

Work Stress Category Work Stressors

Job Conditions • Excessive workload

• Heavy patient load
Demands associated with being in • Too much documentation
a nursing role, as well as the • Constant interruptions
conditions associated with • No meal/toilet breaks
working as a nurse in a hospital • Staff shortages (including shortage of experienced staff)
environment.
• Being responsible for running the department, as well as
the patients’ well-being
• Dealing with multiple pressures at the one time
• Trying to meet the needs of different people.
• Being expected to do things you haven’t done before
• Shortage of beds, space and medical equipment
• Noise

Job Uncertainty • Patient load

• Patient crises
Unexpected stressful events that • Breakdown of equipment
are often beyond a nurse’s • Rostering schedule
control. • Unfamiliar problems (e.g., medical conditions, operating
new technology)
• Doctors not being available for consultation
• Inability of doctors to make decisions
• Lack of communication between departments
• Fear of litigation

Interpersonal Conflict • Demanding patients

• Abusive/violent patients
Conflicts that arise from working • Difficult and/or demanding relatives
so closely with patients and their • Demanding doctors
relatives, as well as doctors, all in
times of acute stress conditions.

Lack of Professional Recognition • Disrespectful doctors

and Support • Lack of support
• Insufficient wages
The undervaluing of nurse’s
skills, experience and
qualifications by other health
professionals, especially doctors,
and the insufficient provision of
support.
 162

Step 3: Classification of Work Stress Subcategories

Following Step 2, the nurses work stressors were collapsed further into

subcategories (see Table 6.3). In Step 3, the frequency of comments, the percentage of

comments, and the inter-rater reliability coefficient were tabulated for each work

stress subcategory. In cases where the work stress subcategory was reported

infrequently (i.e., less than five times), the subcategory was removed.

For job conditions, resource shortages (e.g., beds, space, and medical

equipment) as well as noise on the ward were removed. For job uncertainty, fear of

litigation and patient crises were removed. For interpersonal conflict, difficult and

demanding doctors were removed. For lack of respect and recognition, insufficient

wages were removed. Only work stress subcategories that were mentioned by nurses

more than 5 times were deemed by the researcher to be common sources of work

stress for nurses.

In Table 6.3, the most commonly reported work stressors for Australian nurses

have been highlighted. The key work stressors are: 1) work overload, 2) lack of

respect and recognition from doctors, 3) difficult/demanding patients, 4)

difficult/demanding relatives, and 5) lack of support.

 163

Table 6.3

Main Sources of Work Stress for a Sample of Australian Nurses

f % IR

Main Sources of Work Stress

Job Conditions
Work Overload (e.g., staff shortages, too much documentation, interruptions, 63 17.52 0.92
too many patients, too many responsibilities).
Role Conflict (e.g., dealing with multiple pressures at the one time) 8 3.42 0.80
Role Ambiguity (e.g., being expected to do things you haven’t done before) 6 2.56 0.75
Job Uncertainty
Unpredictability (e.g., rostering, patient load, interruptions, treatment mistakes) 18 7.69 1.00
Unfamiliar Problems (e.g., medical conditions, operating new technology) 10 4.27 0.91
Doctor Availability 6 2.56 0.83
Interpersonal Conflict
Difficult/Demanding Patients (e.g., abusive, violent, dependant) 27 11.54 1.00
Difficult/Demanding Relatives 27 11.54 0.86
Lack of Professional Recognition and Support
Lack of Respect and Recognition from Doctors 45 11.97 1.00
Lack of Support 24 10.26 1.00

Total 234 100 .91

Note. f = frequency of comments; % = percentage of comments; IR = inter-rater reliability coefficient.

Step 4: Comparing Australian Nurses’ Sources of Stress to the NSS

At Step 4, the main sources of stress for the current sample are compared to the

seven factors measured by Gray-Toft and Anderson’s (1981a) NSS (see Table 6.4).

Some similarities and notable differences are identified. Similar to Gray-Toft and

Anderson, participants recognised that workload, conflict with physicians, and lack of

support were major sources of stress for nurses. However, Australian nurses did not

recognise death and dying, inadequate preparation, conflict with other nurses, and

uncertainty concerning treatment as chronic sources of work stress.

 164

Table 6.4

Australian Nurses’ Major Sources of Stress Compared to the NSS

Seven Sources of Stress from the NSS Five Sources of Stress for Australian Nurses

1) Death and dying 1) Work overload

2) Conflict with physicians 2) Lack of respect and recognition from
3) Inadequate preparation doctors
4) Lack of support 3) Difficult/demanding patients
5) Conflict with other nurses 4) Difficult/demanding relatives
6) Work load 5) Lack of support
7) Uncertainty concerning treatment

Step 5: Description of Australian Nurses’ Main Sources of Work Stress

In the section below, typical statements underpinning the major sources of

stress for the current sample are detailed.

Work overload. Work overload may result from the employee being assigned

too many tasks or insufficient time to accomplish the assigned tasks (Van Sell et al.,

1981). Work overload was the highest predictor of work stress and accounted for

approximately 17% of comments regarding prominent sources of work stress. Work

overload was mentioned substantially more than any other stressor. Each ward that

participated in the focus groups identified work overload as a significant source of

stress. Participant’s reported that their workload was excessive as a result of high

patient loads, other health professionals (e.g., doctors) readily off-loading work onto

nurses, constant interruptions, and increased paper work and time pressures. A

typical submission covered a number of points as depicted below:

• There are too many patients to handle during a shift. We manage 10-20 patients per
shift.
• Experienced staff are overworked because they are expected to carry the load of new
staff as well as their own.
 165

• We don’t have enough staff to cope with the workload.

• Other staff (e.g., doctors) off-load work and extra responsibilities onto to us. They
often ask us to do extra things.
• We have to be able to handle a number of different things at once. There’s not enough
time to do everything that is required of us. When attending to 10 to 20 patients,
there’s not enough time to complete all the documentation and paper work during our
shift. We often have to work longer than eight hours on a shift because we can’t get
everything done that is expected of us. We are faced with constant interruptions -
telephone calls, other staff asking us to do things for them, dealing with relatives’
enquiries. We can’t even get a meal break or go to the toilet during our shift. We just
can’t stretch ourselves any further.

The excessive workload endured by nursing staff appears to be strongly linked to

a shortage of staff to manage organisational demands. Specifically, participants

indicated that there is a lack of experienced staff on each shift, especially night duty.

Experienced nurses are expected to familiarise new staff to the ward and to ensure

that they are performing their job correctly at the same time as performing their

normal duties. Some of the key phrases were:

• Our patient loads may increase, but there is no change in staff levels.

• We are short-staffed, especially after-hours.

• You just can’t readily call in extra staff on night duty.

• We need more staff. There are not enough people to cover you if you need a break.

• There is not enough experienced staff on each shift. Sometimes half the staff are

agency staff who don’t know the patients or the procedures on the ward. It just makes

the day so much harder for regular staff when we have to take on extra work as well

as orientate them to the ward.

Lack of respect and recognition from doctors. According to the nursing literature,

doctors and nurses have a collaborative working relationship (Craven & Hirnle,
 166

2000). This is because nurses are required to work closely with doctors and assist

them by ensuring that the treatment plan directed by the doctor is implemented and

followed correctly. Participants reported, however, that they felt that some doctors

did not treat them with respect and did not recognise their skills, experience, or

qualifications. Lack of respect and recognition from doctors accounted for

approximately 12% of comments raised by nurses. Some participants indicated that

there is a sense of inequality between nurses and doctors. Although they were

expected to be respectful towards doctors they indicated that doctors were not

expected to demonstrate respect towards nurses. Nurses indicated that doctors

sometimes behave in a superior manner and abuse the power associated with their

position. The way in which doctors demonstrated their disrespect and lack of

recognition for nurses is depicted below.

• Doctors like to feel like they’re in charge. They barge in and say “I want this done
now.” They expect you to do what they want immediately. They want you to drop
everything no matter what you’re doing.
• They command respect. They think they’re God!
• Their tone of voice makes you feel incompetent – “Well, why don’t you know?” They
don’t listen to you. They just bark orders at you. They ‘chew your pants off’ if you
don’t document everything, yet if they do something wrong, they don’t want to hear
about it from nurses. Instead they blame us for their mistakes!
• They don’t recognise our skills, our qualifications, or experience.
• They couldn’t care less about us. They couldn’t give a toss.

Difficult/demanding patients. A nurse’s role is to assist patients in meeting their

health care needs. They help the patient to adapt to their situation, to cope, to work

through feelings and concerns, and to solve problems. Nurses suggested, however,

that some patients are too dependent on them and are not willing to take responsibility

for their own improvement in health. Other patients are difficult to manage because
 167

they are abusive or violent, especially those with personality or psychiatric disorders.

Difficult and/or demanding patients are a major source of stress for nurses. They

account for approximately 11% of the responses related to work stress. Typical

statements regarding patients are outlined below.

• We have to give attention to those who are acutely ill, as well as attend to the needs
of all the other patients.
• Some patients engage in attention-seeking behaviour. They always want your
attention.
• Patients with psychiatric problems are very difficult to care for and manage.
• Certain patients stress you out, especially intoxicated or abusive patients. They
demand to be seen straight away and then they blame you when they’re not. They give
us a mouthful and blame everything on us. They’re always ready to ‘have a go’ at the
nurse.
• They expect nurses to have all the answers. They expect nurses to be doctors!

Difficult/demanding relatives. Nurses also work closely with patients’ relatives.

They inform the relatives about the patient’s illness and discuss with them the

patient’s treatment plan. It would appear that interacting with relatives is equally as

difficult and demanding as working with patients. Approximately 11% of the

participants’ responses related to the stress associated with dealing with relatives.

Nurses reported that relatives can sometimes be demanding because they have very

little understanding or knowledge of the nurse’s role and as a result they may have

unrealistic expectations of nurses. In addition, nurses are regularly exposed to families

in which there is conflict or where there are long standing feuds between family

members. They suggested that managing these ‘dysfunctional’ families was stressful.

Some families are difficult to manage because they refuse to follow the nurses’

advice, instructions, or hospital rules. Some of the stress associated with interacting

with patients’ relatives is outlined below:

 168

• Some relatives have unrealistic expectations about the patient’s extent of recovery or
speed of recovery. If the patient isn’t improving as fast as they thought, they often
blame us for being incompetent and not doing enough. We have to constantly explain
the patient’s condition to them but they don’t seem to listen, or don’t want to listen.
They don’t comprehend that we’re here to help.
• Some relatives have outlandish expectations of nurses. They treat you like a servant.
• Some families are very dysfunctional and can unsettle the patient. We constantly have
to readjust our care to deal with family conflict.
• Relatives believe they are entitled to see the patient at any time they like and for as
long as they like, regardless of visiting hours.
• Families can be manipulative. They try to challenge your boundaries. They often try
to catch you out or they will play one nurse off another.

Lack of support. Nurses are required to be a source of support for patients and

their families, however there is not enough support for nurses at work to help them

cope with everyday demands. Nurses indicated that due to shortages in nursing staff,

there was very limited support available to nurses during their shift, especially night

duty. Inexperienced staff appeared to feel less supported than experienced staff. Less

experienced nurses suggested that there is very little support for new staff because

everyone is so busy. Typical statements reflecting nurses’ perceptions of an

unsupportive work environment are reflected below:

• When you’re new, you don’t get that support. It’s difficult to try and get someone to

help when everyone is so busy and you’re still learning the ropes.

• We have to operate new medical equipment, but there is little supervision or guidance

to help us operate the machines and to deal with any technical mishaps. Whoever put

the new machinery in isn’t there, and nobody knows how to use it.

• There is a lack of support from medical staff. They don’t give us any encouragement,

help, or support. They don’t care.

 169

• We don’t have the staff to be able to pull them from other areas to come and help us

when we need it. We don’t have any support.

• Other nursing staff can be unsupportive. If you ask them a question, they’ll pull faces

at you and humiliate you.

• On night duty there’s no-one to call. There’s no avenue to out.

• We don’t have a big support system. There’s no time to debrief and nobody cares.

In summary, it would appear that fiscal tightening in the health care sector, as

discussed in Chapter 1, has had a substantial impact on the working conditions of

Australian nurses. Nurses are confronting ever-increasing workloads without

adequate resources (e.g., staff) and workplace support to contend with these demands.

Furthermore, patients and relatives readily project their fears and frustrations on

nurses by being aggressive and verbally abusive. Nurses are being asked to do more

with less and it is little wonder that many nurses are opting to work part-time or leave

the profession. Despite the enormous pressures associated with nursing, doctors do

not seem to work as a team with nurses. Instead, doctors appear to off-load further

responsibilities onto staff. The above findings provide sufficient evidence that nursing

is a stressful occupation. In the following section, participants’ comments regarding

social support are classified.

Classification of Comments into Social Support Categories

Sources of Support

The qualitative data obtained from the focus group discussions was used to

determine whether nurses’ perceptions of workplace support were consistent with the

way in which social support has been conceptualised in recent occupational stress

literature. The process for identifying the major sources of support from work
 170

involved a one-step process. The nurses’ key supportive persons were classified into

broad categories and the frequency for each source of support was calculated. The

major sources of support were deemed to be those that were mentioned by

participants ten times or more.

Identification of the Major Sources of Work Support

Analysis of the data suggested that the supportive persons available to

Australian nurses in their work environment could be classified under four

overarching categories. The four categories were: 1) colleagues, 2) supervisor, 3)

health professionals and 4) non-health professionals. The frequency of comments, the

percentage of comments, and the inter-rater reliability coefficients were tabulated for

each category. In Table 6.5, the major sources of support for Australian nurses have

been highlighted. The key sources of support for public hospital nurses are their

colleagues, and their supervisor.

Table 6.5

Sources of Workplace Support for a Sample of Australian Nurses

Sources of Workplace Support f % IR

Colleagues 30 49.19 .98

Supervisor (e.g., CNC/NPC) 15 24.59 .90
Health Professionals (i.e., doctors, staff counsellors, staff 8 13.11 .81
educators, physiotherapists)
Non-Health Professionals (i.e., ward persons, patients, 8 13.11 .79
relatives)

Total 61 100 .87

Note. f = frequency of comments; % = percentage of comments; IR = inter-rater reliability coefficient.

CNC = Central Nurse Coordinator; NPC = Nursing Practice Coordinator

Table 6.5 suggests that Australian nurses have access to a number of sources

support at work. For the most part, however, nurses indicated that their primary

source of support at work was their nursing colleagues. Coworker support was
 171

reported three times more frequently than supervisor support and accounted for 49%

of the comments. Support from their supervisor, namely their CNC or NPC,

accounted for approximately 25% of the comments. Although accounting for a

smaller percentage of the comments, other health professionals that were considered

sources of support were doctors, physiotherapists, staff development educators, and

staff counsellors. Non-health professionals (ward persons, patients and their relatives)

were also perceived to be supportive.

Nurses indicated that they often received support from their coworkers without

requesting it. They said their coworkers related well to the pressures and problems

they face at work because they are regularly exposed to the same stressful situations.

Sometimes, however, they sought support from their supervisor or other hospital staff,

someone a little more removed from the stressful situation because they could offer

another perspective or solution to the stressful situation. The researcher observed that

when the participants were discussing the support they receive from their supervisor,

it was primarily in relation to instrumental support (e.g., seeking advice). In contrast,

when the nurses were discussing the support they receive from their coworkers, it

primarily related to emotional support (listening and empathising with the problem).

This may indicate that different work sources fulfil different support needs for the

stressed individual. It is therefore reasoned that different sources of support should be

considered when examining the stressor-support matching model.

Types of Support

Similar to classifying the data for work stress, a multi-step process was used to

categorise the types or functions of social support. Step 1 involved identifying the

range of ways in which nurses felt supported in their work environment. Step 2

consisted of categorising the various support functions into broad categories. The
 172

specific supportive behaviours were aligned to the key types of support. Step 3

involved identifying the main supportive behaviours by calculating the frequency in

which the supportive behaviours were mentioned. In cases where the supportive

behaviour was mentioned infrequently (i.e., less than ten times), the supportive

behaviour was not recorded. Supportive behaviours that were mentioned ten times or

more were deemed by the researcher to be the most common types of support nurses

receive at work. At Step 4, typical phrases or statements are provided to help illustrate

the nurses’ perceptions of how support is demonstrated at work.

Step 1: Identification of Supportive Behaviours

Analysis of the qualitative data revealed that nurses felt supported in a range

of different ways. Specific acts of support are depicted in Table 6.6.

Table 6.6

A Range of Supportive Behaviours Identified by a Sample of Australian Nurses

• Listening to problems or concerns • Helping somebody lift or manoeuvre a

• Reassuring the person that they are doing physically challenging patient
their job well • Helping during busy periods
• Showing appreciation when somebody • Getting advice to help solve a problem
does something for you. • Offering assistance
• Identifying with or understanding the • Sharing nursing experiences
problem • Sharing knowledge
• Recognising and appreciating an • Having social get-togethers outside of work
individual’s efforts • Helping somebody with their patient load
• Discussing problems or concerns • Sharing a joke with each other
• Trusting another person’s judgement or
decision

The number of different types of work support perceived by participants

seems to suggest that social support is a multidimensional construct. Given the range

of supportive behaviours that are associated with social support, it is not surprising

that the definitions of social support have been inconsistent, diverse, vague and

sometimes contradictory. As Vaux (1988) noted, “people assist each other in an

 173

astonishing variety of ways” (p.17), and most of these have been labelled social

support at one time or another.

Step 2: Classification of the Social Support Dimensions

As shown in Table 6.7, the supportive behaviours were classified into two

broad categories – emotional support and instrumental support. Emotional support

refers to the emotional comfort an individual receives during a stressful situation that

leads the person to believe that they are cared for and valued by others. Instrumental

support refers to the instrumental assistance that a person receives as a result of being

given the necessary resources (e.g., physical help with a task) to cope with a stressful

situation, or the guidance or advice to help solve a problem. Therefore work support

may be defined in the present study as the emotional and instrumental assistance an

individual receives through his or her interpersonal relationships at work.

Table 6.7

Classification of the Social Support Dimensions

Types of Social Support Supportive Behaviours

Emotional Support (ES) • Listening

• Identifying/understanding the problem
ES refers to the emotional comfort an (empathy)
individual receives during a stressful situation • Respect
that leads the person to believe that they are • Appreciation
cared for and valued by others. • Reassurance
• Laughing together

Instrumental Support (IS)

• Help or assistance
IS refers to the instrumental assistance that a • Getting advice
person receives as a result of being given the • Sharing knowledge/experiences
necessary resources to cope with a stressful
situation, or the guidance or advice to help
solve a problem.
 174

Step 3: Identification of the Main Types of Support for Nurses

The frequency of comments, the percentage of comments, and the inter-rater

reliability coefficients were tabulated for each type of support category (see Table

6.8). Supportive behaviours that were mentioned less than five times were removed.

Two of these emotionally supportive behaviours were removed - reassurance and

laughing together. No instrumentally supportive behaviours were removed. In Table

6.8, the key supportive behaviours identified by Australian nurses have been

highlighted. The two most frequently reported supportive behaviours are listening and

helping or offering assistance.

Table 6.8

Main Types of Work Support for a Sample of Australian Nurses

Type of Support f % IR

Emotional Support
Listening 19 24.07 .91
Showing Appreciation 12 15.19 .83
Identify/Understanding the Problem 11 13.92 .81
Respect 6 7.59 .75
Instrumental Support
Helping or Offering Assistance 19 24.05 .90
Sharing Information 6 7.59 .84
Problem Solving 6 7.59 .96

Total 79 100 .86

Note. f = frequency of comments; % = percentage of comments; IR = inter-rater reliability coefficient.

Step 4: Description of the Main Types of Support for Nurses

In the next section, typical statements underpinning the major types of support

are detailed.

Emotional support. In the present study, emotional support refers to the

emotional comfort an individual receives during a stressful situation. Receiving

 175

emotional support leads the person to believe that they are cared for and valued by

others. Emotional support included specific behaviours such as being listened to,

being appreciated, identifying with the problem, and demonstrating respect.

Interestingly, 60% of the comments regarding the functions of support were related to

emotional support. This may suggest that nurses are able to recognise emotionally

supportive behaviours more readily than instrumentally supportive behaviours.

Alternatively, nurses’ may view emotional support to be more important than

instrumental support. Typical statements reflecting nurses’ perceptions of emotional

support are outlined below.

• Knowing that you can vent your frustrations and they will completely understand

where you are coming from. They identify with you. People outside of here can’t do

that.

• They listen to you. Having someone to listen to you whinge, bitch and moan when

you’ve had a bad day.

• You can say things that are considered socially unacceptable and they won’t bat an

eyelid.

• We appreciate one another. We take care of each other and we try to look out for

each other.

• We have a lot of respect for each other. We try to let our colleagues know that they

are supported.

Instrumental support. Instrumental support refers to the instrumental assistance

that a person receives as a result of being given the necessary resources to cope with a

stressful situation, or the guidance or advice to help solve a problem. Instrumental

support included behaviours such as helping or offering assistance, sharing

knowledge, and helping a person to rectify the problem. Nurses’ perceptions of

instrumental assistance are detailed below.

 176

• If we can, we try and help each other out. They don’t even need to ask. We just step in

and give each other a hand.

• When things get tough for someone, we chip in and help. We might just do a few

things for them and assist them where we can.

• Sometimes, someone will take over for you and give you an exit. Sometimes you just

need that when it all gets too much.

• When you don’t know what you’re doing and someone shows you what to do or talks

you through it step by step, it’s really helpful.

• When someone helps you fix a problem, or gives you some advice so you can solve

the problem yourself.

In summary, participants identified two major types of support nurses’ receive at

work – emotional support and instrumental support. These types of support are

consistent with current social support literature. In the following section, the overall

findings will be considered in relation to Study 1’s principal research questions and

sub-problems.

Discussion

The qualitative data from Study 1 was able to provide responses to the two

principal research questions. First, the data obtained from the focus groups provided

valuable insight into the stressful conditions Australian nurses regularly confront

working in a public hospital setting. The findings revealed that nurses are exposed to

both job-specific stressors and generic role stressors. It could therefore be reasoned

that both job-specific and generic role stressors should be explored simultaneously

when examining nurses’ work stress.

Second, the study confirmed that the way in which nurses perceive work

support is consistent with current social support literature. More specifically, nurses

reported that their coworkers, followed by their supervisor, were their main sources of
 177

support at work. Furthermore, nurses discussed work support from a multidimensional

perspective, recognising various support components that could be classified broadly

as emotional and instrumental support. In the section below, the main sources of

stress for Australian nurses are delineated.

Chronic Sources of Stress Among Australian Nurses

In line with its image as a stressful occupation, considerable attention has been

paid to pinpointing specifically intrinsic sources of stress within the nursing

profession (Kirkcaldy & Martin, 2000). Recent empirical literature (e.g., Janssen, De

Jonge, et al., 1999; Janssen, Schaufeli, et al., 1999; Kennedy & Grey, 1997; Kirkcaldy

& Martin, 2000; Wheeler & Riding, 1994) has established ‘dealing with death’ and

‘workload’ as particularly potent sources of stress for hospital nurses. An excessive

workload has featured prominently among the leading occupational stressors for

nurses, together with patient care, relationships with colleagues, and bureaucratic

constraints.

Similarly, Study 1 also sought to identify the chronic sources of stress

commonly experienced by Australian nurses. Previous empirical studies investigating

work stress among nurses have generally relied upon Gray-Toft and Anderson’s

(1981a) NSS. Since this tool was validated using American nurses at a private

hospital, it was considered important to clarify whether the working conditions and

the associated demands experienced by Australian nurses from public hospitals were

similar to American nurses from private hospitals.

In the present study, nurses’ work stressors were classified according to four

broad stress categories. These included: 1) job conditions; 2) job uncertainty; 3)

interpersonal conflict; and 4) lack of professional recognition and support. Five key

work stressors commonly experienced among Australian nurses were identified.

 178

These included: 1) work overload; 2) lack of respect and recognition from doctors; 3)

difficult/demanding patients; 4) difficult/demanding relatives; and 5) lack of support.

Although examining job stressors according to specific areas of nursing clinical

practice has been of appeal to researchers (e.g., Keane, Ducette, & Adler, 1985;

Sullivan, 1993), the findings from Study 1 seem to suggest that the similarities among

nurses working in different wards are more apparent than the differences between

them. Thus, the current thesis confines its study to work stressors experienced by the

majority of nurses.

Work overload was clearly the most reported source of stress for Australian

nurses. This finding is consistent with other Australian empirical studies (Bryant et

al., 2000; Healy & McKay, 2000) as well as international empirical studies (e.g.,

Edwards et al., 2000a, 2000b; Guppy & Gutteridge, 1991; Hillhouse & Adler, 1997;

Landsbergis, 1988; Wheeler, 1998; Wheeler & Riding, 1994). The nursing sample

indicated that work loads were increasing due to staff shortages. Regular staff

indicated that they were being ‘stretched to their limits’ as they took on heavier

workloads. Nurses reported that there was a shortage of experienced staff to cover

shifts, especially night duty. According to Lumby (1996), shortages in experienced

staff is due in part to an ageing nursing population in which older, more experienced

staff, are leaving the profession or are choosing to work part-time.

Work overload resonated most acutely with experienced staff. Experienced

nurses spoke in depth about the number of tasks they are expected to perform during a

shift. It seems that the number of responsibilities given to experienced staff has

grown disproportionately in recent years. Not only are they expected to perform their

regular duties but they are also required to familiarise new staff (e.g., first year nurses

and agency nursing staff) to the ward. It is their responsibility to teach less
 179

experienced nurses new skills and procedures, oversee that they are performing their

tasks competently, and take on extra duties to compensate for their inability to

perform higher-level duties.

Despite heavier workloads and inadequate staffing, there seems to be limited

support available to nurses to contend with these organisational demands.

Inexperienced staff reported feeling unsupported at work because they are not always

able to ask for assistance or talk through their concerns with more experienced nurses

as they are already overloaded with work. Experienced staff felt unsupported as there

is nobody to assist them when their workload is excessive and unmanageable.

Furthermore, nurses indicated that there was insufficient technical support available to

them. For example, when new technological equipment is installed on the wards there

is often no-one to provide guidance and assistance to operate the machines properly.

Interestingly, this finding is not consistent with Healy and McKay’s (2000) study in

which lack of support was the least reported stressor among Australian nurses. One

possible reason for this is that nurses in the present study were more open to

disclosing that they required more support because the purpose of the study was to

investigate work support. Alternatively, it could be that the nurses who participated in

the study were not typical of the wider nursing population. Participants in Study 1

may have felt more stressed about the available support at work than their nursing

counterparts.

Another source of stress for Australian nurses was the lack of respect and

professional recognition demonstrated by some doctors. Nurses spoke of being

undervalued and being made to feel inferior as doctors issued orders and appeared to

question their abilities. Lumby (1996) also found in her study of Australian nurses

that doctors did not provide support to nurses and that verbal abuse by some doctors
 180

was increasing. Similarly, Stehle (1981) also found the working relationship between

nurses and doctors to be a major source of stress for nurses.

As well as contending with disrespectful and unsupportive doctors, nurses also

have to manage difficult and demanding patients and their relatives. Participants

reported that some patients and their relatives have unrealistic expectations of nurses.

They noted that patients seem reluctant to take responsibility for their own healing

process and relatives blame nurses when a patient’s health does not improve as

quickly as they would expect. In addition, nurses have to contend with abusive

patients who become frustrated when they unable to be seen by a doctor immediately.

According to Cherniss (1995), regularly confronting patients who do not follow

advice, make impossible demands, resist change, and who sometimes lie, cheat and

manipulate is an everyday reality for health professionals. In contrast, Bryant et al.

(2000) reported that Australian nurses did not perceive patients and relatives as

significant contributors to their stress at work. In their study, many expressed positive

statements about their roles and the nurse-patient relationship. Wheeler (1994) and

Riding and Wheeler (1995a, 1995b), however, have also found pressure from patients

to be a major stress factor for nurses.

Overall, it seems that the role of the Australian nurse is changing. Some of

these changes appear to be a direct result of cut-backs in the government’s funding of

hospitals. The behaviours expected of nurses are being extended, whilst

simultaneously the amount of support available to help nurses cope with these

increasing demands is diminishing. In addition, it seems that patients and their

relatives are not receiving the level of attention they desire and as a result, they are

taking out their frustration on nurses by being abusive and aggressive. According to

Laschinger, Finegan, Shamian, and Almost (2001), many nurses currently working in
 181

hospitals are in conflict. They are often faced with the need to provide a much greater

volume and intensity of care than they are able to deliver, yet they remain responsible

and accountable as professionals for practicing within the standards of the profession.

Many of the resources supporting nursing practice, such as nurse educators and unit

administrators with nursing backgrounds, have been reduced or eliminated. Not

surprisingly, many nurses believe that these changes have led to an erosion of patient

care, decreased job satisfaction and burnout among nurses (Laschinger, Finegan,

Shamian, & Wilk, 2001). The current study suggests that for Australian nurses, it is

not the work itself, rather the working conditions that cause stress. Similarly, Adomat

and Killingworth (1994) concluded that stressors in nursing were more related to

“…the organisational pressures of the environment in which care was delivered,

rather than the involvement of caring for critically ill patients” (p. 912).

Australian Nurses’ Sources of Stress Compared to the NSS

The major sources of stress identified by the current nursing sample were

compared to the seven sources of stress measured by the NSS (Gray-Toft &

Anderson, 1981a). Both similarities and notable differences were found. Similar to the

present study’s findings, Gray-Toft and Anderson also recognised that workload was

a significant source of stress for nurses. The NSS measures work load using five

items. These stressful events include: the breakdown of a computer, unpredictable

rostering, staff shortages, too many non-nursing tasks such as clerical work, and not

enough time to provide emotional support to patients. Although they fit under the

rubric of workload, these work stressors are somewhat different from Australian

nurses’ perceptions of workload. Gray-Toft and Anderson primarily examined

workload from a qualitative perspective which takes into consideration job

complexity, time constraints, and resources. In the present study, workload is

 182

discussed from both a qualitative and quantitative perspective, with a greater focus on

quantitative workload.

In the current sample, nurses reported that their workload was excessive as a

result of high patient loads, other health professionals (e.g., doctors) readily off-

loading work onto nurses, constant interruptions, and increased paper work.

Participants indicated that the organisational demands placed upon them were

growing and many nurses worked longer than the specified eight-hour shift to

complete the work expected of them. It seems that the nursing tasks themselves are

not indicative of stress. Instead, it is the magnitude of tasks combined with the lack of

time to complete the tasks that is stressful. Australian nurses perceive workload not

only in terms of the actual tasks they perform, but also in terms of the amount of the

work they are expected to perform in a given time.

It is evident that public hospitals in Australia are expecting nurses to

accommodate additional work responsibilities at a time in which resources are already

stretched. As national employment of nurses continues to fall (AIHW, 1999), patient

admissions continue to rise. In 1998-99, there was a 2.3% increase in patient

admissions. With an ageing population, the number of elderly and acutely ill patients

has increased and the length of hospital stays has decreased. Furthermore, hospital

beds have been reduced which has led to overcrowding in the waiting rooms as

patients wait to see a doctor. All these factors suggest that the role of a nurse is

changing and it would appear unlikely that their workload is going to improve in the

near future. Taking this into consideration, it could be argued that it is important to

examine not only workload, but also role overload in Study 2.

Gray-Toft and Anderson also recognised poor relationships with doctors as a

major determinant of stress among nurses. The NSS measures ‘conflict with doctors’
 183

using four items. Items examine criticism by a physician, conflict with a physician,

fear of making a mistake in treating a patient, and disagreement concerning the

treatment of a patient when the physician is unavailable. In the current study, nurses

also commented on the way in which some doctors undervalued their work and

treated them with a lack of respect. Nurses reported that they spend more time with

patients than doctors, yet doctors were not prepared to listen to their advice on

treating patients if it conflicted with their own opinion. Nurses indicated that they felt

more responsible for a patient’s welfare than a doctor because if a mistake is made,

they can be blamed for a doctor’s error. Nurses stated that they double-check

medication dosages because they fear mistakes could be made in treating a patient.

The NSS also recognises lack of support as a source of stress for nurses. There

are three items that address nurses’ lack of opportunity to be able to talk openly to

other unit personnel about problems, to share experiences and feelings, and to express

negative feelings toward patients. A limitation of this subscale, however, is that the

items primarily tap into emotional support. In the present study, nurses also found the

inability to be able to talk openly to other nurses due to a lack of time to be stressful,

but they also revealed that not being able to ask for extra help or assistance (i.e.,

instrumental support), was also stressful.

In the current study, the qualitative data revealed that staff shortages were a

prominent source of stress for Australian nurses with participants stating that there is

not enough staff to adequately cover a unit. Nurses suggested that despite their

increasing workload, nursing staff levels have remained the same, or in some cases,

decreased. Gray-Toft and Anderson also identified shortages in staff as a potential

source of nurse stress. On the NSS, lack of staff support is measured using one item

on the workload subscale.

 184

A major limitation of Gray-Toft and Anderson’s NSS is that it does not

measure the stress related to dealing with difficult and demanding patients and their

relatives. The majority of items concerning patients on the NSS are associated with

treating a dying patient, and other death related issues. Several studies have also

found that nurses perceive treating a dying patient as a major source of stress (e.g.,

Adomat & Killingworth, 1994; Farrington, 1995; Hingley & Cooper, 1986; Hipwell et

al., 1989; Numerof & Abrams, 1984). In the current study, however, nurses did not

indicate that this was a prominent work stressor. When the researcher enquired more

about this, participants reported that they knew that treating dying patients was an

inevitable aspect of nursing and that they were adequately prepared to cope with this

aspect of the job. According to Peeters and Le Blanc (2001), stressors that are

considered typical to the job (e.g., treating critically ill patients) are appraised as the

least significant. Apparently, employees expect that some stressors are indissolubly

connected with their profession and, as a result, they do not perceive them to be very

significant. In addition, there were two items addressing the patient’s family on the

NSS. They include feeling inadequately prepared to help with the emotional needs of

the patient’s family, and not knowing what a patient’s family should be told about the

patient’s condition and treatment. The items on the NSS do not adequately capture the

stress related to the problematic behaviours (e.g., difficult, demanding, abusive,

aggressive) and unrealistic expectations of patients and their relatives.

Although there are similarities in the types of work stressors Australian nurses

and American nurses experience, there are also some notable differences. Both

recognise that workload, conflict with physicians, and lack of support are important

sources of stress. For Australian nurses, staff shortages, and dealing with difficult and

demanding patients and their relatives are also common sources of stress. Unlike
 185

American nurses, nurses working in Australian public hospitals did not identify

inadequate preparation, conflict with other nurses, and uncertainty concerning

treatment as major sources of stress. These findings would seem to suggest that Gray-

Toft and Anderson’s NSS may not be the most appropriate measure of Australian

nurses’ work stress. This prompted the researcher to consider alternative measures of

nurses’ work stress. In the section below, the measures chosen by the researcher to

investigate nurses’ work stressors in Study 2 are briefly discussed.

Measures of Nurses’ Work Stress

The findings from Study 1 revealed that nurses are exposed to both job-

specific stressors and role stressors. The four broad work stress categories (i.e., job

conditions, job uncertainty, lack of professional recognition and interpersonal

conflict) identified in the current study would appear to align more closely to the four

factors measured by Wolfgang’s (1988a) HPSI: job conflicts, professional

uncertainty, professional recognition and patient care responsibilities. As discussed

earlier in Chapter 5, the HPSI is a 30 item survey designed to measure perceptions of

potentially stressful job situations in health care professionals, including nurses. The

HPSI’s depiction of stressful events commonly experienced among health care

professionals is similar to the stressful events identified by the sample of nurses in the

present study. For instance, examination of the HPSI reveals that items relating to the

work stressor - work overload underpin the factor ‘job conflicts.’ The HPSI assesses

quantitative and qualitative aspects of work overload. The stressful events measuring

work overload include: shortages in staff, having too much work to do, interruptions

by phone calls and people, and supervising the performance of coworkers.

The work stressors – lack of respect and recognition from doctors and lack of

support appear to align to the HPSI factor ‘professional recognition.’ Stressful events
 186

include: not receiving the respect and recognition you deserve, not being allowed to

participate in decision making, not utilising your abilities to their full extent, not

having the opportunities to share feelings and experiences with colleagues, not

receiving adequate feedback on job performance, and insufficient pay.

Dealing with difficult/demanding patients is adequately assessed by the HPSI

factor ‘patient care responsibilities.’ Stressful events include: dealing with difficult

patients, trying to meet society’s expectations for high quality medical care, caring for

the emotional needs of patients, feeling ultimately responsible for patient outcomes,

and caring for terminally ill patients. A limitation of the HPSI and most commonly

used nurse stress scales, such as Gray-Toft and Anderson’s (1981a) NSS and

Numerof and Abrams’ (1984) Nurse Stress Inventory, is that they do not examine the

stress associated with dealing with difficult/demanding relatives. For the purpose of

the present research, the item “Dealing with difficult/demanding relatives” was added

to the HPSI as it was reasoned that the structure of the factor - ‘patient care

responsibilities’ would not be significantly altered. Since the items on the HPSI

appeared to closely resemble the current sample’s perceptions of work stress, the

HPSI was administered to nurses in Study 2.

Finally, the findings also provided evidence that the nursing profession is

characterised by role stress. Although mentioned less frequently than the job-specific

stressors, nurses also identified chronic generic stressors that could be classified as

role overload, role conflict and role ambiguity. The inclusion of Osipow and

Spokane’s (1987) Occupational Roles Questionnaire (ORQ) to measure role stress in

Study 2 was therefore deemed appropriate. The psychometric properties and

administration details of the ORQ are discussed earlier in Chapter 5.

 187

Conceptualisation of Social Support

A primary objective of Study 1 was to reduce some of the confusion

surrounding the support construct by examining the way in which nurses perceive the

support they receive at work. The data obtained from the focus groups was used to

reconcile some of the inconsistencies in the way support has been defined and

operationalised in earlier literature.

In the present study, support was considered from two perspectives: sources of

support and types of support. The findings were consistent with most studies

investigating work support in that support from supervisors and colleagues were

identified as the two most salient sources of support for nurses. Most studies suggest

that supervisor support is the most important because the supervisor is the only person

who has the position of power to alter or change the work situation at hand (Peeters &

Le Blanc, 2001). However, in the current study, nurses identified their coworkers as

the most salient source of support. According to Peeters and Le Blanc (2001), nurses

are more likely to share their experiences with other colleagues whom they know also

encounter similar situations on a regular basis. An alternative explanation could be

that nurses turn to their coworkers because they work in close proximity to each other

and therefore their colleagues are easily accessible.

The researcher observed that nurses spoke primarily about emotional support

when discussing coworker support and instrumental support when discussing

supervisor support. This observation provides some validation for the stress-support

matching theory. The stressor-support matching theory proposes that specific events

elicit particular salient coping requirements and that social support is more likely to

buffer the negative effects of stress, if available support is able to address the needs of

the situation (Cohen & Wills, 1985). Thus in the present thesis, both supervisor and
 188

coworker support will be investigated separately when considering the effects of

social support on the stress-burnout relationship.

The study’s findings also suggested that global conceptualisations of social

support (e.g., Cohen & McKay, 1984; Heller, 1979; Heller & Swindle, 1983;

Shumaker & Brownell, 1984) should be abandoned in favour of more precise models

that depict specific support components. Nurses recognised that work support served a

multitude of functions. In the present study, it was deemed useful to organise these

various functions into two basic categories - emotional support and instrumental

support. Other empirical studies support this view (Fenlason & Beehr, 1994; Beehr,

1995). The nurses identified behavioural components of emotional support including

listening, showing appreciation, identifying with the problem and demonstrating

respect. The specific acts associated with instrumental support included helping or

offering assistance, sharing advice, and problem-solving. Whilst it is recognised that

there may be other types of supportive behaviours, the functions of support identified

in the current study are considered salient to nurses. Furthermore, the supportive

behaviours generated in the current study were consistent with other researchers’

conceptualisations of emotional and instrumental support. For example, Beehr (1995)

defined emotional support as the provision of sympathy, evidence of liking, caring

and listening. House (1981) identified instrumental support as behaviours that directly

assist another individual in need such as taking on part of a colleague’s workload.

The two most frequently reported supportive behaviours were listening and

helping or offering assistance.

 189

Pines, Aronson, and Kafrey (1980, p.125) also recognised the importance of

listening for individuals working in stressful occupations. They state:

Everyone has occasions when they need one or more people who will actively listen

to them without giving advice or making judgements. They need someone with

whom they can share the joys of success as well as the pain and frustration of failure.

They need someone they can share conflicts as well as everyday trivial incidents.

People working in a stressful occupation occasionally need to let off steam. A good

active listener listens with understanding and sympathy.

Beehr (1995) proposed that tangible aid, such as helping or offering

assistance, is most effective for individuals where there is high workload. He argued

that helping stressed people to get their job tasks done at a time when they are

overloaded will help relieve the stressor (work overload). Nurses in the present study

were also exposed to excessive workloads. It is therefore not surprising that these

nurses perceived the receipt of help or assistance with tasks as an important type of

support. It could be argued that instrumental support will be the most effective type of

support in reducing nurses’ workloads.

In recent literature, several studies incorporate House’s (1981)

conceptualisation of social support in which support is broken into four main

components: emotional, instrumental, appraisal and informational support. Some

studies (e.g., Cutrona & Russell, 1990; Wills & Shinar, 2000), however, have found

sizeable correlations among these types of support, suggesting significant overlap

among the dimensions. It was reasoned that since emotional support can be clearly

distinguished from instrumental support, classifying social support into two categories

may be more useful. Furthermore, distinguishing between four different types of

support could further complicate analyses and could result in unreliable findings.
 190

In summary, the findings suggest that while nurses are at work, it is their

supervisors and their colleagues who are able to provide important support functions.

The study also confirms that social support is a multidimensional construct. Taking

this into consideration, a measure of work support that distinguishes between different

sources and types of support will be used in Study 2 to examine the influence of

support on work stress and burnout.

Limitations

Finally, while the focus groups provided some interesting results, this study is

not without limitations. First, the findings from the focus groups may be open to some

bias. Although the researcher attempted to gain information from each nurse who

participated in the focus groups, there were times in which a few members of the

group tended to dominate the discussion. The dominating members may have

influenced the opinions of other nurses in the group. Those nurses who felt their

opinions were in the minority may have been reluctant to express their views.

Furthermore, some nurses may have felt uncomfortable sharing their views with other

members in the group and some may have been concerned about confidentiality

between participants. Finally, since only a small number of nurses from each hospital

ward participated in this study, the results cannot be generalised to the wider nursing

population without further replication. In the following chapter, the psychometric

properties of the HPSI and the work support scales are examined in more detail.
 191

CHAPTER 7

Study 2

An Assessment of the Psychometrics Properties of the HPSI and the Work Support Scales

Study 2 utilises a survey methodology to measure the main variables of interest in this

research program. The main objective of Study 2 is to provide evidence of the psychometric

soundness of the scale used to measure nurses’ job-specific stress (i.e., the HPSI) and the

scales designed to assess nurses’ work support. Verifying that these scales are

psychometrically sound is a necessary step before examining the variables further in Study 3.

In this chapter, the rationale for investigating the psychometric properties of these scales is

provided. Next, the development of the quantitative survey used in the present study to

measure nurses’ levels of job-specific stress, role stress, support at work, and burnout is

briefly described. Following this, the psychometric properties of the job-specific stress

measure and the work support scale are examined.

Rationale for Study 2

In the present study, a multi-measure questionnaire was constructed to assess work

stress, work support and burnout in a sample of Australian nurses working in a public

hospital setting (see Appendix A). The qualitative findings in Study 1 were used as a basis

for choosing the scales to measure nurses’ work stress. In particular, the HPSI (Wolfgang,

1988a) was chosen to measure sources of job-specific stress. This was because the four broad

stress factors identified by the sample of Australian nurses appeared to align closely to the

four factors measured by the HPSI. Use of the HPSI, however, was not without limitations.

One item was added to the inventory to assess “dealing with difficult and demanding

relatives.” Furthermore, one item was slightly adapted to tap into doctors’ lack of
 192

professional recognition and support for nurses. The ORQ (Osipow & Spokane, 1987) was

chosen to measure sources of role stress as the scale demonstrates adequate psychometric

properties, it has been frequently used in the occupational stress literature, and it provides

normative data for a large sample of adults across a broad range of occupations.

The qualitative data from Study 1 also provided insight into Australian nurses’

perceptions of support received from work. The nursing sample identified two primary

sources of support at work, their nursing colleagues and their supervisor (e.g., CNC or NPC).

The supportive behaviours provided by their coworkers and their supervisor were broadly

classified as emotional and instrumental support. The findings enabled the researcher to

clearly conceptualise the social support construct before compiling a concise, contextually

relevant measure of work support for nurses.

Lastly, the MBI (Maslach et al., 1996) was chosen to measure nurses’ levels of

emotional exhaustion, depersonalisation, and reduced personal accomplishment. The MBI is

the most frequently used measure of burnout in the occupational strain literature, enabling

comparisons between studies to be easily made. Furthermore, satisfactory levels of reliability

and validity have been shown (e.g., Maslach et al., 1996). Normative data for a variety of

human service occupations is also provided in the MBI administration manual.

While there is sufficient evidence in the empirical literature demonstrating that the

ORQ and MBI possess adequate psychometric properties (see Maslach et al., 1996; Osipow

& Spokane, 1987; Schaufeli & Enzmann, 1998), less is known about the HPSI and the work

support scales developed for this research program. It was therefore considered appropriate

to provide some evidence of their construct validity and reliability. This is demonstrated

using a number of statistical procedures.

 193

To assess the factor composition of these scales, the statistical technique known as

factor analysis is conducted. Factor analysis is deemed a suitable technique for determining

whether the HPSI’s structure has changed as a result of the modifications that were made to

the inventory for the purpose of this research. Factor analysis is also a useful procedure for

establishing whether the supervisor and coworker support scales adequately discriminate

between emotional and instrumental support. Furthermore, inter-factor correlations as well as

inter-scale correlations are generated for each scale. To estimate the internal consistency of

each scale, Cronbach’s coefficient alpha is calculated. Clarifying the psychometric soundness

of these measures ensures that any findings of statistical significance in Study 3 are not the

result of errors in measurement.

Method
Subjects

Participants were drawn from three public hospitals. A total of 273 nursing

professionals (235 females, 38 males) completed the survey. The response rate for the

questionnaire was 67.74%. Further information regarding the sociodemographic

characteristics of the sample can be found in Chapter 5.

Materials

In Study 2, a survey was developed to examine the key variables of interest in this

research program, that is, work stress, social support, and burnout (see Appendix A). The

survey incorporated measures of work stress that assessed stressful events that are considered

to be characteristic of the nursing industry, as well as generic stressors that are often

associated with white-collar occupations (i.e., role overload, role conflict, role ambiguity).

Social support was measured from within the workplace, namely coworker and supervisor

support. The coworker support and supervisor support scales contained both emotional and
 194

instrumental support items. The burnout instrument tapped into three components - emotional

exhaustion, depersonalisation, and reduced personal accomplishment.

Work Stress

One of the central purposes of Study 1 was to identify the major sources of stress that

are common to most Australian nurses. The findings indicated that nurses are exposed to

several work stressors that are unique to the health profession industry, as well as role

stressors that are common to most occupations. In order to establish which work stressors are

the most influential determinants of burnout, it was considered important to measure work

stressors that are specific to nursing, as well as generic work stressors.

Job-specific stress. In the present study, nurses’ job-specific stress is measured using

the HPSI (Wolfgang, 1988a). The HPSI comprises 30 items that reflect stressful situations

frequently encountered by professionals working in the health care industry. The items on the

HPSI have been found to load onto four broad stress factors: job conflict, professional

uncertainty, professional recognition and patient care responsibilities (Akhtar & Lee, 2002;

Gupchup & Wolfgang, 1994). The HPSI provides a measure of the amount and sources of

stress experienced by health professionals. Respondents answer how often they find each

situation to be stressful in their work setting. A five-point likert scale is used ranging from 0

(never/rarely) to 4 (very often). In the present study, a global measure of job stress is

generated by summing the scores for each item. Item scores are also totalled for the four

subscales. Higher scores represent higher levels of stress.

For the purpose of the current research, item 4 on the HPSI was slightly modified to

examine doctor’s lack of respect and recognition for nurses. The item - “Not receiving the

respect or recognition that you deserve from the general public” - was modified with the
 195

words ‘the general public’ being replaced by the word ‘physicians.’ It was expected that this

item would still load onto the stress factor – professional recognition and therefore would not

alter the factor structure of the HPSI. Furthermore, one extra item was added to the HPSI to

measure the stress associated with difficult and demanding relatives. It was assumed that the

additional item – “dealing with difficult and demanding relatives” would load onto the stress

factor - patient care responsibilities, and therefore would not alter the factor structure of the

survey. Further details regarding the rationale for choosing the HPSI and its psychometric

properties are provided in Chapter 5.

Role stress. Role stress is assessed using one measure from Osipow and Spokane’s

(1981, 1987) OSI. The OSI consists of three separate questionnaires. The Personal Strain

Questionnaire, the Personal Resources Questionnaire and the ORQ. Only the ORQ was used

in the present study. The ORQ comprises six subscales, three of which were applicable to

this study. The three subscales measured: 1) role overload, 2) role ambiguity, and 3) role

boundary or role conflict as it is defined in the present study. Each subscale contains 10

items. Responses are made on a 5-point likert scale ranging from 1 (never or rarely) to 5

(most of the time). Each subscale scores in a positive direction, with higher sores indicating

higher levels of stress. Further details regarding the psychometric properties of the ORQ are

presented in Chapter 5.

Work Support

In the nursing literature, no one scale exists that adequately measures work support as

it is operationalised in the present study. This prompted the researcher to construct a measure

of supervisor support and coworker support that is relevant to nurses by adapting previous

existing measures. The work support scales were designed to measure nurses’ perceptions of
 196

the emotional and instrumental assistance they receive from their supervisor and their

coworkers. The Supervisor Support Scale consists of seven emotional support items and five

instrumental support items. The items taken from established social support scales aligned to

the key supportive behaviours identified by the nurses in Study 1. Items were selected from

Shinn et al.’s (1989) Supervisor Support Scale, Ray and Miller’s (1994)

Supervisor/Coworker Support Scale, and King et al.’s (1995) Family Support Inventory for

Workers. The items selected from the Family Support Inventory for Workers required some

modifications. Specifically, the word ‘family’ was replaced by the word ‘supervisor’ and the

word ‘home’ replaced by the word ‘work.’ The Coworker Support Scale contained the same

items as the supervisor support scale however, the word ‘supervisor’ was replaced by the

word ‘coworkers.’ Responses are made on a 5-point likert scale ranging from 1 (strongly

disagree) to 5 (strongly agree). Items are totalled to give a global measure of supervisor

support and a global measure of coworker support. As there are 12 items, the total scores

range between 12 and 60 with higher scores representing higher levels of support. The items

comprising the emotional and instrumental support subscales are also aggregated to give a

more specific measure of emotional and instrumental support. As there are seven items

comprising the emotional support subscale, total scores range between 7 and 35. As there are

five items on the instrumental support subscale, total scores range between 5 and 25. For

further details regarding the work support items and their corresponding authors, see Table

5.2 in Chapter 5.
 197

Burnout

Burnout is measured by the MBI-HSS (Maslach et al., 1996), The MBI-HSS is a 22-

item self-report instrument which yields three separate sub-scores reflecting emotional

exhaustion (nine items), depersonalisation (five items), and feelings of reduced personal

accomplishment (eight items). Participants rate on a 6-point response format how often they

feel a particular way about their job, with the range being never (0) to every day (6). The

MBI manual defines high levels of burnout by high scores on emotional exhaustion and

depersonalisation and by low scores on the personal accomplishment subscale. Low levels of

burnout are reflected by low scores on the emotional exhaustion and depersonalisation

subscales and by high scores on the personal accomplishment subscale. Further details

regarding the psychometric properties of the MBI are presented in Chapter 5.

Procedure

Permission was granted from the hospitals’ ethics committee to administer the survey

to ward nurses. The Executive Director of Nursing Services from each hospital provided a

list of the wards in which potential participants could be recruited. The researcher contacted

the head nurse from each ward to gain their consent for nurses to be involved in the study.

Following this, the researcher arranged with the head nurse to meet with staff to disseminate

information about the study. At the introductory meeting, an overview of the research

program was given and the study’s main objectives were explained. Nurses were informed

about survey confidentiality and anonymity. They were also made aware that participation

was voluntary and that they could withdraw from the study at any time. They were advised

that the questionnaire would take approximately 30 minutes to complete. In addition, an

 198

information sheet clearly detailing the nature of the study and the proposed time and venue

for administering the survey in their ward was distributed (see Appendix A).

Questionnaires were administered to nurses directly after staff hand-over meetings.

The nursing supervisor from each ward was not invited to attend the meeting because items

on the survey related to supervisor support. To improve chances of obtaining an adequate

response rate and to address participants’ questions throughout the survey, the researcher

stayed with respondents while they completed the surveys. The researcher attached a

chocolate to each survey as a token of her appreciation for staff taking the time to complete

the survey. Once respondents had completed the questionnaire, they were instructed to place

their survey in the envelope provided and to return it in person to the researcher. Additional

questionnaires were also left in common rooms for nurses who were unable to attend the

meeting but had expressed interest in participating in the study. A self-addressed, reply-paid

envelope accompanied the questionnaire. In some cases, the researcher arranged with nurses

to return a week later to collect the completed questionnaires.

Data Analysis

In the present study, data was analysed by the Statistical Package for the Social

Sciences (SPSS) Version 11.0. The psychometric properties of the HPSI and the work

support scales are examined using the following statistical techniques. Specifically, factor

analysis is performed to examine the factorial composition of each scale. Cronbach’s alpha

reliability coefficient is generated to assess the internal consistency of the scales. In addition,

inter-factor and inter-scale correlations are examined using Pearson’s product-moment

correlations to provide further evidence of the scales’ construct validity.

 199

Results

Factor structure of the HPSI

Factor analysis of the HPSI was conducted using principal axis factoring. Based on

previous studies (Akhtar & Lee, 2002; Eells et al., 1994; Gupchup & Wolfgang, 1994), it was

expected that a four-factor solution would be identified. It was therefore specified that four

factors were to be extracted with a loading cut of .30. The factors were rotated to a simple

structure by employing both orthogonal (VARIMAX) and oblique (OBLIMIN) rotations.

While the orthogonal and oblique solutions were very similar, with only one item loading

differently, the oblique solution was judged to be most appropriate. This was because the

loadings in the oblique solution showed a clearer pattern than the orthogonal solution, which

showed more cross-loadings. Table 7.1 delineates the four factor structure of the HPSI which

was obtained using principal axis factor analysis followed by oblique (oblimin) rotations.
 200

Table 7.1

Pattern Matrix for Oblique Factors of the HPSI

Factors and Items Factors

I II III IV
I. Lack of Professional Recognition and Support
25. Not being allowed to participate in making decisions about .69
your job.
29. Not being able to use your abilities to the fullest extent on the .65
job.
20. Not receiving adequate feedback on your job performance. .60
26. Not being challenged by your work. .59
13. Feeling that opportunities for advancement on the job are .58
poor.
17. Not being recognised or accepted as a true health professional .54
by other health professionals.
27. Feeling that you are inadequately paid as a health .47
professional.
4. Not receiving the respect or recognition that you deserve from .43
physicians.
8. Not having opportunities to share feelings and experiences .42
with colleagues.
22. Having non-health professionals determine the way you must .36
practice your profession.
_
23. Not knowing what type of job performance is expected.
II. Patient Care Uncertainty
5. Being uncertain about what to tell a patient or their family -.72
about the patient’s condition and/or treatment.
31. Dealing with difficult/demanding relatives. -.65
16. Dealing with difficult patients. -.56
28. Caring for terminally ill patients. -.53
11. Allowing personal feelings or emotions to interfere with the -.50
care of patients.
18. Being inadequately prepared to meet the needs of patients. -.49 .41
19. Possessing inadequate information regarding a patient’s -.46
medical condition.
30. Fearing that a mistake will be made in the treatment of a -.44
patient.
7. Disagreeing with other health professionals concerning the -.43
treatment of a patient.
III. Job Conditions
24. Being interrupted by phone calls or people while performing .62
job duties.
14. Trying to meet society’s expectations for high quality medical .56
care.
3. Feeling ultimately responsible for patient outcomes. .55
15. Supervising the performance of less experienced workers. .51
6. Caring for the emotional needs of patients. .50
1. Having so much work to do so that everything cannot be done .49 .38
well.
21. Not having enough staff to adequately provide necessary .49
services.
12. Keeping up with new developments in order to maintain .37
professional competence.
 201

Table 7.1 (cont)

IV. Interpersonal Conflict
_
10. Having job duties which conflict with family responsibilities.
2. Experiencing conflicts with supervisors or administrators. .43
9. Experiencing conflicts with coworkers .36
Note. Item numbers are from the original questionnaire. Professional Recognition: eigenvalue = 8.22, total

variance explained = 26.52%. Professional Uncertainty: eigenvalue = 2.48, total variance explained = 8%. Job

Expectations: eigenvalue = 1.82, total variance explained = 5.88%. Interpersonal Conflict: eigenvalue = 1.49,

total variance explained = 4.8%.

In the present study, the four-factor solution explained 45.2 % of the total variance.

Items loading onto Factor 1, accounting for 26.52% of the total variability, pertain to the

absence of opportunities for nurses to use their abilities fully and to contribute to important

decisions concerning their job, and the lack of recognition and support provided to nurses by

other health professionals. It was therefore labelled Lack of Professional Recognition and

Support. Previous studies have similarly labelled this factor Professional Recognition

(Akhtar & Lee, 2002; Gupchup & Wolfgang, 1994) or Lack of Perceived Enrichment

Potential (Eells et al., 1994) and have also found that this factor explained most of the

inventory’s variance. Factor analysis of the HPSI, confirmed that item 4 which was slightly

modified by replacing the word ‘general public’ by the word ‘physicians’ performed as

expected and loaded moderately onto this factor with a loading of .43.

Items loading onto Factor 2, accounting for 8% of the total variability, primarily

relate to the unpredictability and uncertainty associated with caring for patients and the

patient’s family. This factor has been labelled Patient Care Uncertainty. Previous studies

(Akhtar & Lee; Gupchup & Wolfgang), however, found that these items loaded onto two

separate factors labelled Patient Care Responsibilities and Professional Uncertainty. It can be

seen that item 31 – “Dealing with difficult and demanding relatives,” an item that was added
 202

to the HPSI for the purpose of the current study, performed well and obtained a factor

loading of -.65 on Patient Care Uncertainty.

Most items loading onto Factor 3, which account for 5.88% of the inventory’s

variability, relate to the tasks and responsibilities that nurses’ undertake in addition to caring

for the physical needs of patients. It was therefore labelled Job Conditions. Akhtar and Lee

(2002) and Gupchup and Wolfgang (1994) referred to this factor as Job Conflicts.

Factor 4, which accounted for 4.8% of the variability, consists of two items and refers

to the conflict that may arise as a result of working closely with other nurses. Factor 4 was

therefore labelled Interpersonal Conflict. Eells et al. (1994) also found that these items best

loaded onto the factor labelled Interpersonal Conflict.

In the present study, item 10 – “Not knowing what type of job performance is

expected” and item 23 – “Having job duties which conflict with family responsibilities” did

not load onto any of the factors and therefore will not be included in further statistical

analyses. Item 1 – “Having so much work to do that everything cannot be done well,” and

item 18 – “Being inadequately prepared to meet the needs of patients,” have moderate cross-

loadings. These items however have been retained because their placements in the respective

factors (i.e., Job Conditions and Patient Care Uncertainty) are conceptually consistent with

the wider nursing literature. In addition, the internal consistency of the factors was not

substantially altered by their removal. For instance, Cronbach’s coefficient alpha for Job

Conditions is .79 with the inclusion of item 1. If item 1 is removed, Cronbach’s coefficient

alpha is .78. Similarly, Cronbach’s coefficient alpha is .83 for Patient Care Uncertainty when

item 18 is included. However, if item 18 is removed, Cronbach’s coefficient alpha is .81.

Furthermore, these items appear to accurately reflect Australian nurses’ sources of stress. For
 203

example, the focus group discussions in Study 1 revealed that nurses were indeed concerned

that their high workload and lack of available resources may be compromising the quality of

patient care.

In terms of factor composition, the present study identified factors that are similar,

but not identical to those previously identified by Akhtar and Lee, Eells et al., and Gupchup

and Wolfgang. Eells et al. surveyed a sample of 92 nurses specialising in geriatrics. Principal

factor analysis followed by orthogonal (varimax) rotation yielded four factors accounting for

52% of the variability. These factors were labelled Lack of Perceived Enrichment Potential,

Patient Care, Interpersonal Conflict, and Work-Family Conflict, accounting for 30, 9, 8, and

5% of the total variance, respectively.

Gupchup and Wolfgang surveyed a random sample of 573 practicing pharmacists

from across the United States. Using principal axis factor analysis followed by oblique

(oblimin) rotations a four-factor solution was yielded, accounting for 38.5% of the

variability. The factors were labelled Professional Recognition, Patient Care Responsibilities,

Job Conflicts, and Professional Uncertainty and accounted for 24.1, 7.1, 4.6 and 2.8% of the

total variance, respectively. However, Gupchup and Wolfgang noted that Patient Care

Responsibilities had weak but significant correlations with measures of coworkers’ social

support, career commitment, and organisational commitment.

Akhtar and Lee surveyed a random sample of 2,267 nurses across 43 public hospitals

in Hong Kong to assess whether the structure of the HPSI was generalisable to nurses.

Confirmatory factor analysis provided support for Gupchup and Wolfgang’s four-factor

structure of the stress inventory. The model obtained a Goodness of Fit Index of .93, a

Comparative Fit Index of .89, an Adjusted Goodness of Fit Index of .90, and a Root Mean
 204

Square Residual of .05, indicating statistically acceptable goodness of fit indices.

Professional Recognition, Job Conflicts, and Professional Uncertainty had moderate to high

coefficient alphas ranging from .70 to .80. These factors had moderately positive correlations

with Emotional Exhaustion and Depersonalisation, and weak negative, but statistically

significant correlations with Personal Accomplishment.

Similar to Gupchup and Wolfgang, they also found that the factor Patient Care

Responsibilities obtained a low coefficient alpha (α =.61, p ≤ .01) and weak estimates of

concurrent validity. In addition, it appeared to be a mixed factor with three of the items

contributing little to its measurement. Contrary to their expectation, it had a positive and

significant correlation with Personal Accomplishment, however the correlation was weak and

accounted for little common variance. They recommended further factor analysis be

conducted to determine whether the items regroup into a different structure.

In the current study, a different pattern structure did indeed emerge. The items

identified as loading onto Patient Care Responsibilities in previous studies, loaded onto two

factors in the present study – Patient Care Uncertainty and Job Conditions. For instance, item

16 – “Dealing with difficult patients,” item 28 – “Dealing with terminally ill patients,” and

item 7 – “Disagreeing with other health professionals concerning the treatment of a patient,”

loaded onto Patient Care Uncertainty. Item 14 – “Trying to meeting society’s expectations

for high quality medical care,” item 3 – “Feeling ultimately responsible for patient

outcomes,” item 6 – “Caring for the emotional needs of patients,” and item 12 – “Keeping up

with new developments to maintain professional competence,” loaded onto Job Conditions.

Furthermore, there were two items in the current study that loaded differently from

previous studies. In contrast to Gupchup and Wolfgang, item 17 – “Not being recognised or
 205

accepted as a true health professional by other health professionals,” and item 22 – “Having

non-professionals determine the way you must practice your profession” loaded onto Lack of

Professional Recognition and Support. Gupchup and Wolfgang however, found that these

items loaded onto Professional Uncertainty and Job Conflict, respectively.

The present study yielded a Cronbach coefficient alpha of .83 for Lack of

Professional Recognition and Support, .83 for Patient Care Uncertainty, and .78 for Job

Conditions, indicating good to very good internal consistency across these three factors.

Interpersonal Conflict, however, demonstrated lower internal consistency with a Cronbach

coefficient alpha of .62. Inter-factor total score correlations, as well as correlations of total

scores on the three burnout dimensions are exhibited below in Table 7.2.

Table 7.2

Pearson Product-Moment Correlations Between the HPSI Subscales, ORQ, and the MBI (N

= 273)

Recog Uncerty JCon Intconf Ro Rc Ra Ee Dp Pa

Recog 1.00
Uncerty .49** 1.00
JCon .51** .56** 1.00
Intconf .42** .25** .21** 1.00
Ro .37** .29** .48** .28** 1.00
Rc .55** .27** .24** .36** .33** 1.00
Ra .41** .12* .15* .32** .27** .61** 1.00
Ee .41** .37** .49** .17** .53** .43** .26** 1.00
Dp .30** .47** .27** .11 .24** .43** .25** .51** 1.00
Pa -.09 -.02 -.06 -.06 -.06 -.39** -.37** -.22** -.27** 1.00
Note. Recog = Lack of Professional Recognition and Support; Uncerty = Patient Care Uncertainty; JCon = Job

Conditions; Intconf = Interpersonal Conflict; Ro = Role Overload; Rc = Role Conflict; Ra = Role Ambiguity;

Ee = Emotional Exhaustion; Dp = Depersonalisation; Pa = Personal Accomplishment.

** Correlation is significant at the 0.01 level.

As expected, the four factors comprising the HPSI were significantly positively

correlated. Correlations were low to moderate, indicating that none of the four factors were
 206

statistically redundant. Providing evidence of convergent validity, scores on the four HPSI

factors correlated moderately positively and significantly with the three role stressors – Role

Overload, Role Conflict, and Role Ambiguity. The four HPSI factors also correlated

positively and significantly with scores on the central component of burnout - Emotional

Exhaustion. Lack of Professional Recognition and Support, Patient Care Uncertainty, and Job

Conditions correlated positively and significantly with Depersonalisation. Although

Interpersonal Conflict correlated positively with Depersonalisation, it was not significant. All

four factors on the HPSI correlated negatively with Personal Accomplishment, but not

significantly. It would seem that job-specific stressors are not theoretically related to

Personal Accomplishment.

In summary, similar to previous research, Lack of Professional Recognition and

Support, Patient Care Uncertainty, and Job Conditions demonstrate adequate levels of

reliability and construct validity. Interpersonal Conflict obtained a moderately low

coefficient alpha. Although this may cast some doubt on the reliability of this factor,

Interpersonal Conflict did demonstrate adequate construct validity. For instance,

Interpersonal Conflict was significantly and moderately correlated to Role Overload, Role

Conflict, and Role Ambiguity with reliability coefficients of .28, .36, and .32, respectively.

Interpersonal Conflict was also significantly positively correlated with Emotional Exhaustion

(r = .17, p < .01).

Taking into consideration the factor structure generated in the present study, it was

deemed appropriate to remove items 10 and 23 from the inventory as they did not align to

any of the four factors on the HPSI. Item 31, which was added to the HPSI for the purpose of

the current study, performed well in the factor analysis and was therefore retained. Although
 207

item 1, which loaded onto Job Conditions and item 18 which loaded onto Patient Care

Uncertainty demonstrated moderate cross-loadings, they were kept because they were

conceptually consistent with each factor’s definition. Thus in the present study, the HPSI

consists of 29 items that measure four main sources of stress specific to the nursing

profession: Lack of Professional Recognition and Support, Patient Care Uncertainty, Job

Conditions, and Interpersonal Conflict.

Finally, although the factor structure differs slightly from the studies discussed above,

it is assumed that the factors obtained in the present study are more closely aligned to

Australian nurses’ perceptions of job stress. It is suggested that the generalisability of the

inventory to nurses has been enhanced by the inclusion of an additional item (item 31) and by

the removal of two items (items 10 and 23). These revisions strengthen the HPSI’s factor

loadings and serve to provide a more stable factor structure.

Factor Structure of the Work Support Scales

Coworker support. Factor analysis of the Coworker Support Scale was conducted

using principal axis factoring. No specification was given as to the number of factors to be

extracted however, theoretically, it was expected that the Coworker Support Scale would

yield a two-factor solution comprising of both emotional and instrumental support

components. Two factors with eigenvalues above 1.00 were extracted. The factors were

rotated to a simple structure by employing both orthogonal (VARIMAX) and oblique

(OBLIMIN) rotations. While the orthogonal and oblique solutions were very similar, the

loadings in the oblique solution showed a more defined pattern than the orthogonal solution,

which showed more cross-loadings. Table 7.3 delineates the two-factor structure of the
 208

Coworker Support Scale which was obtained using principal axis factor analysis followed by

oblique (oblimin) rotations.

Table 7.3

Pattern Matrix for Oblique Factors of the Coworker Support Scale

Factors and Items Factors

I II
I. Emotional Coworker Support
1. My coworkers always seem to make time for me if I need to discuss work. .86
2. My coworkers respect me. .83
3. My coworkers listen to my problems. .82
4. My coworkers appreciate the work I do. .80
5. When I am frustrated by some aspect of my work, my coworkers try to .78
understand.
6. My coworkers are understanding and sympathetic. .70
7. I feel comfortable asking my coworkers for help if I have a problem. .57
II. Instrumental Coworker Support
8. My coworkers can be relied on to help when things get tough at work. .87
9. If my job duties become very demanding, my coworkers will take on extra .87
work responsibilities.
10. My coworkers share useful ideas or advice with me. .73
11. My coworkers cooperate with me to get things done at work. .64
12. My coworkers will help me figure out a work problem. .41
Note. Emotional Coworker Support: eigenvalue = 7.08, total variance explained = 58.97%. Instrumental

Coworker Support: eigenvalue = 1.51, total variance explained = 9.6%.

The two-factor solution explained 68.56 % of the total variance. Items loading onto

Factor 1, accounting for 58.97% of the total variability, clearly reflect emotional coworker

support. Items loading onto Factor 2, accounting for 9.6% of the variability, reflect

instrumental coworker support.

Supervisor support. Factor analysis of the Supervisor Support Scale was conducted

using principal axis factoring. No specification was given as to the number of factors to be

extracted but again it was expected that the Supervisor Support Scale would yield a two-

factor solution comprising of emotional and instrumental support components. Contrary to

expectations, a one-factor solution was generated.

 209

Table 7.4

Factor Matrix of the Supervisor Support Scale

Factors and Items Factor

I
I. Supervisor Support
1. When I am frustrated by some aspect of my work, my supervisor tries to understand. .87
2. My supervisor cooperates with me to get things done at work. .87
3. My supervisor is understanding and sympathetic. .87
4. My supervisor listens to my problems. .86
5. My supervisor will help me figure out a work problem. .86
6. My supervisor respects me. .85
7. My supervisor appreciates the work I do. .83
8. I feel comfortable asking my supervisor for help if I have a problem. .83
9. My supervisor always seems to make time for me if I need to discuss work. .81
10 My supervisor can be relied on to help when things get tough at work. .80
11. My supervisor shares useful ideas or advice with me. .78
12. If my job duties become very demanding, my supervisor will take on extra work .70
responsibilities.
Note. Factor 1: eigenvalue = 8.56, total variance explained = 71.34%.

A one-factor solution accounted for 71.34% of the total variability. Based on this

finding it would appear that nurses find it difficult to distinguish between emotional and

instrumental support when it is provided by their supervisor. Coyne and Lazarus (1980)

assert that social support can simultaneously function as a problem-focused and an emotion-

focused coping strategy. For instance, talking to the head nurse about a stressful work event

can function as a problem-focused coping strategy if the nurse receives tangible information

that aids to resolve the event. However, talking to the head nurse may also function as an

emotion-focused coping strategy when the responses obtained also regulate emotional

responses arising from the event. This is one possible reason why nurses may find it difficult

to differentiate between emotional and instrumental support from their supervisors.

Emotional Coworker Support and Instrumental Coworker Support yielded Cronbach

coefficient alphas of .92, and .88 respectively, indicating very good internal consistency.

Cronbach’s coefficient alpha for Supervisor Support is .96, also indicating very good internal

consistency. Inter-factor total score correlations, as well as correlations between the support
 210

scales and the total scores on the work stress factors and burnout subscales are exhibited in

Table 7.5 and Table 7.6, respectively.

Table 7.5

Pearson Product-Moment Correlations Between the Work Support Scales and the HPSI and

ORQ Subscales (N = 273)

Ecs Ics Sups Recog Uncerty JCon Intconf Ro Rc Ra

Ecs 1.00
Ics .72** 1.00
SupS .39** .36** 1.00
Recog -.37** -.41** -.42** 1.00
Uncerty -.07 -.04 -.08 .49** 1.00
JCon -.08 -.17** -.13** .51** .56** 1.00
Intconf -.30** -.24** -.36** .42** .25** .21** 1.00
Ro -.10 -.25** -.20** .37** .29** .48** .28** 1.00
Rc -.35** -.36** -.47** .55** .27** .24** .36** .33** 1.00
Ra -.37** -.40** -.56** .41** .12* .15* .32** .27** .61** 1.00
Note. Ecs = Emotional Coworker Support; Ics = Instrumental Coworker support; SupS = Supervisor Support;

Recog = Lack of Professional Recognition and Support; Uncerty = Patient Care Uncertainty; JCon = Job

Conditions; Intconf = Interpersonal Conflict; Ro = Role Overload; Rc = Role Conflict; Ra = Role Ambiguity.

Untransformed variables were used in the correlation table.

** Correlation is significant at the .01 level. * Correlation is significant at the .05 level.

Table 7.6

Pearson Product-Moment Correlations Between the Work Support Scales and the MBI (N =

273)

Ecs Ics Sups Ee Dp Pa

Ecs 1.00
Ics .72** 1.00
SupS .39** .36** 1.00
Ee -.15* -.25** -.12* 1.00
Dp -.04 -.05 -.04 .51** 1.00
Pa .12 .18** .06 -.22** -.27** 1.00
Note. Ecs = Emotional Coworker Support; Ics = Instrumental Coworker Support; SupS = Supervisor Support;

Ee = Emotional Exhaustion; Dp = Depersonalisation; Pa = Personal accomplishment.

Untransformed variables were used in the correlation table.

** Correlation is significant at the .01 level. * Correlation is significant at the .05 level
 211

In terms of inter-factor correlations, it can be seen that Emotional Coworker Support

was strongly positively correlated with Instrumental Coworker Support (r = .72). Squaring

0.72 indicates that the Coworker Support subscales share approximately 52% variance.

These high correlations were not unexpected. In theory, instrumental support is clearly

distinguished from emotional support. However, when individuals help other people to do

their work (i.e., instrumental support), it shows the recipient of that support that they are

cared for (i.e., emotional support). Thus a purely instrumental act can also have

psychological influences. Therefore, a person being helped with their work may interpret

this act as a sign of caring (Kong & Wertheimer, 1994). Since 48% of the variance remains

unique, however, it was still considered appropriate to investigate the influence of emotional

and instrumental support on the stress-burnout relationship among nurses.

Table 7.5 provides some evidence for the construct validity of the work support

scales. For instance, Emotional Coworker Support was significantly negatively correlated

with Lack of Professional Recognition and Support (r = -.37, p = .01), Interpersonal Conflict

(r = -.30, p = .01), Role Conflict (r = -.35, p = .01) and Role Ambiguity (r = -.37, p = .01).

This suggests the higher the level of emotional support provided by coworkers, the lower the

levels of work stress. Instrumental Coworker Support was significantly negatively related to

Lack of Professional Recognition and Support (r = -.41, p = .01), Job Conditions (r = -.17, p

= .01), Interpersonal Conflict (r = -.24, p = .01), Role Overload (r = -.25, p = .01), Role

Conflict (r = -.36, p = .01) and Role Ambiguity (r = -.40, p = .01). This also suggests the

higher the level of instrumental support provided by coworkers, the lower the levels of work

stress. Furthermore, instrumental coworker support would seem to be more strongly

negatively associated with work stressors than emotional coworker support.

 212

Supervisor Support was significantly negatively related to Lack of Professional

Recognition and Support (r = -.42, p = .01), Job Conditions (r = -.13, p = .01) Interpersonal

Conflict (r = -.36, p = .01), Role Overload (r = -.20, p = .01), Role Conflict (r = -.47, p = .01)

and Role Ambiguity (r = -.56, p = .01). Similarly, higher levels of Supervisor Support are

associated with lower levels of stress. Emotional Coworker Support, Instrumental Coworker

Support, and Supervisor Support were correlated very weakly and negatively to Patient Care

Uncertainty. The correlations were not significant. This result implies that the

unpredictability and uncertainty associated with caring for patients is not conceptually related

to work support. The relationship between the work support scales and the work stressors

(i.e., job-specific stressors and generic role stressors) were all in the expected direction.

Further evidence of construct validity is portrayed in Table 7.6. Emotional Coworker

Support is significantly negatively related to Emotional Exhaustion (r = -.15, p = .05).

Instrumental Coworker Support is also significantly negatively related to Emotional

Exhaustion (r = -.25, p = .01). This indicates that the higher the level of coworker support,

the lower the level of Emotional Exhaustion. Although Emotional and Instrumental

Coworker Support and Supervisor Support correlated negatively with Depersonalisation, it

was not significant. Instrumental Coworker Support is also significantly positively related to

Personal Accomplishment (r = .18, p = .01). This suggests that the higher the level of

instrumental support provided by coworkers, the higher the level of Personal

Accomplishment. Supervisor Support is significantly negatively related to Emotional

Exhaustion (r = -.12, p = .01). Emotional Coworker Support and Supervisor Support were

positively correlated with Personal Accomplishment, however the relationship was not

significant.
 213

In summary, factor analysis of the 12-item Coworker Support Scale revealed that the

items clearly loaded onto two factors, accounting for 68.56% of the total variance. These

factors are labelled emotional and instrumental support. Factor analysis of the 12-item

supervisor support scale revealed a one-factor solution, accounting for 71.34%. The results

demonstrated that nurses are unable to clearly distinguish between emotional and

instrumental supervisor support. It could be inferred that nurses perceive emotional and

instrumental assistance are closely related if their supervisor provides it.

Discussion

In Study 2, a self-report survey was developed to measure the main variables of

interest in this research program: job-specific stress, role stress, work support and burnout.

The qualitative findings generated in Study 1 were used as a basis for choosing appropriate

measures of work stress and in designing a work support scale for nurses. Study 1 provided

some insight into the main sources of stress for a sample of Australian nurses working in

public hospitals. Both job-specific stressors and role stressors were identified. In the present

study, job-specific stress was measured using Wolfgang’s (1988a) HPSI. Three role stressors

(role overload, role conflict, and role ambiguity) were measured using Osipow & Spokane’s

(1987) ORQ. In addition, Study 1’s findings suggested that both sources and types of support

are important aspects to consider when examining the adequacy of social support at work.

Nurses’ main sources of support were their supervisor and their coworkers. The supportive

behaviours provided by their supervisor and their colleagues were classified as emotional and

instrumental support. In the current study, the Supervisor Support Scale and the Coworker

Support Scale were designed to measure the level of emotional and instrumental support

nurses receive at work.

 214

The main objective of Study 2 was to provide some evidence of the psychometric

soundness of the HPSI (Wolfgang, 1988a) and the work support scales designed specifically

for nurses. The HPSI has not been commonly used to measure sources and levels of stress

among nurses. The HPSI assesses 30 stressful events commonly associated with the health

care profession. The items have been found in previous studies to load onto four factors:

professional recognition, patient care responsibilities, professional uncertainty, and job

conflict, (Akhtar & Lee, 2002; Gupchup & Wolfgang, 1994).

In Study 2, the HPSI’s factor structure was assessed using factor analysis. Similar to

previous research, a four-factor structure emerged. The four broad work stress factors were

more fittingly labelled: Lack of Professional Recognition and Support, Patient Care

Uncertainty, Job Conditions, and Interpersonal Conflict. The four-factor solution accounted

for 45.2% of the variance, with Lack of Professional Recognition and Support explaining

most of the variance (26.5%). The four factors demonstrated adequate internal consistency

with Cronbach coefficient alphas of .83 for Lack of Professional Recognition and Support,

.83 for Patient Care Uncertainty, .79 for Job Conditions and .62 for Interpersonal Conflict.

Examination of the inter-factor correlations revealed that none of the four factors

were statistically redundant. Inter-scale correlations provide some evidence of the scale’s

convergent validity. Scores on the four HPSI factors correlated moderately positively and

significantly with the three role stressors. The four HPSI factors also correlated positively

and significantly with scores on the central component of burnout - Emotional Exhaustion.

Lack of Professional Recognition and Support, Patient Care Uncertainty, and Job Conditions

correlated positively and significantly with Depersonalisation. The four factors on the HPSI
 215

correlated weakly and negatively with Personal Accomplishment although the correlations

were not significant.

As expected, the modification of item 4, which taps into a doctors’ lack of respect and

professional recognition for nurses loaded onto Lack of Professional Recognition and

Support. Furthermore, the inclusion of the item 31, which measures the stress associated with

managing difficult and demanding relatives loaded onto the factor Patient Care Uncertainty.

The results indicated however, that the item composition of the HPSI differed

somewhat from previous studies. In particular, two items (items 10 and 23) did not load onto

any of the four factors, and these items were removed from subsequent analyses in Study 3.

The items identified as loading onto Patient Care Responsibilities in previous studies, loaded

onto two factors in the present study – Patient Care Uncertainty and Job Conditions. In the

present study, items 17 and 22 loaded onto the factor Lack of Professional Recognition and

Support. Gupchup and Wolfgang (1994) found that these items loaded onto two separate

factors. Although the factor structure of the HPSI differs from previous studies, it is assumed

that the factors obtained in the present study are more fittingly aligned to Australian nurses’

perceptions of job stress. It was therefore deemed appropriate to retain the four factors and

their corresponding job stress items that emerged in the current study for further statistical

analyses in Study 3.

Factor analysis of the 12-item Coworker Support scale revealed that the items clearly

loaded onto two factors, accounting for 68.56% of the total variance. The factors clearly

represented emotional and instrumental support. The emotional and instrumental coworker

subscales demonstrated very high internal consistency with Cronbach’s coefficient alphas of

.92 and .88 respectively. The major limitation of the Coworker Support Scale is that the
 216

emotional and instrumental support subscales are strongly correlated, sharing 52% variance.

Social support scales in the occupational stress literature such as the Caplan Social Support

Instrument (Caplan et al., 1975) and the Worksite Support Scale (LaRocco et al., 1980), have

also found modest to strong correlations between emotional and instrumental support

subscales. Despite this, differential relationships were observed when emotional coworker

support and instrumental coworker support were correlated with the job-specific stressors

and role stressors. For instance, although Emotional and Instrumental Coworker Support

were both significantly negatively correlated with Lack of Professional Recognition and

Support and Interpersonal Conflict, only Instrumental Coworker Support was significantly

negatively correlated with Job Conditions. Furthermore, while Emotional and Instrumental

Coworker Support were both significantly negatively correlated with Role Conflict and Role

Ambiguity, only Instrumental Coworker Support was significantly negatively correlated with

Role Overload. These results indicate that emotional and instrumental support from

coworkers may influence nurses’ perceptions of work stress in different ways. In particular, it

would seem that only instrumental support from coworkers alters nurses’ perceptions of their

workload.

Similarly, when emotional coworker support and instrumental coworker support were

correlated with the burnout dimensions, differential relationships were observed. For

example, while Emotional and Instrumental Coworker Support were both significantly

negatively correlated to Emotional Exhaustion, only Instrumental Coworker Support was

significantly positively related to Personal Accomplishment. The results provide further

evidence to support the uniqueness of the emotional and instrumental support subscales.
 217

Based on these findings, it was considered feasible to examine emotional and instrumental

coworker support separately in Study 3.

In contrast, factor analysis of the 12-item Supervisor Support Scale revealed a one

factor-solution, accounting for 71.34%. The results demonstrated that nurses are unable to

clearly distinguish between emotional and instrumental support when it is provided by their

supervisor. For instance, if a nurse complained to his/her supervisor about his/her rostering

schedule, the supervisor may first listen to the nursing subordinate’s concerns (i.e., emotional

support) before offering to change their roster (instrumental support). In this instance, the

source of the support may be more important than the type of support given. Supervisor

support demonstrated some evidence of convergent validity. For example, Supervisor

Support was significantly negatively correlated with Lack of Professional Recognition and

Support, Job Conditions and Interpersonal Conflict. In addition Supervisor Support was

significantly negatively correlated with the three role stressors. Furthermore, Supervisor

Support was significantly negatively correlated with Emotional Exhaustion. The Supervisor

Support Scale demonstrated high internal consistency with a Cronbach coefficient alpha of

.96. Taking into consideration these findings, it was not appropriate to examine emotional

and instrumental supervisor support in Study 3. Instead, the scores for the 12 items that

comprise the Supervisor Support Scale will be totalled to give a global measure of supervisor

support.

Summary

In conclusion, the findings from Study 2 provide supporting evidence to suggest that

the job-specific stress measure (i.e., the HPSI) and the new work support measures (i.e.,

coworker support and supervisor support) possess adequate levels of internal consistency and
 218

construct validity. More specifically, factor analysis of the HPSI revealed a four-factor

structure: Lack of Professional Recognition and Support (10 items), Patient Care Uncertainty

(9 items), Job Conditions (8 items) and Interpersonal Conflict (2 items). A two-factor

structure emerged for the Coworker Support Scale. These factors were labelled: Emotional

Coworker Support (7 items) and Instrumental Coworker Support (5 items). Factor analysis of

the Supervisor Support Scale revealed a one-factor solution comprising 12 items. The four

subscales that comprise the HPSI and the two subscales that encompass the Coworker

Support Scale will be explored independently in Study 3. Due to the high inter-item

correlations, different types of supervisor support will not be examined further in Study 3.

Instead, the 12 items will be aggregated to give a global measure of supervisor support.

Limitations

A significant limitation of the present study’s research design was the inability to

pilot the scales before administering the full survey to nurses. In addition, the survey was not

re-administered to nurses over a period of time. Although the researcher desired to evaluate

the scales by conducting a pilot study with a smaller sample of nurses and to utilise a

longitudinal research methodology, permission was not granted by the hospitals’ ethics

committees due to a recent trend of over-surveying of nurses. The researcher was therefore

limited to a cross-sectional survey methodology. Furthermore, the HPSI and work support

scales were validated using only one sample and therefore it is not known whether the same

factor structure would result using other nursing samples. Despite these limitations, the

evaluation of the psychometric properties of the HPSI and the work support scales provided

sufficient evidence of their reliability and validity. In the following chapter, the relationships

between the main variables of interest to this research program will be explored further.
 219

CHAPTER 8

Study 3

The Influence of Social Support on the Stress-Burnout Relationship

Overview

Study 3 builds upon previous nursing literature by examining the effect of social

support on the stress-burnout relationship. In particular, two models of social support are

tested – the main effect and the buffering effect model. The main effect model proposes that

regardless of the level of work stress, when social support is high, burnout is lower than

when social support is low (Kessler & McLeod, 1985). The buffering effect proposes that

social support buffers the impact of stressors on manifestations of strain such that the

relations between stress and strain is stronger for persons with low levels of social support

than those with high levels of support. From this perspective, social support does not

necessarily lower the level of experienced stress but instead aids the employee to cope with

the stressful situation. Hence, the impact of social support is expected to be greater for those

experiencing high levels of stress. For those who are not experiencing stressful events the

benefits of social support may be minimal (Hall & Wellman, 1985).

According to the stressor-support matching theory (Cohen & Wills, 1985), social

support is more likely to buffer the negative effects of stress if the support received is able to

address the needs of the situation. The present study explores how certain types of supports

(i.e., emotional/instrumental support) and specific sources of support (i.e.,

coworker/supervisor support) interact with particular stressors to produce stress-buffering

effects on burnout. The thesis also extends previous stressor-support matching studies by

taking into consideration a nurse’s control over a stressful situation. Cutrona and Russell
 220

(1990) assert that controllable and uncontrollable stressful events require different types of

support to be optimally effective in buffering strains. Their optimal matching model proposes

that when a stressful situation is primarily controllable, the most beneficial support is

instrumental. However, when a stressful situation is essentially uncontrollable, the most

effective type of support is emotional. Thus in the present study, nurses’ main sources of

stress are classified as controllable or uncontrollable events by two independent raters prior

to examining the main and moderating effect of social support.

In this chapter, the main variables of interest in this research program will first be

explored independently. Specifically, the average level of work stress experienced by public

hospital nurses and the key stressors causing nurses the most concern are identified. The

prevalence of burnout for the sample population is estimated and compared to appropriate

normative groups in an attempt to verify whether Australian nurses experience similar levels

of burnout to other nurses and other human service professionals. The average level of

support nurses receive at work is established. Next, the inter-relationships between these

variables are explored. The study ascertains which aspects of nursing work are related to

burnout. Finally, the relative contribution to burnout made by the independent variables (i.e.,

sociodemographic factors, work stressors and work support) is established.

Rationale for Study 3

The final study brings this research program to its logical conclusion. Study 3 builds

upon the results reported in the first two studies. First, a range of stressful work events were

identified by public hospital nurses in Study 1. In Study 2, these job-specific stressors and

generic role stressors were measured using the HPSI and three subscales on the ORQ,

respectively. Factor analysis of the HPSI revealed that the items loaded onto four broad work
 221

stress factors: Lack of Professional Recognition and Support, Patient Care Uncertainty, Job

Conditions and Interpersonal Conflict. In Study 3, the effects of the four job-specific stress

variables and the three role stress variables on burnout are examined. Furthermore, the

average level of work stress and the key determinants of stress for Australian nurses are

established.

Second, the results in Study 1 enabled the researcher to clearly operationalise the

social support construct. In the second study, a Supervisor Support Scale and a Coworker

Support Scale was constructed using items from pre-existing social support scales to tap into

nurses’ perceptions of the support they receive at work. Factor analysis of the Coworker

Support Scale revealed that items loaded onto two broad factors labelled emotional and

instrumental support. Factor analysis of the Supervisor Support Scale however, revealed a

one-factor solution, suggesting that nurses are unable to clearly distinguish between

emotional support and instrumental support when it is provided by their supervisor. Thus in

Study 3, an examination of Cutrona and Russell’s (1990) proposition that uncontrollable and

controllable stressful events require different types of support, will be confined to

investigating only emotional and instrumental support provided by coworkers. The main and

moderating effects of sources of support (coworker/supervisor) on burnout will also be

explored.

In addition, Study 3 addresses major gaps in the nursing literature regarding burnout.

Most nursing studies investigating burnout have limited their study to emotional exhaustion

or to the first two components of burnout – emotional exhaustion and depersonalisation.

Furthermore, studies have employed different measures of burnout making comparisons

between studies difficult. The present study, however, views burnout as a three-component
 222

syndrome and therefore assesses burnout using the widely recognised MBI (Maslach et al.,

1996). Furthermore, despite research suggesting that nurses are at high risk of experiencing

burnout throughout their nursing career, empirical studies have generally reported moderate

levels of emotional exhaustion, depersonalisation and reduced personal accomplishment

(e.g., Iacovides et al., 1997; Turnipseed & Turnipseed, 1997). An extensive search has

revealed that there are no recent studies that have investigated the degree of burnout among

Australian nursing staff working in public hospitals. In the present study, the mean MBI

scores will be generated for each component of burnout. Study 3 will determine whether

Australian nurses experience similar levels of burnout to other nursing professionals, as well

as other human service professionals. Finally, the current study will enhance our

understanding of the major antecedents of burnout among Australian nurses. Similar to

previous nursing studies, both sociodemographic factors and specific work stressors are

explored in relation to the three burnout dimensions. Most studies investigating nurses’ levels

of burnout, however, have not examined the effect of work support on burnout. The present

study extends our knowledge of burnout by exploring the influence of work support on

burnout for nurses. In the section below, the research hypotheses to be addressed in Study 3

are outlined.

Research Hypotheses

In Study 3, the following hypotheses are examined.

Hypothesis 1: Nurses’ levels of work stress will not be significantly different based

on gender.

Hypothesis 2: Burnout consists of three related, but empirically distinct components –

emotional exhaustion, depersonalisation, and personal accomplishment.

 223

Hypothesis 3: Australian nurses will report moderate levels of emotional exhaustion,

depersonalisation and reduced personal accomplishment.

Hypothesis 4: Both job-specific stressors and role stressors will be related to burnout

in nurses.

Hypothesis 5: Job-specific stressors will be significantly and positively related to

emotional exhaustion and depersonalisation and significantly negatively related to personal

accomplishment.

Hypothesis 6: Role stressors will be significantly positively related to emotional

exhaustion and depersonalisation and significantly negatively related to personal

accomplishment.

Hypothesis 7: Younger nurses will report higher levels of burnout than older nurses.

Hypothesis 8: Less experienced nurses will report higher levels of burnout than more

experienced nurses.

Hypothesis 9: Nurses will report higher levels of coworker support than supervisor

support.

Hypothesis 10: Work support will be significantly negatively correlated to the

emotional exhaustion and depersonalisation and significantly positively related to personal

accomplishment.

Hypothesis 11: Work support will be significantly negatively correlated to work

stressors.

Hypothesis 12: Work support will significantly buffer the relationship between work

stress and burnout.

 224

Hypothesis 13: For stressful events deemed to be uncontrollable, emotional support

will have a significant buffering effect on burnout.

Hypothesis 14: For stressful events deemed to be controllable, instrumental support

will have a significant buffering effect on burnout.

Method
Subjects

The participant profile of Study 3 is the same as the participant profile in Study 2.

This is because the data collected in Study 2 has been subjected to further statistical analyses

in Study 3. Briefly, the sample comprises 273 nursing professionals (235 females, 38 males)

from a range of nursing wards from three public hospitals in the Queensland. A complete

description of the sociodemographic characteristics of the sample is presented in Chapter 5.

Materials

In Study 2, the researcher developed a questionnaire to examine the main variables of

interest in this research program: job-specific stress, role stress, coworker support, supervisor

support, and burnout. In particular, the HPSI (Wolfgang, 1988a) was chosen to measure

nurses’ sources of job-specific stress. The ORQ (Osipow & Spokane, 1987) was chosen to

measure role stress (role overload, role conflict, and role ambiguity). The Coworker Support

Scale and the Supervisor Support Scale was developed by the researcher. Items were taken

from established measures including: Shinn et al.’s (1989) Supervisor Support Scale, Ray

and Miller’s (1994) Supervisor/Coworker Support Scale, and King et al.’s (1995) Family

Support Inventory for Workers and slighted adapted to make them contextually relevant to

nurses. The MBI (Maslach et al., 1996) was chosen to measure nurses’ levels of emotional

exhaustion, depersonalisation, and reduced personal accomplishment. These measures are

described in detail in Chapter 5.

 225

It was the researcher’s intention to use the data collected in Study 2 for further

analyses in Study 3. Factor analysis of the job-specific stress measure and the work support

scales in Study 2 revealed however, that some modifications to these scales was necessary

prior to statistical analysis in Study 3. These alterations are outlined in Chapter 7. Previous

empirical studies demonstrate adequate factorial validity for the ORQ and the MBI and thus

further investigation into these scales item composition was not required. Study 3 also

examines nurses’ perceptions of work stress controllability. Two independent raters were

used to classify nurses work stressors as controllable and uncontrollable stressful events. The

rating exercise used to categorise the work stress variables according to situation

controllability is described in the following section.

Work Stress Controllability

As previously discussed, Cutrona and Russell’s (1990) optimal matching model

proposes that an individual’s perception of controllability over a stressful situation is an

important factor in determining which type of support will be effective in reducing or

preventing strain outcomes. They suggested that emotional support is most effective in

uncontrollable situations. In contrast, instrumental support is most effective in controllable

situations. It was therefore necessary to clarify nurses’ sources of stress as controllable or

uncontrollable events before examining the buffering effects of support on burnout.

Following factor analysis of the HPSI in Study 2, two independent raters were asked

to categorise nurses’ major sources of stress as controllable and uncontrollable events (see

Appendix C). Two female nurses aged 53 years and 55 years, with approximately 25 years

nursing experience participated in the rating exercise. The independent raters were employed

at Hospital A and Hospital C. Raters were provided with a definition for each job-specific
 226

stress factor and each role stress factor. The items that aligned to each broad stress category

were also supplied. Raters were asked to decide whether each job-specific stress factor and

each role stress factor was primarily controllable or primarily uncontrollable. Controllable

sources of nurses’ stress were defined as “situations in which the occurrence or consequence

of the stressful event is preventable” (Cutrona, 1990, p. 8). Uncontrollable sources of nurse

stress were defined as “situations in which the occurrence or consequence of the stressful

event is not preventable” (Cutrona, 1990, p. 8). To aid their decision making, the raters were

first asked to identify whether the stressful events that aligned to each broad stress category

were controllable or uncontrollable. If the total number of controllable events exceeded the

uncontrollable events, it was inferred that the stress factor was primarily controllable. If the

number of uncontrollable events exceeded the controllable events, it was inferred that the

stress factor was primarily uncontrollable. Where raters did not agree on the classification of

a particular source of stress, a third rater arbitrated. The third rater was a female nurse (58

years) with 30 years experience. The consistency of the rating exercise was checked by

calculating an inter-rater reliability coefficient. An inter-rater reliability coefficient was

calculated by dividing the number of coding agreements by the number of coding agreements

plus the number of coding disagreements (Goodwin & Goodwin, 1985). The overall inter-

rater reliability estimate for the coding of comments was .79. Categorisation of the four job-

specific stress factors and the three generic role stress factors based on controllability are

presented in Table 8.1.

 227

Table 8.1

Nurses’ Perceptions of Work Stress Controllability

Source of Stress Perception of IR

Controllability

Lack of Professional Recognition and Support Controllable .88

Patient Care Uncertainty Controllable .77

Job Conditions Uncontrollable .77

Interpersonal Conflict Controllable .71

Role Overload Controllable .66

Role Conflict Controllable .73

Role Ambiguity Controllable 1.00

Note. IR = Inter-rater reliability coefficient.

Table 8.1 would seem to indicate that nurses perceive that they have a high degree of

control over stressful situations at work. More specifically, nurses believe they can reduce or

prevent the occurrence and/or the outcome of most stressful situations at work. Nurses do not

consider, however, that they can control the conditions associated with their job. Nurses are

expected to contend with heavy workloads without adequate staffing support. Not only are

they responsible for patient outcomes, they are also responsible for supervising less

experienced nursing colleagues. They constantly strive to meet societal expectations for high

quality medical care and to meet the emotional needs of their patients. They are required to

update their skills and qualifications in order to keep up with new developments and to

maintain professional competence. Nurses’ job conditions appear to be primarily controlled

by external sources (e.g., the government, the hospital, and society). It is therefore not

surprising that Job Conditions are perceived to be demands that are beyond a nurse’s control.

Applying the above findings to Cutrona and Russell’s (1990) optimal matching

theory, it could be inferred that:

 228

• Emotional Coworker Support will interact with Job Conditions to prevent burnout.

• Instrumental Coworker Support will interact with the remaining sources of job-

specific stress (i.e., Lack of Professional Recognition and Support, Patient Care

Uncertainty and Interpersonal Conflict) to prevent burnout.

• Instrumental Coworker Support will interact with the three sources of role stress (i.e.,

Role Overload, Role Conflict, and Role Ambiguity) to prevent burnout among nurses.

In Study 3, Cutrona and Russell’s optimal matching theory is examined within a nursing

context.

Procedure

The procedure used to collect the quantitative data is described earlier in Chapter 7,

Study 2.

Data Analysis

A description of the statistical analysis techniques used to examine the present study’s

research hypotheses is outlined earlier in Chapter 5.

 229

Results

Descriptive Statistics for the Independent and Dependent Variables

Descriptive statistics, including internal reliability coefficients using Cronbach’s

alpha for each independent and dependent variable are presented in Table 8.2.

Table 8.2

Descriptive Data for the Independent and Dependent Variables (N=273)

Scale Mean S.D. S.E. Skew Kurtosis Alpha No. of

Mean Level Items

RECOG 15.51 7.80 0.47 2.74 -1.08 .83 10

UNCERTY 11.24 6.12 0.37 5.14 2.00 .82 9
JOB CON 15.71 6.27 0.38 1.48 -2.59 .79 8
INTCONF 1.88 1.63 0.10 10.61 10.38 .62 2

RO 26.02 6.82 0.41 3.08 -1.19 .81 10

RC 22.94 6.58 0.40 2.48 -0.79 .73 10
RA 19.77 5.57 0.38 3.77 0.15 .71 10

EE 23.01 11.17 0.68 1.76 -2.12 .90 9

DP 7.75 5.89 0.36 5.86 0.54 .71 5
PA 34.62 7.83 0.47 - 3.63 -0.36 .75 8

ECS 28.03 5.14 0.31 - 5.53 3.40 .92 7

ICS 19.75 3.62 0.22 - 5.07 4.34 .88 5
SUPS 44.11 11.21 0.68 - 4.31 0.54 .96 12

Note. RECOG = Lack of Professional Recognition and Support; UNCERTY = Patient Care Uncertainty; JOB

CON = Job Conditions; INTCONF = Interpersonal Conflict; RO = Role Overload; RC= Role Conflict; RA =

Role Ambiguity; EE = Emotional Exhaustion; DP = Depersonalisation; PA = Personal Accomplishment; ECS

= Emotional Coworker Support; ICS = Instrumental Coworker Support; SUPS = Supervisor Support.

* Scores for skew and kurtosis for each variable are recorded as z-scores.
 230

Based on Cronbach’s alpha, it can be seen that high internal reliability coefficients

were found for Lack of Professional Recognition and Support (α = .83), Patient Care

Uncertainty (α = .82), and Job Conditions (α = .79). A lower internal reliability coefficient

was generated for Interpersonal Conflict (α = .62).

Although Osipow and Spokane (1987) reported slightly higher reliability coefficients

of .83 for Role Overload, .82 for Role Conflict and .78 for Role Ambiguity, the internal

reliability coefficients for the role stressors in the present study are adequate with coefficients

ranging from .71 to .81.

The reliability coefficients reported for the burnout subscales in the current study

were similar to those reported by Maslach et al. (1996) of .90 for Emotional Exhaustion, .79

for Depersonalisation, and .71 for Personal Accomplishment.

As discussed earlier in Chapter 7, high internal reliability coefficients were found for

Emotional Coworker Support (α = .92), Instrumental Coworker Support (α = .88), and

Supervisor Support (α = .96).

The data were screened for normality, linearity, and homescedasticity prior to

statistical analyses. The data were also screened for multicollinearity and singularity, as well

as outliers. Z-scores were calculated to test for skewness and kurtosis in the sample

population. The distributions for Patient Care Uncertainty, Interpersonal Conflict, Role

Ambiguity, and Depersonalisation were positively skewed. Square root transformations were

performed to improve the distribution of these variables. Following the transformation, high

levels of Patient Care Uncertainty, Interpersonal Conflict, Role Ambiguity, and

Depersonalisation are reflected by high scores on these scales. Conversely, low levels of
 231

Patient Care Uncertainty, Interpersonal Conflict, Role Ambiguity, and Depersonalisation are

reflected by low scores on these scales.

The distributions for the burnout dimension - Personal Accomplishment, and the

work support scales (i.e., Emotional Coworker Support, Instrumental Coworker Support, and

Supervisor Support) were negatively skewed. Appropriate transformations for negative skew

(i.e., reflect and square root) were therefore conducted (Tabachnick & Fidell, 1989).

Following the transformation of Personal Accomplishment, a high total score relates to low

levels Personal Accomplishment (or high levels of burnout) and a low total score relates to

high levels of Personal Accomplishment (or low levels of burnout). Following the

transformation of the work support variables (Emotional Coworker Support, Instrumental

Coworker Support, and Supervisor Support), high total scores reflect low levels of support

and low total scores reflect high levels of support.

In the section below, the nature of nurses’ work stress is explored. It is important to

note that the variable Role Ambiguity is not transformed in the descriptive analysis below.

By using the untransformed variable, comparisons between the present study’s results and

the normative data provided in the ORQ administration manual could be made.
 232

Level of Work Stress Among Australian Nurses

Table 8.3 depicts the average level of job-specific stress and role stress reported by

the total sample population. In addition, total mean scores for the work stress scales are

reported separately for male and female nurses. For the untransformed variable – Role

Ambiguity, a high total score reflects a high level of Role Ambiguity.

Table 8.3

Mean Levels of Work Stress for the Total Sample Population and Across Gender

Total Males Females Test of

Sample (n = 38) (n = 235) Differences
(n =273)

Work Stress Mean SD Mean SD Mean SD t p

HPSI 44.35 17.31 48.01 21.05 43.76 16.61 1.41 .16

(score ranges:0-116)
Role Overload 26.02 6.82 25.82 6.50 26.06 6.89 -.20 .84
(score ranges:10-50)
Role Conflict 22.95 6.58 24.27 7.10 22.73 6.48 1.34 .18
(score ranges:10-50)
Role Ambiguity * 19.77 5.57 19.86 5.30 19.75 5.62 .12 .91
(score ranges:10-50)
Note. HPSI = Health Professions Stress Inventory. * Role Ambiguity has not been transformed.

As shown in Table 8.3, the mean score on the HPSI for the total sample population

indicates that nurses report moderately low levels of job-specific stress (M = 44.35). Males

(M = 48.01) reported higher levels of job-specific stress than females (M = 43.76). However,

an Independent Samples t-test found that there were no significant differences between male

and female nurses on perceptions of job-specific stress (t (271) = 1.41, p = .16, two-tailed).

The sample population also reported moderately high levels of Role Overload, and

moderate levels of Role Conflict and moderately low levels of Role Ambiguity. The mean

total scores for males and females across the role stressors were compared to normative data

provided in Osipow and Spokane’s (1987) OSI administration manual. The normative data is
 233

based on 909 professionals employed in the human services industry as well as the technical

and manufacturing industry. For males, Role Overload total scores ranging between 22 and

35 are indicative of normal levels of stress. For females, total scores ranging between 18 and

33 are indicative of normal levels of stress. It could therefore be inferred that both male (M =

25.82) and female (M = 26.06) nurses report normal levels of stress associated with Role

Overload. An Independent Samples t-test revealed that there were no significant differences

between male and female nurses’ perceptions of Role Overload (t (271) = -.20, p = .84, two-

tailed).

For males, Role Conflict total scores ranging between 15 and 26 indicate normal

levels of stress. For females, total scores ranging between 14 and 29 indicate normal levels of

stress. It could therefore be inferred that both male (M = 24.27) and female (M = 22.73)

nurses report normal levels of stress associated with Role Conflict. An Independent Samples

t-test revealed that there were no significant differences between male and female nurses’

perceptions of Role Conflict (t (270) = 1.34, p = .18, two-tailed).

For males, Role Ambiguity total scores ranging between 15 and 26 indicate normal

levels of stress. For females, total scores ranging between 14 and 25 indicate normal levels of

stress. It could therefore be inferred that both male (M = 19.86) and female (M = 19.75)

nurses report normal levels of stress associated with Role Ambiguity. An Independent

Samples t-test revealed that there were no significant differences between male and female

nurses’ perceptions of Role Ambiguity (t (270) = .12, p = .91, two-tailed).

In summary, the level of role stress reported by male and female nurses is within the

normal range. While female nurses report slightly higher levels of Role Overload and slightly

lower levels of Role Conflict and Role Ambiguity than males, there are no significant
 234

differences in levels of role stress based on gender. These results provide support for

Hypothesis 1 that nurses’ levels of work stress are not significantly different based on

gender.

Sources of Stress for Australian Nurses

Major job-specific stressors. The total mean scores for the 29-item HPSI are

presented in Figure 8.1. The items highlighted in red, are those items perceived as often

stressful by nurses. Those highlighted in yellow, depict the items that are considered

occasionally stressful by nurses. Those highlighted in green, depict the items that are rated

as never or rarely stressful to nurses.

3
Mean Total Score

0
hp11
hp2
hp18
hp26
hp9
hp5
hp8
hp30
hp7
hp19
hp17
hp28
hp29
hp25
hp12
hp20
hp15
hp22
hp13
hp31
hp14
hp4
hp6
hp1
hp16
hp3
hp21
hp27
hp24

HPSI Item

Figure 8.1. Total mean item scores for the Health Professions Stress Inventory.

___________________________________________________________________________
Note. Item numbers are from the original questionnaire. Items 10 and 23 were removed.
 235

Figure 8.1 suggests that there are very few work events that often cause nurses stress.

The majority of work conditions are perceived to be ‘occasionally’ stressful. Since these

stressors do not contribute to our knowledge of Australian nurses’ a) major sources of job-

specific stress, or b) least stressful job-specific conditions, these events will not be discussed.

The most stressful job-specific conditions, the total item mean scores and the corresponding

HPSI subscale for each item, are presented in Table 8.4.

Table 8.4

The Highest Rated Job-Specific Stressors for Australian Nurses

Item Number and Description Item HPSI Subscale

Mean
24. Being interrupted by phone calls or people 2.67 Job Conditions
while performing job duties.
27. Feeling that you are inadequately paid as a 2.57 Lack of Professional Recognition
health professional. and Support
21. Not having enough staff to adequately 2.34 Job Conditions
provide necessary services.
3. Feeling ultimately responsible for patient 2.03 Job Conditions
outcomes.
16. Dealing with difficult patients. 2.02 Patient Care Uncertainty

The job-specific events rated as ‘often’ stressful included: being interrupted by phone

calls or people while performing job duties (M = 2.67); feeling that you are inadequately paid

as a health professional (M = 2.57); not having enough staff to adequately provide necessary

services (M = 2.34); feeling ultimately responsible for patient outcomes (M = 2.03); and

dealing with difficult patients (M = 2.02). Wolfgang (1988b) reported similar results for a

sample of 379 registered nurses using the HPSI. The results indicate that Job Conditions are a

significant source of stress for Australian nurses.

In contrast, the job-specific events that are considered to be ‘never’ or ‘rarely’

stressful included: allowing personal feelings or emotions to interfere with the care of
 236

patients (M = 0.60); experiencing conflicts with supervisors and/or administrators (M =

0.84); being inadequately prepared to meet the needs of patients (M = 0.95) and not being

challenged by your work (M = 0.99).

Major role stressors. The total mean scores for the 30 items comprising the three role

stress subscales are presented in Figure 8.2. The items highlighted in red, depict items rated

as often stressful by nurses. Those highlighted in yellow, depict the items that are rated

occasionally stressful by nurses. Those highlighted in green, depict the items that are rated

as never or rarely stressful to nurses.

4
Mean Total Score

1
ra7
ra4
ro6
ra8
ra9
ra6
rc6
rc7
rc5
ro4
ra5
ro3
rc9
ra3
ra10
rc2
rc1
ro5
rc8
rc3
ro10
ro8
ra2
rc10
ro1
ro9
ro2
ra1
ro7
rc4

Role Stress Items

Figure 8.2. Total mean item scores for the Occupational Roles Questionnaire.
________________________________________________________________________________
Note. ro = Role Overload; rc = Role Conflict; ra = Role Ambiguity
 237

Figure 8.2 indicates that the majority of role stressors are ‘never’ or ‘rarely’ stressful,

with a much smaller number of events associated with ‘often’ causing role stress. Since role

stressors rated ‘occasionally’ stressful do not contribute to our knowledge of Australian

nurses’ a) major sources of role stress, or b) least stressful role conditions, these events will

not be discussed. The most stressful role stressors, the total item mean scores and the

corresponding subscale for each item, are presented in Table 8.5.

Table 8.5

The Highest Rated Role Stressors for Australian nurses

Item Description Mean ORQ Subscale

I feel I have a stake in the success of this hospital. * 3.61 Role Conflict
I work under tight deadlines. 3.51 Role Overload
My supervisor provides me with useful feedback about my 3.50 Role Ambiguity
performance. *
I feel that my job responsibilities are increasing. 3.43 Role Overload
My job requires me to work in several equally important 3.17 Role Overload
areas at once.
At work I am expected to do too many different things in too 3.06 Role Overload
little time.
Note. * = reverse scored item

The role stress events rated as ‘often’ causing nurses’ stress included: not having a

stake in the success of the hospital (M = 3.61); working under tight deadlines (M = 3.51); not

receiving useful feedback about job performance by their supervisor (M = 3.50); increasing

job responsibilities (M = 3.43); working in several equally important areas at once (M = 3.17)

and being expected to do too many different tasks in too little time (M = 3.06). The results

demonstrate that role overload is a major source of stress for Australian nurses.

In contrast, the role stress events that are considered to be ‘never’ or ‘rarely’ stressful

included: determining acceptable behaviour on the job (e.g., dress, interpersonal relations)

(M = 1.17); knowing which task should be done first when presented with several tasks (M =
 238

1.46); feeling competent in doing their job (M = 1.48) and knowing what constitutes as a

high priority (M = 1.48). It could be inferred that Australian nurses are relatively clear about

what is expected of them as a nurse and are confident in their ability to fulfill job

requirements.

The Nature of Burnout Among Australian Nurses

Intercorrelations between the burnout dimensions. Table 8.6 reports the inter-

correlations between the three burnout subscales. Transformed variables included

Depersonalisation and Personal Accomplishment. It should be noted that a high total score on

Depersonalisation relates to a high level of Depersonalisation and a low total score reflects a

low level of Depersonalisation. A high total score on Personal Accomplishment relates to a

low level of Personal Accomplishment (or high levels of burnout) and a low total score

relates to a high level of Personal Accomplishment (or low levels of burnout). Therefore, all

burnout subscales are expected to have positive correlations.

Table 8.6

Pearson Product-Moment Correlations between the MBI Subscales (N = 273)

Emotional Depersonalisation Personal

Exhaustion Accomplishment

Emotional 1.00
Exhaustion

Depersonalisation .48** 1.00

Personal .24** .29** 1.00

Accomplishment
Note. ** Correlation is significant at the 0.01 level.

Pearson’s product-moment correlation coefficients were used to examine the inter-

relationships between the burnout components. Table 8.6 demonstrates that Emotional
 239

Exhaustion is positively correlated with Depersonalisation (r = .48, p = .01), suggesting that

high levels of Emotional Exhaustion are associated with high levels of Depersonalisation.

Emotional Exhaustion is significantly positively associated with Personal Accomplishment (r

= .24, p = .01), indicating that high levels of Emotional Exhaustion are associated with low

levels of Personal Accomplishment. Similarly, Depersonalisation was also significantly

positively correlated with Personal Accomplishment (r = .29, p = .01), suggesting that high

levels of Depersonalisation are associated with low levels of Personal Accomplishment.

These results confirm Hypothesis 2 that burnout consists of three related, but empirically

distinct components.

Average level of burnout. Table 8.7 presents the MBI subscale means and standard

deviations for the current sample. The total mean MBI scores are compared to an appropriate

normative sample reported in the MBI administration manual (Maslach et al., 1996). The

normative sample consists of 1104 medicine professionals, including physicians and nurses.

In establishing the overall level of burnout, the scores were examined in relation to the ranges

of burnout defined by Maslach et al. (1996). The score ranges reflect low, moderate, and high

levels of burnout. Burnout scores are considered high if they fall within the upper third of the

normative distribution, moderate if they fall within the middle third, and low if the fall within

the lower third of the distribution (Maslach et al., 1996). It should be noted that the variables

Depersonalisation and Personal Accomplishment are not transformed in the descriptive

analysis below. By using the untransformed variables, comparisons can be made between the

present study’s results and the normative data provided in the MBI administration manual.

Hence, the higher the mean score on Depersonalisation, the higher the level of
 240

Depersonalisation. In contrast, the lower the mean score on Personal Accomplishment, the

lower the level of Personal Accomplishment.

Table 8.7

Means, Standard Deviations, and Reliability Coefficients for a Sample of Australian Nurses

and a MBI Normative Sample

Sample MBI Norms

n = 273 n = 11067
Variable Mean SD α Mean SD α

Emotional Exhaustion 23.01 11.17 .90 22.19 9.53 .90

(0-78)
Depersonalisation* 7.75 5.89 .71 7.12 5.22 .79
(0-96)
Personal 34.62 7.83 .75 36.53 7.34 .71
Accomplishment*
(0-84)
Note. MBI = Maslach Burnout Inventory. Depersonalisation and Personal Accomplishment

have not been transformed.

Range of Experienced Burnout

: EE – High: 27 or > Moderate: 19-26 Low: 0-18
: DP – High: 8 or > Moderate: 5-7 Low: 0-4
: PA – High: 0-33 Moderate: 34-39 Low: 40 or >

The data presented in Table 8.7 indicates that overall, nurses’ report moderate levels

of Emotional Exhaustion, Depersonalisation, and Personal Accomplishment, thus confirming

Hypothesis 3. More specifically, the sample of Australian nurses reported moderate levels of

Emotional Exhaustion, moderately high levels of Depersonalisation, and moderately low

levels of Personal Accomplishment. The mean burnout scores are similar to those reported in

Maslach et al.’s (1996) MBI manual for the medicine norm group. They reported a mean

score of 22.19 (SD = 9.53) for Emotional Exhaustion, 7.12 (SD = 5.22) for

Depersonalisation, and 36.53 (SD = 7.34) for Personal Accomplishment. This finding

suggests that Australian nurses experience similar levels of burnout to other medical staff.
 241

Although comparing nurses’ levels of burnout with the medicine profession used to

construct the norms for the MBI is useful, much of Maslach et al.’s work was conducted

more than a decade ago. It therefore seems appropriate to compare the current sample’s

burnout levels with more recent nursing studies as shown in Table 8.8.

Table 8.8

Mean Scores for the MBI Subscales for Nursing Studies Between 1991 and 2001

Researchers Type of Location N EE DP PA

Nursing
Present study General Australia 273 23.01 7.75 34.62
Iacovides et al. (1997) General Greece 368 22.40 7.59 36.18
Schmitz et al. (2000) General Germany 361 19.60 10.60 31.00
Bakker et al. (2000) General Germany 207 22.21 8.01 33.25
Schaufeli & Janczur (1994) General Poland 200 20.00 8.70 27.30
Schaufeli & Janczur (1994) General Holland 183 16.20 5.40 32.70
Turnipseed & Turnipseed (1997) General United States 129 24.50 7.75 37.70
Turnipseed & Turnipseed (1997) General Philippines 71 17.50 6.76 33.80
Payne (2001) Hospice United 89 17.19 3.91 35.70
Kingdom
Edwards et al. (2000b) Mental United 301 21.20 5.20 34.80
Health Kingdom
Mallet et al. (1991) Critical Care United States 167 22.20 10.60 56.40
Ogus (1992) Medical Canada 62 45.90 18.00 25.44
Ogus (1992) Surgical Canada 66 29.74 10.18 20.32
Note. EE = Emotional Exhaustion; DP = Depersonalisation; PA = Personal Accomplishment.

Range of Experienced Burnout

: EE – High: 27 or > Moderate: 19-26 Low: 0-18

: DP – High: 8 or > Moderate: 5-7 Low: 0-4
: PA – High: 0-33 Moderate: 34-39 Low: 40 or >

Table 8.8 demonstrates that the mean MBI subscale scores for the present study are

relatively consistent to the mean MBI subscale scores reported by nurses in foreign studies.

Several studies (e.g., Bakker et al., 2000; Edwards et al., 2000b; Iacovides et al., 1997;

Mallet et al., 1991; Turnipseed & Turnipseed, 1997) have also found that overall, nurses

report moderate levels of Emotional Exhaustion. There seems to be much greater variation in

the levels of Depersonalisation across nursing studies. Some nurses reported low levels of
 242

Depersonalisation (Payne et al., 2001) whilst others reported high levels of Depersonalisation

(Bakker et al., 2000; Mallet et al., 1991; Ogus, 1992; Schmitz et al., 2000). Other research

studies (Edwards et al., 2000b; Iacovides et al., 1997; Turnipseed & Turnipseed, 1997) have

found similar findings to the present study in that nurses reported moderate levels of

Depersonalisation. Finally, the majority of studies have found that nurses report moderate

levels of Personal Accomplishment (Edwards et al., 2000b; Iacovides et al., 1997; Schmitz et

al., 2000; Turnipseed & Turnipseed, 1997).

Comparison with Other Human Service Professions

Table 8.9 compares the present study’s mean MBI subscale scores to other human

service professions.

Table 8.9

Burnout Levels Across Occupational Fields

EE DP PA
Human Service
Profession N Mean SD N Mean SD N Mean SD
Present Study 273 23.01 11.17 273 7.75 5.89 273 34.62 7.83
Teaching 5481 28.15 11.99 5481 8.68 6.46 6191 36.35 7.41
Social Workers 628 20.82 10.17 628 6.94 5.60 628 35.15 7.80
Physicians 479 24.03 10.77 479 10.59 6.46 475 39.36 5.93
Psychologists 1382 19.75 9.77 1382 6.14 4.45 1382 40.94 6.07
Counsellors 422 20.52 8.97 422 6.64 4.26 422 35.35 6.75
Note. Data taken from The Burnout Companion to Study and Practice: A Critical Analysis (p.62), by W.

Schaufeli and D. Enzmann, 1998, London: Taylor & Francis. EE = Emotional Exhaustion; DP =

Depersonalisation; PA = Personal Accomplishment.

Range of Experienced Burnout

: EE – High: 27 or > Moderate: 19-26 Low: 0-18

: DP – High: 8 or > Moderate: 5-7 Low: 0-4
: PA – High: 0-33 Moderate: 34-39 Low: 40 or >

Table 8.9 indicates that when the sample’s mean level of burnout is compared to

other human service professions, nurses report moderate levels of Emotional Exhaustion,
 243

Depersonalisation and Personal Accomplishment. In comparison to physicians, the current

sample reported slightly less Emotional Exhaustion, much less Depersonalisation, and

substantially lower levels of Personal Accomplishment. In comparison to teachers, nurses

report much lower levels of Emotional Exhaustion, similar levels of Depersonalisation and

slightly lower levels of Personal Accomplishment. When compared to psychologists and

counsellors, nurses report higher levels of Emotional Exhaustion and Depersonalisation and

substantially lower levels of Personal Accomplishment.

Prevalence of Burnout

Table 8.10 presents the proportion of sample respondents that fall within the high,

moderate and low categories of burnout.

Table 8.10

Categorisation of the MBI Scores

Range of Experienced Burnout

Low Moderate High

MBI Subscale n % n % n %

Emotional Exhaustion 101 37.1 64 23.5 107 39.3

(EE)
Depersonalisation 120 44.6 60 22.3 89 33.1
(DP)*
Personal 85 31.3 73 26.8 114 41.9
Accomplishment (PA)*
Note. * Untransformed variables.

The frequency data presented in Table 8.10 indicate that a high proportion of nurses

are experiencing some aspect of burnout. For instance, approximately 40% of nurses are

reporting high levels of Emotional Exhaustion, the core dimension of burnout. In addition,

approximately 33% of nurses reported high levels of Depersonalisation. Furthermore,

approximately 42% of nurses reported low levels of Personal Accomplishment. The MBI
 244

defines high levels of burnout by high scores on Emotional Exhaustion and

Depersonalisation and low scores on Personal Accomplishment. Overall, 5.9% of the sample

population reported high levels of burnout.

Correlates of Burnout

Pearson’s product-moment correlations for the work stress variables and the burnout

variables in Study 3 are presented in Table 8.11. When interpreting the relationships between

the variables, it is important to keep in mind the transformed variables. The transformed job-

specific stress variables included Patient Care Uncertainty and Interpersonal Conflict. For

these variables, high scores represent high levels of job-specific stress. Role Ambiguity is the

only transformed role stress variable. A high total score on Role Ambiguity represents a high

level of Role Ambiguity. The burnout dimensions – Depersonalisation and Personal

Accomplishment are also transformed. A high total score on Depersonalisation relates to a

high level of Depersonalisation and a low total score reflects a low level of

Depersonalisation. A high total score on Personal Accomplishment relates to a low level of

Personal Accomplishment and a low total score relates to a high level of Personal

Accomplishment. Therefore, the work stress and burnout subscales are expected to have

positive correlations.
 245

Table 8.11

Correlations Between the Work Stress Subscales and the MBI Subscales

Recgn Uncert Conf JCon Ro Rc Ra Ee Dp Pa

Recgn 1.00
Uncert .47** 1.00
Conf .41** .22** 1.00
JCon .51** .54** .24** 1.00
Ro .37** .23** .24** .48** 1.00
Rc .55** .24 .30** .24** .33** 1.00
Ra .41** .10 .24** .15* .26** .61** 1.00
Ee .41** .36** .15* .49** .53** .43** .26** 1.00
Dp .30** .41** .08 .23** .21** .43** .26** .48** 1.00
Pa .10 .01 .03 .07 .07 .39** .37** .24** .29* 1.00
Note. Recogn = Lack of Professional Recognition and Support; Uncert = Patient Care Uncertainty; Conf =

Interpersonal Conflict; JCon = Job Conditions; Ro = Role Overload; Rc = Role Conflict; Ra = Role Ambiguity;

Ee = Emotional Exhaustion; Dp = Depersonalisation; Pa = Personal Accomplishment.

** Correlation is significant at the 0.01 level. * Correlation is significant at the 0.05 level.

Table 8.11 provides partial support for Hypothesis 4 that both job-specific stressors

and role stressors are related to burnout in nurses. All four job-specific stressors and all three

role stressors are significantly positively correlated to Emotional Exhaustion. Three job-

specific stressors and the three role stressors were significantly positively related to

Depersonalisation. Only Interpersonal Conflict was not related to Depersonalisation. Job-

specific stressors were not significantly related to Personal Accomplishment. Role Conflict

and Role Ambiguity were the only role stress variables significantly positively related to

Personal Accomplishment. In the section below, the relationships between the job-specific

variables and the three burnout components are explored in further detail.
 246

The Relationship Between Job-Specific Stressors and Burnout

All four job-specific stressors correlated positively and significantly with Emotional

Exhaustion. Specifically, Job Conditions was most strongly correlated with Emotional

Exhaustion with a correlation coefficient of .49 (p = .01), followed by Lack of Professional

Recognition and Support (r = .41, p = .01), Patient Care Uncertainty (r = .36, p = .01) and

Interpersonal Conflict (r = .15, p = .05). Patient Care Uncertainty was most strongly

correlated with Depersonalisation with a correlation coefficient of .41 (p = .01), followed by

Lack of Professional Recognition and Support (r = .30, p = .01) and Job Conditions (r = .23,

p = .01). All four job-specific stressors were related very weakly and positively to Personal

Accomplishment, however, the correlation coefficients were non-significant. Taken together,

these findings provided partial support for Hypothesis 5. As expected the job-specific

stressors were significantly and positively related to Emotional Exhaustion. Except for

Interpersonal Conflict, all job-specific sources were related to Depersonalisation. None of the

job-specific stressors however, were significantly positively related to Personal

Accomplishment.

The Relationships Between Role Stressors and Burnout

In accordance with Hypothesis 6, the three role stressors were positively and

significantly correlated with Emotional Exhaustion and Depersonalisation, suggesting that

high levels of role stress are associated with high levels of Emotional Exhaustion and

Depersonalisation. Specifically, Role Overload was most strongly correlated with Emotional

Exhaustion with a correlation coefficient of .53 (p = .01), followed by Role Conflict (r = .43,

p = .01) and Role Ambiguity (r = .26, p = .01). Role Conflict was most strongly correlated

with Depersonalisation with a correlation coefficient of .43 (p = .01), followed by Role

 247

Ambiguity (r = .26, p = .01) and Role Overload (r = .21, p = .01). Role Conflict and Role

Ambiguity were positively and significantly correlated to Personal Accomplishment with

correlation coefficients of .39 and .37 (p = .01) respectively. Role Overload correlated

weakly and positively with Personal Accomplishment, however the correlation coefficient

was non-significant. Overall, Hypothesis 6 is partially supported.

The Relationship Between Sociodemographic Variables and Burnout

Table 8.12 depicts the Pearson’s product-moment correlations for the

sociodemographic variables (employment status, age, gender, professional qualifications,

nursing experience and division of nursing) and the three burnout components – Emotional

Exhaustion, Depersonalisation, and Personal Accomplishment. Transformed variables for

Depersonalisation and Personal Accomplishment were used. A high total score on

Depersonalisation relates to a high level of Depersonalisation and a low total score reflects a

low level of Depersonalisation. A high total score on Personal Accomplishment relates to a

low level of Personal Accomplishment and a low total score relates to a high level of

Personal Accomplishment.
 248

Table 8.12

Pearson’s Product-Moment Correlations Between Burnout and the Sociodemographic

Variables

EE DP PA Emply Age Gender PQual Nursing Nursing

Status Exp Division
EE 1.00

DP .480** 1.00

PA .243** .292* 1.00

Emply -.137* -.067 .003 1.00

Status
Age -.138* -.321** .033 .032 1.00

Gender .007 -.091 -.003 .087 .105 1.00

Pqual .030 -.044 .035 -.011 -.089 -.055 1.00

Nursing -.085 -.228** .077 -.023 .673** .174** -.108* 1.00

Exp
Nursing .013 -.012 -.106 .089 .116 -.060 .125* .242** 1.00
Division
Note. EE = Emotional Exhaustion; DP = Depersonalisation; PA = Personal Accomplishment; Emply Status =

employment status; Pqual = Professional Qualifications; Nursing Exp = Nursing Experience in Years.

** Correlation is significant at 0.01 level. * Correlation is significant at the 0.05 level.

Table 8.12 demonstrates that overall, sociodemographic factors are not highly

correlated with burnout among nurses. For instance, only Age (r = -.14, p = .05) and

Employment Status (r = -.14, p = .05) are negatively but weakly correlated with Emotional

Exhaustion. More specifically, younger nurses report higher levels of Emotional Exhaustion

than older nursing staff. Nurses working full-time report higher levels of Emotional

Exhaustion, than part-time/casual nurses. Age (r = -.32, p = .01) and Nursing Experience (r

= -.23, p = .01) are moderately correlated to Depersonalisation. Younger, less experienced

nurses report higher levels of Depersonalisation than older, more experienced nurses.

Sociodemographic factors are not significantly related to Personal Accomplishment. The

 249

findings provide support for Hypothesis 7 that younger nurses report higher levels of burnout

than older nurses, and Hypothesis 8 that less experienced nurses report higher levels of

burnout than more experienced nurses.

Level of Social Support for Nurses Working at Public Hospitals

Total mean scores were generated for the main sources of work support. Table 8.13

depicts the average level of Coworker and Supervisor Support reported by the total sample

population. In addition, total mean scores for the work support scales are reported separately

for male and female nurses. The Coworker Support Scale and the Supervisor Support Scale

have not been transformed to ensure that comparisons between the sources of support can be

made.

Table 8.13

Mean Levels of Coworker and Supervisor Support

Total Males Females Test of

Sample Differences
(N = 273) (n = 38) (n = 235)

Source of Support Mean SD Mean SD Mean SD t p

Coworker Support 47.85 8.10 47.21 7.79 47.95 8.17 -.52 .60
(score ranges:12-60)

Supervisor Support 44.11 11.21 46.60 10.95 43.71 11.22 1.48 .14
(score ranges:12-60)

Note. Untransformed variables have been used.

Table 8.13 suggests that overall, the sample population report moderate levels of

Supervisor and Coworker Support. Nurses’ reported receiving higher levels of Coworker

Support than Supervisor Support, thus confirming Hypothesis 9. An Independent Samples t-

test revealed that there were no significant differences in nurses’ levels of Coworker Support

based on gender (t (269) = -.52, p = .60, two-tailed). Similarly, there were no significant
 250

differences in nurses’ levels of Supervisor Support based on gender (t (267) = 1.48, p = .14,

two-tailed).

The Relationship Between Social Support and Burnout

Table 8.14 depicts the Pearson’s product-moment correlations for the work

support variables (i.e., Emotional Coworker Support, Instrumental Coworker Support and

Supervisor Support) and the three burnout variables - Emotional Exhaustion,

Depersonalisation, and Personal Accomplishment. The burnout variables - Depersonalisation

and Personal Accomplishment are transformed. A high total score on Depersonalisation

relates to a high level of Depersonalisation and a low total score reflects a low level of

Depersonalisation. Similarly, a high total score on Personal Accomplishment relates to a low

level of Personal Accomplishment and a low total score relates to a high level of Personal

Accomplishment. The work support variables are also transformed. For the work support

scales, a high total score relates to low levels of support and a low total score relates to high

levels of work support.

 251

Table 8.14

Pearson’s Product-Moment Correlations Between the Work Support Scales

and the MBI Subscales

Ecs Ics Sups Ee Dp Pa

Ecs 1.00
Ics .69** 1.00
Sups .43** .39** 1.00
Ee .11 .24** .11 1.00
Dp .02 .06 .03 .48** 1.00
Pa .10 .16** .07 .24** .29** 1.00
Note. Ecs = Emotional Coworker Support; Ics = Instrumental Coworker Support;

SupS = Supervisor Support; Ee = Emotional Exhaustion; Dp = Depersonalisation;

Pa = Personal Accomplishment.

** Correlation is significant at the 0.01 level. * Correlation is significant at the 0.05 level.

Table 8.14 demonstrates that only Instrumental Coworker Support is significantly

related to nurses’ levels of burnout. Specifically, Instrumental Coworker Support was

significantly positively related to Emotional Exhaustion (r = .24, p = .01) and positively

related to Personal Accomplishment (r = .16, p = .01). This indicates that low levels of

instrumental assistance from coworkers is associated with higher levels of Emotional

Exhaustion. Furthermore, low levels of instrumental coworker support is associated with

lower levels of Personal Accomplishment. The remaining support variables were not related

to burnout. These findings provide partial support for Hypothesis 10 that work support will

be significantly negatively correlated to Emotional Exhaustion and Depersonalisation and

significantly positively related to Personal accomplishment.

 252

The Relationship Between Work Stressors and Social Support

Table 8.15 depicts the Pearson’s product-moment correlations for the work stress

variables and the work support variables. The job-specific stress variables Patient Care

Uncertainty and Interpersonal Conflict are transformed. For these variables, high scores

represent high levels of job-specific stress. The role stress variable Role Ambiguity is also

transformed. A high total score on Role Ambiguity represents a high level of Role

Ambiguity. The work support variables Emotional Coworker Support, Instrumental

Coworker Support, and Supervisor Support are all transformed. For the work support scales,

a high total score relates to low levels of support and a low total score relates to high levels

of work support.

Table 8.15

Pearson’s Product-Moment Correlations Between the Work Stress and Work Support

Variables

Recgn Uncert Conf JCon Ro Rc Ra Ecs Ics SupS

Recgn 1.00
Uncert .47** 1.00
Conf .41** .22** 1.00
JCon .51** .54** .24** 1.00
Ro .37** .23** .24** .48** 1.00
Rc .55** .24 .30** .24** .33** 1.00
Ra .41** .10 .24** .15* .26** .61** 1.00
Ecs .37** .08 .28** .08 .06 .34** .36** 1.00
Ics .41** .04 .24** .18** .23** .36** .40** .69** 1.00
SupS .41** .06 .28** .12* .17** .46** .55** .43** .39** 1.00
Note. Recogn = Lack of Professional Recognition and Support; Uncert = Patient Care Uncertainty; Conf =

Interpersonal Conflict; JCon = Job Conditions; Ro = Role Overload; Rc = Role Conflict; Ra = Role Ambiguity;

Ecs = Emotional Coworker Support; Ics = Instrumental Coworker Support; SupS = Supervisor Support.

** Correlation is significant at the 0.01 level. * Correlation is significant at the 0.05 level.
 253

Table 8.15 demonstrates that Emotional Coworker Support is significantly positively

correlated with Lack of Profession Recognition and Support (r = .37, p = .01) and

Interpersonal Conflict (r = .28, p = .01). This suggests that lower levels of Emotional

Coworker Support are associated with higher levels of job-specific stress. Emotional

Coworker Support was not significantly related to Patient Care Uncertainty (r = .08, p > .05)

or Job Conditions (r = .08, p > .05). Emotional Coworker Support is also significantly

positively correlated with Role Ambiguity (r = .36, p = .01) and Role Conflict (r = .34, p

=.01) indicating that lower levels of Emotional Coworker Support are associated with higher

levels of role stress. Emotional Coworker Support was not significantly related with Role

Overload (r = .06, p > .05).

Similarly, Instrumental Coworker Support is significantly positively correlated with

Lack of Profession Recognition and Support (r = .41, p = .01), Interpersonal Conflict (r =

.24, p = .01), and Job Conditions (r = .18, p = .01). This suggests that low levels of

Instrumental Coworker Support are associated with higher levels of job-specific stress.

Instrumental Coworker Support was not significantly correlated with Patient Care

Uncertainty (r = .04, p > .05)

Instrumental Coworker Support is also significantly positively correlated with Role

Ambiguity (r = .40, p = .01), Role Conflict (r = .36, p = .01) and Role Overload (r = .23, p =

.01) indicating that lower levels of Instrumental Coworker Support are associated with higher

levels of role stress. Instrumental Coworker Support is not significantly correlated with

Patient Care Uncertainty (r = .04, p > .05).

Supervisor Support influences stress in a similar manner to Coworker Support.

Supervisor Support is positively correlated with Lack of Professional Recognition and

 254

Support (r = .41, p = .01), Interpersonal Conflict (r = .28, p = .01), and Job Conditions (r =

.12, p = .05). This suggests that lower levels of Supervisor Support are associated with higher

levels of job-specific stress. Supervisor Support was not significantly correlated with Patient

Care Uncertainty (r = .06, p > .05). Supervisor Support is also significantly positively

correlated to Role Ambiguity (r = .55, p = .01), Role Conflict (r = .46, p = .01), and Role

Overload (r = .12, p = .05), indicating that lower levels of Supervisor Support are associated

with higher levels of role stress.

These findings provide partial support for Hypothesis 11 as not all work support

variables were significantly negatively correlated with work stressors. The results provide

supporting evidence that work support directly affects nurses’ perceptions of stress.

Supervisor Support is most strongly related to Role Ambiguity, indicating that the supervisor

plays an important role in reducing the ambiguity associated with the nursing role. Coworker

Support is most strongly related to Lack of Professional Recognition and Support. Finally,

Supervisor Support is more strongly correlated with work stressors than Coworker Support,

indicating that supervisors have a stronger influence in lowering nurses’ levels of perceived

work stress than their nursing colleagues. Interestingly, although the work support variables

(i.e., Emotional Coworker Support, Instrumental Coworker Support, and Supervisor Support)

are weakly and positively related to Patient Care Uncertainty, they are not significant. It

could therefore be concluded that work support has very little influence over how nurses

perceive the unpredictability and uncertainty associated with treating patients.

 255

Work Stress Determinants of Burnout

Standard multiple regression analyses were performed to identify the major

determinants of burnout. One equation was computed for each component of burnout. Only

the work stress variables that were significantly correlated to each burnout component were

entered into the equation.

Antecedents of Emotional Exhaustion

Using Emotional Exhaustion as the dependent variable, the four job-specific stressors

and the three role stressors were entered as the independent variables into the equation. One

outlier was removed. Transformed variables included Patient Care Uncertainty, Interpersonal

Conflict, and Role Ambiguity. When interpreting these variables, it is important to keep in

mind that high scores represent high levels of job-specific stress and low scores represent low

levels of job-specific stress.

Table 8.16

Standard Multiple Regression of Work Stress on Emotional Exhaustion

Variables Ee Recgn Uncer JCon Conf Ro Rc Ra B β

(DV)
Recgn .41** .048 .03
Uncer .36** .47** 1.181 .10
JExp .49** .51** .54** .407** .23
Conf .15* .41** .22** .24** -1.495 -.09
Ro .53** .37** .23** .48** .24** .563** .34
Rc .43** .55** .24** .24** .30** .33** .459** .27
Ra .26** .41** .10 .15* .24** .26** .61** -.650 -.04
Intercept = -.835
Mean 23.01 15.51 3.21 15.72 1.20 26.02 22.94 4.40
SD 11.18 7.80 0.98 6.28 0.67 6.82 6.58 0.62 R2 = .45
Adj R2 = .43
R = .67
Note. Ee = Emotional Exhaustion; Recgn = Lack of Professional Recognition and Support; Uncer = Patient

Care Uncertainty; JCon = Job Conditions; Conf = Interpersonal Conflict; Ro = Role Overload; Rc = Role

Conflict; Ra = Role Ambiguity.

** p < .001. * p < .05.

 256

The regression coefficients for the work stressors were significantly different from

zero [F(7, 262) = 30.20, p < .001]. Altogether, 43.2% of the variability in Emotional

Exhaustion was accounted for by the work stressors. Table 8.16 demonstrates that Role

Overload (β = .34, p < .001), Role Conflict (β = .27, p < .001) and Job Conditions (β = .23, p

< .001) are significant positive predictors of Emotional Exhaustion, with Role Overload

explaining most of the variance.

Antecedents of Depersonalisation

Using Depersonalisation as the dependent variable, all the job-specific variables

except Interpersonal Conflict, and the three role stressors were entered into the equation.

Interpersonal Conflict is not significantly related to Depersonalisation (see Table 8.17). Two

outliers were removed. Transformed variables included: Depersonalisation, Patient Care

Uncertainty, and Role Ambiguity. A high total score on Depersonalisation relates to a high

level of Depersonalisation and a low total score relates to a low level of Depersonalisation.

For Patient Care Uncertainty and Role Ambiguity, a high total score relates to a high level of

stress and a low total score represents a low level of stress.

 257

Table 8.17

Standard Multiple Regression of Work Stress on Depersonalisation

Variables DP Recgn Uncer JCon Ro Rc Ra B β

(DV)
Recgn .30** -.018 -.13
Uncer .41** .47** .462** .10
JCon .23** .51** .54** -.007 -.04
Ro .21** .37** .23** .48** .013 .08
Rc .43** .55** .24** .24** .33** .063** .37
Ra .26** .41** .10 .15* .26** .61** .124 .07
Intercept = -.856
Mean 2.54 15.51 3.21 15.72 26.02 22.94 4.40
SD 1.15 7.80 0.98 6.28 6.82 6.58 0.62 R2 = .33
Adj R2 = .31
R = .57
Note. Dp = Depersonalisation; Recgn = Lack of Professional Recognition and Support; Uncer = Patient Care

Uncertainty; JCon = Job Conditions; Ro = Role Overload; Rc = Role Conflict; Ra = Role Ambiguity.

* p < .05. ** p < .001.

The regression coefficients for the work stressors were significantly different from

zero [F(6,259) = 21.26, p < .001]. Altogether, 31% of the variability in Depersonalisation

was accounted for by the work stressors. Table 8.17 demonstrates that Patient Care

Uncertainty (β = .40, p < .001) and Role Conflict (β = .37, p < .001) are significant positive

predictors of Depersonalisation.

Antecedents of Personal Accomplishment

Using Personal Accomplishment as the dependent variable, Role Conflict and Role

Ambiguity were entered as the independent variables into the regression equation as these

were the only work stressors to correlate significantly with Personal Accomplishment (Refer

Table 8.18). There were no outliers. Transformed variables included Personal

Accomplishment, and Role Ambiguity. When interpreting the variable Personal

Accomplishment, it is important to note that a high total score relates to a low level of
 258

Personal Accomplishment and a low total score relates to a high level of Personal

Accomplishment. A high total score on Role Ambiguity represents a high level of Role

Ambiguity and a low total score represents a low level of Role Ambiguity.

Table 8.18
Standard Multiple Regression of Work Stress on Personal Accomplishment
Variables PA RC RA B β
(DV)
RC .39** .042** .26
RA .37** .61** .370** .21
Intercept = 1.04
Mean 3.63 22.94 4.40
SD 1.08 6.58 0.62 R2 = .18
Adj R2 = .17
R = .42
Note. PA = Personal Accomplishment; RC = Role Conflict; RA = Role Ambiguity.

** p < .001.

The regression coefficients for the role stressors were significantly different from

zero [F(2, 268) = 28.74, p < .001]. Altogether, 17% of the variability in Personal

Accomplishment was accounted for by the work stressors. Table 8.18 demonstrates that

Role Conflict (β = .26, p < .001), and Role Ambiguity (β = .21, p < .001) are significant

positive predictors of Personal Accomplishment, with Role Conflict explaining most of the

variance. More specifically, higher levels of Role Conflict and Role Ambiguity are related to

lower levels of Personal Accomplishment.

The Influence of Social Support on the Stress-Burnout Relationship

Hierarchical multiple regression analysis was conducted to examine the main and

interaction effects of social support on the stress-burnout relationship. A regression was

performed for each component of burnout. Using each burnout component as the dependent

variable, control variables and independent variables are entered into the regression equation.

Sociodemographic factors that were found to be significantly correlated to burnout were

 259

entered as the first step of the analyses. This enabled the researcher to control for the

potential confounding effects of sociodemographic factors. At Step 2, only the work stress

variables that were significant predictors of burnout were entered into the hierarchical

multiple regressions. In Step 3, the work support factors (Emotional Coworker Support,

Instrumental Coworker Support, and Supervisor Support) were entered into the equation.

Finally, Step 4 added the interaction terms. These were created by multiplying the work

stress factors by the work support factors. In line with recommendations for dealing with

problems of collinearity that arise from the use of cross-product terms (Aiken & West, 1991;

Jaccard, Turrisi, & Wan, 1990), variables were centred before calculating their cross-product

terms and conducting the analysis. Each regression equation was found to satisfy the

assumptions of multiple regression which includes normality, linearity, equality of variance

and the absence of collinearity.

The Influence of Work Support on Emotional Exhaustion

Using Emotional Exhaustion as the dependent variable, the control variables of Age

(29 years or less and 30 years or more) and Employment Status (full-time and part-

time/casual) were entered in the first step of the equation. Role Overload, Role Conflict and

Job Conditions were entered at the second step. In the third step, Emotional Coworker

Support, Instrumental Coworker Support and Supervisor Support were added. In the fourth

step, the interaction terms were entered. This four-step process is illustrated in Figure 8.3.

Transformed variables were used for each work support variable. A high total score on a

work support scale relates to a low level of support and a low total score relates to a high

level of work support.

 260

Step 1: Step 2: Step 3: Step 4:

Sociodemographic Work Stress Factors Work Support Interaction Terms
Factors Factors

• Age • Role Overload • Emotional • RO x ECS

• Employment • Role Conflict Coworker Support • RC x ECS
Status • Job Conditions • Instrumental • JC x ECS
Coworker Support
• Supervisor • RO x ICS
Support • RC x ICS
• JC x ICS

• RO x SUP
• RC x SUP
• JC x SUP

Figure 8.3. Four-step hierarchical regression process for Emotional Exhaustion.

Note. RO = Role Overload; RC = Role Conflict; JC = Job Conditions; ECS = Emotional Coworker Support;

ICS = Instrumental Coworker Support, SUP = Supervisor Support.

 261

Table 8.19

Hierarchical Regression Analysis for Emotional Exhaustion (Standardised Regression

Coefficients β)

Independent variables Step 1 Step 2 Step 3 Step 4

Age -.129* -.067 -.068 -.068

Employment Status -.128* -.040 -.032 -.025
Role Overload .327** .319** .323**
Role Conflict .227** .268** .257**
Job Conditions .278** .274** .278**
Emotional Coworker Support -.028 -.030
Instrumental Coworker Support .090 .067
Supervisor Support -.131* -.118*
Role Overload x ECS .105
Role Overload x ICS -.210*
Role Overload x Sup S -.019
Role Conflict x ECS -.028
Role Conflict x ICS .127
Role Conflict x Sup S -.051
Job Conditions x ECS -.004
Job Conditions x ICS .002
Job Conditions x Sup S .069

Constant -.092 -.092 -.089 -.297

F Change 4.59* 58.54** 2.25 .95
2 .034 .426 .441 .460
R
2 .027 .415 .423 .422
Adj R
2 .034 .392 .015 .019
R Change
Sig F Change .011 .000 .083 .484
Note. ECS = Emotional Coworker Support; ICS = Instrumental Coworker Support,

Sup S = Supervisor Support.

** p < .001. * p < .05.

Table 8.19 demonstrated that at Step 1, the sociodemographic variables were

significant negative predictors of Emotional Exhaustion. This indicates that younger nurses

(β = -.13, p < .05) report higher levels of Emotional Exhaustion than older nursing staff.

Furthermore, full-time nurses (β = -.13, p < .05) report higher levels of Emotional

Exhaustion than part-time/casual nurses. Step 1 accounted for a small, but significant 3.4% of

the variance in Emotional Exhaustion.

 262

At Step 2, the work-related stressors were significant positive predictors of Emotional

Exhaustion. It can be inferred that nurses reporting higher levels of Role Overload (β = .33, p

< .05), Job Conditions (β = .28, p < .05), and Role Conflict (β = .23, p < .05) have higher

levels of Emotional Exhaustion. Age and Employment Status were not significant predictors

of Emotional Exhaustion when the work stressors were added to the equation at Step 2. The

F change value indicated that R2 change was significant [F(3, 257) = 58.54, p < .001]. The R2

change coefficient indicated that the addition of the work stress variables significantly

improved the prediction of Emotional Exhaustion as the change associated with the step was

large. Adding the work stressors to the regression equation in Step 2 resulted in an R2 change

from R2 = .03 to R2 = .43 and contributed a significant 39.2% increment in the explained

variance.

At Step 3, examination of the standardised beta coefficients suggested that only

Supervisor Support was a significant predictor of Emotional Exhaustion (β = -.13, p < .05).

This indicates that nurses who are reporting higher levels of Supervisor Support are also

reporting lower levels of Emotional Exhaustion, and nurses who are reporting lower levels of

Supervisor Support are also reporting higher levels of Emotional Exhaustion. The work

stressors (i.e., Role Overload, Job Conditions and Role Conflict) remained significant at Step

3. The F change value, however, was not significant [F(3, 254) = 2.25, p > .05]. This

indicates that the addition of work support did not significantly improve the prediction of

Emotional Exhaustion as the change associated with this step was small. The inclusion of

work support variables in Step 3 resulted in a R2 change from R2 = .43 to R2 = .44 and

contributed to a slight, but non-significant, 1.5% increment in the explained variance.

 263

At Step 4, examination of the standardised beta coefficients revealed that the

interaction between Role Overload and Instrumental Coworker Support (β = -.21, p = .05)

was significantly related to Emotional Exhaustion. This suggests that for nurses with low

Instrumental Coworker Support, Emotional Exhaustion is higher when Role Overload is

high. For those nurses with high Instrumental Coworker Support, Emotional Exhaustion is

lower when Role Overload is high. This finding appears to be congruent with Cutrona and

Russell’s (1990) optimal matching theory as Role Overload is classified as a controllable

stressful event and therefore it was expected that Instrumental Coworker Support would be

the optimal type of support in preventing Emotional Exhaustion. The work stressors and

Supervisor Support also remained significant at Step 4. The F change value, however, was

not significant [F(9, 245) = .95, p > .05]. This indicates that the addition of interaction terms

did not significantly improve the prediction of Emotional Exhaustion as the change

associated with the step was small. Adding the interaction terms in Step 4 resulted in a R2

change from R2 = .44 to R2 = .46 and contributed a slight, but non-significant 2% increment

in the explained variance. The final model accounted for 42.2% of the explained variance in

Emotional Exhaustion.

The Influence of Work Support on Depersonalisation

Using Depersonalisation as the dependent variable, the control variables – Age (29

years or less and 30 years or more) and Nursing Experience (2 years or less and greater than

2 years), were entered at the first step. Patient Care Uncertainty and Role Conflict were

entered at the second step. In the third step, Emotional Coworker Support, Instrumental

Coworker Support, and Supervisor Support were added. In the fourth step, the interaction

terms were entered. This four-step process is depicted in Figure 8.4. Transformed variables
 264

included Depersonalisation, Patient Care Uncertainty, and the three work support variables.

A high total score on Depersonalisation relates to a high level of Depersonalisation and a low

total score reflects a low level of Depersonalisation. For Patient Care Uncertainty, high

scores represent high levels of Patient Care Uncertainty and low scores represent low levels

of Patient Care Uncertainty. In contrast, high total scores on the work support scales relate to

low levels of support and low total scores relate to high levels of support.

Step 1: Step 2: Step 3: Step 4:

Sociodemographic Work Stress Factors Work Support Factors Interaction Terms
Factors

• Age • Patient Care • Emotional • Pt Care x ECS

• Nursing Uncertainty Coworker Support • RC x ECS
• Role Conflict • Instrumental
Coworker Support • Pt Care x ICS
• Supervisor • RC x ICS
Support
• Pt Care x SUP
• RC x SUP

Figure 8.4. Four-step hierarchical regression process for Depersonalisation.

__________________________________________________________________________

Note. ECS = Emotional Coworker Support; ICS = Instrumental Coworker Support; SUP = Supervisor Support;

Pt Care = Patient Care Uncertainty; RC = Role Conflict.

 265

Table 8.20

Hierarchical Regression Analysis for Depersonalisation (Standardised Regression

Coefficients β)

Independent variables Step 1 Step 2 Step 3 Step 4

Age -.314** -.171* -.152* -.150*
Nursing Experience -.037 -.037 -.031 -.037
Patient Care Uncertainty .289** .289** .278**
Role Conflict .333** .423** .432**
ECS -.093 -.095
ICS .034 .043
Sup S -.148* -.142*
Pt Care Uncert x ECS .033
Pt Care Uncert x ICS .028
Pt Care Uncert x Sup S -.082
Role Conflict x ECS .105
Role Conflict x ICS -.007
Role Conflict x Sup S -.036

Constant .124 .124 .105 .146

F Change 17.01** 43.08** 3.62** .77
2 .115 .336 .363 .375
R
2 .109 .326 .346 .342
Adj R
2 .115 .221 .027 .012
R Change
Sig F Change .000 .000 .014 .593
Note. ECS = Emotional Coworker Support; ICS = Instrumental Coworker Support,

Sup S = Supervisor Support; Pt Care Uncert = Patient Care Uncertainty.

** p < .001. * p < .05.

Table 8.20 demonstrated that at Step 1, only Age was a significant negative predictor

of Depersonalisation. This indicates that younger nurses (β = -.31, p < .001) report higher

levels of Depersonalisation than older nursing staff. Step 1 accounted for a significant 11.5%

of the explained variance in Depersonalisation.

At Step 2, the work-related stressors were significant positive predictors of

Depersonalisation. It can be inferred that nurses reporting higher levels of Role Conflict (β =

.33, p < .01) and Patient Care Uncertainty (β = .29, p < .01) have higher levels of

Depersonalisation. Age remained a significant predictor of Depersonalisation when the work

 266

stressors were added to the equation at Step 2. The F change value indicated that R2 change

was significant [F(2, 261) = 43.08, p < .001]. The R2 change coefficient indicated that the

addition of the work stress variables significantly improved the prediction of

Depersonalisation as the change associated with the step was large. Adding the work

stressors to the regression equation in Step 2 resulted in an R2 change from R2 = .12 to R2 =

.34 and contributed a significant 22.1% increment in the explained variance.

At Step 3, only Supervisor Support was a significant predictor of Depersonalisation

(β = -.15, p < .05). The main effect for Supervisor Support on Depersonalisation is depicted

in Figure 8.5. Age, Role Conflict, and Patient Care Uncertainty remained significant at Step

3. The F change value indicated that R2 change was significant [F(3, 256) = 3.62, p < .05].

The R2 change coefficient indicated that the addition of work support did significantly

improve the prediction of Depersonalisation. The inclusion of work support variables in Step

3 resulted in an R2 change from R2 = .34 to R2 = .36 and contributed to a slight, but

significant, 2.7% increment in the explained variance.

At Step 4, the interaction terms were not significantly related to Depersonalisation.

Age, Role Conflict, Patient Care Uncertainty, and Supervisor Support remained significant at

Step 4. The F change value indicated that R2 change was not significant [F(6, 250) = .77, p >

.05]. The R2 change coefficient indicated that the addition of interaction terms did not

significantly improve the prediction of Depersonalisation. The R2 change from R2 = .36 to R2

= .38 in Step 4 suggested that the introduction of the interaction terms added a small, but

non-significant, 2% in explained variance. The final model accounted for 34.2% of the

explained variance in Depersonalisation.

 267

0.15

0.1

0.05
Depersonalisation

0 Depersonalisation

-0.05

-0.1

-0.15
High Low
Supervisor Support

Figure 8.5. Main effect of Supervisor Support on Depersonalisation.

Figure 8.5 suggests that lower levels of Supervisor Support are associated with higher

levels of Depersonalisation and higher levels of Supervisor Support are associated with lower

levels of Depersonalisation. This finding provides support for the main effect model of social

support. The main effect model suggests that when social support is high, burnout is lower

than when social support is low, regardless of the level of work stress (Kessler & McLeod,

1985).

The Influence of Work Support on Personal Accomplishment

Using Personal Accomplishment as the dependent variable, the variables were

entered in three steps. Since there were no significant correlations between Personal

Accomplishment and the sociodemographic factors, none were entered into the hierarchical

regression equation. The role stressors Role Conflict and Role Ambiguity were entered at the

first step. Emotional Coworker Support, Instrumental Coworker Support, and Supervisor
 268

Support were added at the second step. Finally, the interaction terms were entered. This

three-step process is presented in Figure 8.6. Transformed variables included: Personal

Accomplishment, Role Ambiguity and the three work support variables. When interpreting

the variable Personal Accomplishment, it is important to note that a high total score on

Personal Accomplishment relates to a low level of Personal Accomplishment and a low total

score relates to a high level of Personal Accomplishment. A high total score on Role

Ambiguity represents a high level of Role Ambiguity and a low total score represents a low

level of Role Ambiguity. High total scores on the work support scales relate to low levels of

support and low total scores relate to high levels of support.

Step 1: Step 2: Step 3:

Work Stress Factors Work Support Factors Interaction Terms

• Role Conflict • Emotional • RC x ECS

• Role Ambiguity Coworker Support • RA x ECS
• Instrumental
Coworker Support • RC x ICS
• Supervisor • RA x ICS
Support
• RC x SUP
• RA x SUP

Figure 8.6. Three-step hierarchical regression process for Personal Accomplishment.

__________________________________________________________________________

Note. ECS = Emotional Coworker Support; ICS = Instrumental Coworker Support; SUP = Supervisor Support;

RC = Role Conflict; RA = Role Ambiguity.

 269

Table 8.21

Hierarchical Regression Analysis for Personal Accomplishment (Standardised Regression

Coefficients β)

Independent variables Step 1 Step 2 Step 3

Role Conflict .256** .300** .322**

Role Ambiguity .211** .312** .313**
ECS -.057 -.081
ICS .064 .082
Sup S -.241** -.245**
Role Conflict x ECS -.025
Role Conflict x ICS .039
Role Conflict x Sup S .051
Role Ambiguity x ECS -.048
Role Ambiguity x ICS .57
Role Ambiguity x Sup S .-.098

Constant .00 .00 .07

F Change 28.41** 4.70** 1.06
2 .177 .219 .238
R
2 .170 .204 .205
Adj R
2 .177 .042 .019
R Change
F Sig Change .000 .003 .390
Note. ECS = Emotional Coworker Support; ICS = Instrumental Coworker Support,

Sup S = Supervisor Support.

** p < .001.

Table 8.21 demonstrated that at Step 1, the role stressors were significant positive

predictors of Personal Accomplishment. This suggests that nurses reporting higher levels of

Role Conflict (β = .26, p < .01) and Role Ambiguity (β = .21, p < .01) also have lower levels

of Personal Accomplishment. Step 1 accounted for a significant 17.7% of the variance in

Personal Accomplishment.

At Step 2, only Supervisor Support was a significant negative predictor of Personal

Accomplishment (β = -.24, p < .01). The main effect for Supervisor Support on Personal

Accomplishment is illustrated in Figure 8.7. The work stressors (i.e., Role Conflict and Role

Ambiguity) remained significant at Step 2. The F change value indicated that the R2 change
 270

was significant [F(3, 262) = 4.70, p < .01]. The R2 change coefficient indicated that the

addition of work support did significantly improve the prediction of Personal

Accomplishment. The inclusion of the work support variables in Step 2 resulted in an R2

change from R2 = .18 to R2 = .22 and contributed to a slight, but significant, 4% increment in

the explained variance.

At Step 3, the interaction terms were not significantly related to Personal

Accomplishment. Role Conflict, Role Ambiguity and Supervisor Support remained

significant at Step 4. The F change value indicated that R2 change was not significant [F(6,

256) = 1.05, p > .05]. The R2 change coefficient indicated that the addition of interaction

terms did not significantly improve the prediction of Personal Accomplishment. The R2

change from R2 = .22 to R2 = .24 in Step 4 suggested that the interaction terms added a small,

but non-significant, 2% in explained variance. The final model accounted for 20.5% of the

explained variance in Personal Accomplishment.

 271

0.15

0.1
Personal Accomplishment

0.05

Personal
0
Accomplishment

-0.05

-0.1

-0.15
High Sup Low Sup
Supervisor Support

Figure 8.7. Main effect of Supervisor Support on Personal Accomplishment

___________________________________________________________________________

Note. A high total score on Personal Accomplishment relates to a low level of Personal Accomplishment. A low

total score on Personal Accomplishment relates to a high level of Personal Accomplishment.

Figure 8.7 suggests that higher levels of Supervisor Support are associated with lower

levels of higher levels of Personal Accomplishment. Lower levels of Supervisor Support are

associated with lower levels of Personal Accomplishment. This finding provides further

support for the main effect model of social support.

Overall, work support did not have a significant buffering on burnout and therefore

Hypothesis 12 is rejected. More specifically, no significant buffering effects were found for

work support on any of the three burnout dimensions. The findings did reveal, however, that

Supervisor Support has a significant main effect on Depersonalisation and Personal

Accomplishment, irrespective of the level of work stress. Emotional Coworker Support and

Instrumental Coworker Support had no direct effects on burnout. Although there was some
 272

indication that Role Overload interacts with Instrumental Coworker Support to buffer

Emotional Exhaustion, the incremental change in R2 for the interaction term was not

significant. Taken together, the findings do not provide any support for Hypothesis 13 that

for uncontrollable stressful events, emotional support has a significant buffering effect on

burnout. Similarly, the findings do not provide any support for Hypothesis 14 that for

controllable stressful events, instrumental support has a significant buffering effect on

burnout.

Discussion

The aim of this study was to examine the relationships among work stress, social

support, and burnout in nurses. In the first instance, Study 3 was designed to examine the

nature and level of work stress and burnout among Australian nurses working in the public

hospital system. Specifically, the average level of work stress and the major sources of

nurses’ stress were identified. The prevalence of burnout in the sample population was

estimated and then compared to appropriate normative groups in an attempt to verify whether

Australian nurses experience similar levels of burnout to other nurses and human service

professionals. Second, the level of support nurses receive from their coworkers and their

supervisor was established. Third, the research was designed to investigate the main, indirect

and buffering effects of social support on nurses’ burnout levels. The principal purpose of

Study 3 was to identify conditions where work support is optimally effective in reducing or

preventing burnout in nurses. In accordance with Cohen and Wills (1985) stressor-support

matching theory, different sources and types of work support were matched to different work

stress factors in an attempt to discover optimal conditions in which work support buffers

burnout.
 273

In addition, the thesis extends the stressor-support matching theory by taking into

consideration a nurse’s situational control during stressful events. Cutrona and Russell’s

(1990) optimal matching theory asserts that when a stressful situation is primarily

controllable, the most beneficial support is instrumental. However, when a stressful situation

is essentially uncontrollable, the most effective type of support is emotional. In the present

study, independent raters classified nurses’ main sources of stress as predominantly

controllable. Only the variable Job Conditions was identified as an uncontrollable stressor. It

was therefore inferred that overall, instrumental support would be more effective than

emotional support in reducing or preventing burnout among nurses. However, for the work

stressor – Job Conditions, emotional support would be more effective than instrumental

support in reducing or preventing burnout. It should be noted, however, that the exploration

of different types of support was confined to coworkers only. A limitation of the present

study is that the sample population was unable to clearly distinguish between emotional and

instrumental support provided by their supervisor. As a result, Supervisor Support was

examined as a global construct in the current research. In the section below, Study 3’s main

findings are discussed in detail.

Work Stress Among Hospital Nurses

Despite numerous recent studies (e.g., Bakker et al., 2000; Healy & McKay, 1999;

Kirkaldy & Martin, 2000; Pikó, 1999) indicating that nursing is, by its very nature, an

occupation subject to a high degree of stress, nurses in the current study reported overall,

moderately low levels of job-specific stress and ‘normal’ levels of role stress. Unfortunately,

it is not possible to make comparisons between nursing studies on the prevalence of stress

given that different research methodologies, nursing samples, and sample sizes have been
 274

used. Furthermore, the majority of nursing studies have focused their research on sources of

stress rather than levels of stress. Despite this, other empirical studies have also found that

overall nurses perceive their jobs to be moderately stressful (e.g., DePew et al., 1999; Lee,

2003; Wheeler & Riding, 1994). Perhaps nurses are relatively resilient to occupational stress.

According to Peeters and Le Blanc (2001), nurses perceive most work stressors to be

indissolubly connected to their profession, and as a result they do not perceive them to be

particularly significant.

The present findings provide support for Peeters and Le Blanc’s (2001) assertion. It

was demonstrated that although stressful conditions are prevalent in the workplace, only

certain events serve as a significant source of stress to nurses on a regular basis. Australian

nurses’ main sources of stress are linked to work overload pressures due to insufficient time

and resources, role conflict pressures due to multiple task demands, pressure due to lack of

involvement in planning and decision-making, pressure related to insufficient pay, pressure

due to insufficient performance feedback, and to a lesser extent, pressures associated with

difficult/demanding patients.

While some causes of stress are related to the caring element of nursing (e.g., dealing

with difficult patients, being responsible for a patient’s recovery), the majority of nurses’

stress results from organisational factors, especially management’s unrealistic expectations of

nurses. The current study revealed that the nursing role is characterised by heavy workloads,

constant interruptions, multiple task demands and inadequate staff support. Runciman (1983)

found that although interruptions are rarely unnecessary, this fragmented approach to work is

a great source of frustration to nurses. Nurses contend with demands from many people (e.g.,

doctors, supervisors, patients, relatives) some of which are incompatible and have been found
 275

to cause role conflict in nurses (McGrath et al., 1989). According to McGrath et al., there is

sometimes uncertainty related to responsibilities, limits of authority, and supervisors’

expectations. Ambiguity about expectations and success in meeting such expectations can

result from inadequate performance feedback from nursing supervisors. In addition, nurses

have little control over their work conditions as they are not in a position to provide input

into important matters and make decisions concerning the functioning of the hospital. It is

little wonder that nurses feel that they are inadequately paid for their efforts. The majority of

nurses today are university trained and often undergo further education in order to specialise

in a specific nursing area. They also regularly engage in extra training to keep up with recent

developments in treating diseases and operating new medical equipment. Drawing from an

effort-reward imbalance theoretical perspective (Siegrist, 1996), it could be inferred that

there is a lack of reciprocity between the costs and gains associated with the nursing role. For

instance, nurses gain poor financial rewards for high levels of effort. In 1997, over two-thirds

(69.8%) of employed Australian nurses fell into the salary range of $28,730 to $41,992, and

the next largest proportion (16.4%) is in the range $41,993 to $50,471 (AIHW, 2002).

Recent findings in the literature support the current study’s findings regarding nurses’

main work stressors. In particular, the majority of studies have also identified excessive

workload as the most salient cause of workplace stress among nurses (Bryant et al., 2000;

Hillhouse & Adler, 1997; Healy & McKay, 2000; Riding & Wheeler, 1995a; Taylor, White,

& Muncer, 1999; Tyler & Cushway, 1995; Wheeler & Riding, 1994). It would seem that

nurses are frequently required to do more work with fewer resources (e.g., staff support).

Wheeler (1998) concluded that this often repeated finding suggests one of two possibilities,

or a combination of these possibilities:

 276

1. Interested parties and those responsible for addressing and finding a

solution to this problem are paying no notice to the research findings.

2. If interventions are being instituted, they are failing to make a positive

impact.

Other studies have also identified lack of involvement in decision-making (Harris,

1989), dealing with difficult patients (Bakker et al., 2000; Lewinson et al., 1981), insufficient

pay (Chung & Corbett, 1998; McGrath et al., 1989), role conflict (Cordes et al., 1997), and

role ambiguity (Edwards et al., 2000a; Gil-Monte et al., 1995; Revicki & May, 1989) to be

sources of significant stress.

Unlike previous studies, however, the current nursing sample did not reveal ‘dealing

with patient death and dying’, and ‘interpersonal conflict’ as major work stressors. Perhaps

the discrepancy in findings may be explained by the researcher’s decision not to employ the

commonly used NSS (Gray-Toft & Anderson, 1981a) to measure nurses’ work stress.

Alternatively, Lazarus and Folkman (1984) suggest that the degree of stress specific events

elicit in a person depends in part on the cognitive appraisal of that demand. It could be

asserted then that dealing with patient death and dying and interpersonal conflict is not

considered stressful by Australian nurses as they have sufficient resources to cope with these

demands.

The current study found no significant differences between nurses’ levels of work

stress based on gender, thus confirming Hypothesis 1. Other studies have also found male

and female nurses’ perceptions of stress do not differ significantly (Deaux, 1984; Greenglass,

1991; Kirkcaldy & Martin, 2000; Martocchio & O’Leary, 1989; Tyler & Ellison, 1994).

According to Frone, Russell, and Cooper (1995), the differences between men and women on
 277

job stress are typically small. Folkman and Lazarus (1980) and Maslach (1982) found that

men and women differ very little in the way they appraise potentially stressful events.

However, it is important to note that the sample size for male nurses in the present study was

typically small and any variation may not be evident from the data.

In summary, the sample of Australian nurses in the present study reported moderately

low levels of job-specific stress and average levels of role stress. Gender had no significant

impact on nurses’ perceptions of work stress, with male and female nurses reporting similar

levels of work stress. Work stress is predominantly related to an excessive workload and

insufficient staffing resources to cope with organisational demands.

Burnout – A Multidimensional Construct

The study provided supporting evidence for Maslach et al.’s (1996) assertion that

burnout consists of three distinct, but conceptually related components, thus confirming

Hypothesis 2. The present study demonstrated that Emotional Exhaustion is moderately

positively related to Depersonalisation. Both Emotional Exhaustion and Depersonalisation

are positively, but less strongly related to reduced Personal Accomplishment. This finding

indicates that high levels of Emotional Exhaustion and Depersonalisation are associated with

low levels of Personal Accomplishment. The low to moderate correlations between these

variables suggests that the variables are sufficiently distinct.

Level of Burnout among Nurses

Overall, the sample population reported moderate levels of Emotional Exhaustion,

Depersonalisation and Personal Accomplishment, thus confirming Hypothesis 3. More

specifically, the sample of Australian nurses reported moderate levels of Emotional

Exhaustion, moderately high levels of Depersonalisation, and moderately low levels of

 278

Personal Accomplishment. Overall 5.9% of the sample population reported high levels of

Emotional Exhaustion and Depersonalisation, and low levels of Personal Accomplishment.

When the mean burnout levels for Australian nurses were compared to other recent

nursing studies, similar findings were reported. Foreign studies in Greece (Iacovides et al.,

1997), Germany (Bakker et al., 2000), Poland (Schaufeli & Janczur 1994), and the United

States (Turnipseed & Turnipseed, 1997) also found that overall, hospital nurses report

moderate levels of burnout.

When compared to other human service professions, the sample reported lower levels

of Emotional Exhaustion and Depersonalisation than physicians and teachers, and lower

levels of Personal Accomplishment. In comparison to psychologists and counsellors

however, nurses report significantly higher levels of Emotional Exhaustion and

Depersonalisation and substantially lower levels of Personal Accomplishment. It could

therefore be inferred that nurses tend to evaluate their accomplishments more negatively than

other human service professions.

In the present research, a substantial proportion of nurses reported experiencing

different aspects of burnout. For instance, 40% of nurses indicated high levels of Emotional

Exhaustion. This suggests that a large percentage of nurses often feel overextended and

depleted of one’s emotional resources. In addition, approximately 30% of nurses reported

high levels of Depersonalisation, suggesting an excessive detachment from other people.

However, the main manifestation of burnout among the current sample was related to

feelings of diminished Personal Accomplishment (42%). This denotes a considerable decline

in one’s feelings of professional competence and successful achievement in one’s work.

According to McGrath et al., lower levels of Personal Accomplishment may be more likely
 279

in a profession which attracts those with idealism which is not realised in practice. It was

somewhat unexpected that nurses are reporting slightly lower levels of Personal

Accomplishment than Emotional Exhaustion and Depersonalisation. This finding may cast

some doubt on the sequential development of burnout and may point to the need to revisit the

burnout process in the future. It is possible that Personal Accomplishment and Emotional

Exhaustion develop in parallel to one another as proposed by Leiter (1993) and Lee and

Ashforth (1996).

Taking into consideration the above findings, it could be concluded that nurses are

indeed at risk of experiencing burnout. Overall, nurses are reporting moderate levels of

burnout. In addition, approximately 6% of the sample population reported experiencing high

levels of burnout (i.e., high levels of Emotional Exhaustion and Depersonalisation and low

levels of Personal Accomplishment). Furthermore, a considerable proportion of the sample

population reported experiencing a particular aspect of burnout. This result was surprising

since the present nursing sample reported moderately low levels of job stress. It may be that

nurses underestimate the effects stressful conditions have on their personal well-being. It is

also plausible that nurses overestimate the degree to which they have control over certain

events. Work stressors that are thought to be controllable are generally considered less

distressing than those that are not perceived to be under an individual’s control (Aldwin,

1994). In the present study, nurses’ work stressors were classified as predominantly

controllable. Research has demonstrated that people both feel and perform better when

perceived control is high. Perceived control appears to enhance confidence, making tasks less

stressful (Parker & Price, 1994). Additionally, perceived control is seen to increase

concentration, task persistence, commitment involvement, motivation and performance

 280

(Parker & Price, 1994). High levels of situational control may motivate nurses to continue to

work hard in the face of difficult circumstances (Ross & Altmaier, 1994). However, when

nurses expend a high degree of effort, they also become more susceptible to burnout. Over

time, nurses may begin to feel emotionally depleted, the first major symptom of burnout.

Correlates of Burnout

In Study 3, the main correlates of burnout were examined. In particular, the

relationships between work stress and burnout, as well as the relationships between

sociodemographic factors and burnout were explored.

Relationships between work stress and burnout. The results provided partial support

for Hypothesis 4 in that both job-specific stressors and role stressors were related to burnout

in nurses. The findings demonstrated that high levels of work stress were associated with

high levels of Emotional Exhaustion. Moderately high positive correlations were found

between the three role stressors and Emotional Exhaustion, with correlation coefficients

ranging from .26 to .53. In addition, all four job-specific stressors were related to Emotional

Exhaustion, with correlation coefficients ranging from .15 to .49. Overall, Role Overload was

most strongly correlated to Emotional Exhaustion, followed by Job Conditions, and Role

Conflict.

Similarly, a number of work stressors are related to Depersonalisation. Moderate

positive correlations were found between the three role stressors and Depersonalisation.

Correlation coefficients ranged from .21 to .43. All the job-specific stressors except for

Interpersonal Conflict were positively associated with Depersonalisation, with correlation

coefficients ranging from .23 to .41. Role Conflict was most strongly correlated to
 281

Depersonalisation, followed by Patient Care Uncertainty, and Lack of Professional

Recognition and Support.

Few work stressors, however, are related to reduced Personal Accomplishment.

Modest positive correlations were found for Personal Accomplishment and the role stressors

Role Conflict (r = .39) and Role Ambiguity (r = .37). Job-specific stressors however, were

not significantly related to Personal Accomplishment.

The above findings demonstrate that work stress is significantly related to burnout

among nurses. Not all job-specific stressors however, were significantly and positively

related to Emotional Exhaustion and Depersonalisation and significantly negatively related to

Personal Accomplishment. Similarly, not all role stressors were significantly and positively

related to Emotional Exhaustion and Depersonalisation and significantly negatively related to

Personal Accomplishment. Thus Hypothesis 5 and 6 were only partially supported.

It was also shown that Emotional Exhaustion and Depersonalisation have more

robust relationships with the work stressors than Personal Accomplishment. This finding

provides support for Shirom’s (1989) view that Emotional Exhaustion and Depersonalisation

are the core components of burnout. Furthermore, role stressors seem to be more strongly

correlated with burnout, than job-specific stressors. It could be assumed from the present

study’s findings, that nurses do not perceive work stressors that are highly specific to the

nursing profession to be as relevant or salient in causing strain. This result contradicted

Beehr et al.’s (2000) finding that job-specific stressors are more strongly related to

psychological strain than generic role stressors. Finally, the results reveal that work stressors

are differentially related to the three burnout components. This finding provides further

support to the argument that burnout is a multidimensional construct and its linkages to
 282

important variables should be tested (Byrne, 1994; Gmelch & Gates, 1998; Maslach et al.,

1996).

Relationships between sociodemographic factors and burnout. The results confirmed

that Age, Employment Status, and Nursing Experience are significantly related to burnout

among nurses. Age correlated negatively with Emotional Exhaustion (r = -.14) and

Depersonalisation (r = -.32), with younger nurses reporting higher levels of Emotional

Exhaustion and Depersonalisation than older nursing staff, thus confirming Hypothesis 7. In

addition, less experienced nurses report higher levels of Depersonalisation (r = -.23) than

more experienced nurses, thus confirming Hypothesis 8.

It could be inferred that older, more experienced nurses have developed more

effective ways to cope with work stress and therefore report lower levels of burnout (Hinds,

Quargnenti, Hickey, & Magnum, 1994). Schaufeli (1999) inferred that the greater incidence

of burnout among younger, less experienced nurses might be caused by a ‘reality shock’ or

by an identity crisis due to unsuccessful occupational socialisation. Furthermore, nurses early

in their careers are faced with the demands of clinical contact with patients as well as the

need to quickly learn practical clinical skills (Baldwin, 1999). Alternatively, a survival bias

cannot be ruled out such that those who ‘burn out’ early in their careers are likely to quit,

leaving behind the survivors who exhibit lower levels of burnout (Schaufeli, 1999).

Furthermore, nurses working on a full-time basis report higher levels of Emotional

Exhaustion than nurses working on a part-time or casual basis. Intuitively this makes sense as

full-time nurses are exposed to the same stressful situations more frequently (i.e., daily) and

for a longer duration than part-time/casual nurses.

 283

Levels of Work Support

Nurses reported receiving moderate levels of support from within the workplace.

More specifically, nurses reported receiving moderately low levels of Supervisor Support and

moderately high levels of Coworker Support. The results supported Hypothesis 9 that nurses

report higher levels of Coworker Support than Supervisor Support. The qualitative findings

from Study 1 also suggested that the primary source of support for nurses is their coworkers.

The sample of respondents in Study 1 indicated that they share their problems with their

nursing colleagues because they identify with the problem and that they can diffuse a

stressful situation through comfort and sometimes humour.

The Influence of Work Support on the Stress-Burnout Relationship

The influence of work support on nurses’ levels of burnout and nurses’ levels of work

stress were explored using Pearson’s product-moment correlation coefficients.

Relationships between work support and burnout. Hypothesis 10 stated that work

support would be significantly negatively correlated to Emotional Exhaustion and

Depersonalisation and significantly positively related to Personal Accomplishment. The

relationship between work support and burnout, however, did not appear to be very strong.

Only Instrumental Coworker Support was significantly related to nurses’ levels of burnout.

Specifically, lower levels of Instrumental Coworker Support was related to higher levels of

Emotional Exhaustion. Similarly, lower levels of Instrumental Coworker Support was

associated with lower levels of Personal Accomplishment. The remaining work support

variables were not significantly related to the burnout dimensions. Thus, only partial support

for Hypothesis 10 was found.

 284

Relationships between work support and work stress. Partial support was found for

Hypothesis 11, that work support would be significantly negatively correlated with work

stressors. Modest positive correlations were found for Emotional Coworker Support and

work stress, with correlation coefficients ranging from .27 to .37. Instrumental Coworker

Support also had modest positive relations with work stress, with correlation coefficients

ranging from .18 to .41. These findings suggest that support from coworkers reduces the

perceived threat or the appraised stressfulness of various work events. Supervisor Support

also had strong positive correlations with work stress, with correlation coefficients ranging

from .12 to .55. The findings indicated that supervisors have a stronger influence in reducing

nurses’ perceptions of work stress than coworkers. While Coworker Support was most

strongly related to Lack of Professional Recognition and Support, Supervisor Support was

most strongly related to Role Ambiguity. Interestingly, work support was not significantly

correlated with Patient Care Uncertainty. It could therefore be concluded that work support

has very little influence in how nurses perceive the unpredictability and uncertainty

associated with treating patients. Overall, the findings supported Cohen and Wills (1985)

assertion that work support reduces the likelihood that work events are perceived as highly

stressful (Kong & Wertheimer, 1994).

Main Predictors of Burnout

For Emotional Exhaustion, predictor variables accounted for 42.2% of the total

variance. Sociodemographic factors explained a small, but significant proportion of the

variance (2.7%). Specifically, younger nurses reported higher levels of Emotional Exhaustion

than older nurses. Furthermore, nurses working full-time reported higher levels of Emotional

Exhaustion than nurses working part-time or casually. Work stressors, however, were the
 285

main predictors of Emotional Exhaustion, explaining 41.5% of the total variance. Role

Overload, Job Conditions, and Role Conflict were the main determinants of Emotional

Exhaustion, with Role Overload explaining most of the variance.

This finding suggests that Emotional Exhaustion is strongly associated with work

demands that directly increase the amount of effort needed to do the job. Support for this

proposition is provided by Cordes et al. (1997). They inferred that individuals with

insufficient time and resources to undertake job duties might expend an excessive amount of

emotional energy in accomplishing role requirements, which over time, may lead to

emotional exhaustion. For instance, if nurses are given extra patients to attend to, or extra

administration work to complete, as well as their normal duties, extra effort is required to

perform their job satisfactorily, which in time may result in emotional exhaustion.

This explanation also seems relevant in explaining the effects of nurses’ Job

Conditions on Emotional Exhaustion. Nurses are regularly exposed to stressful conditions at

work, such as caring for patients, managing a heavy work load, supervising less experienced

nursing staff, keeping up with new developments, and trying to meet the general public’s

expectations for high quality medical care. It could be assumed that nurses must consistently

maintain a high level of effort in order to meet the everyday demands of their job.

Similarly, high levels of Role Conflict imply multiple sources of demands, whereby

high levels of effort are necessary to satisfy those demands. According to Cash (1991),

nurses are educated to believe that their primary responsibility is to meet the needs of the

individual patient, but as employees they are also expected to meet the needs of their

organisation. The conflict between meeting these demands would undoubtedly increase

Emotional Exhaustion (Gil-Monte et al., 1995). Gil-Monte et al. added that nursing work
 286

requires warmth and sympathy, together with objectivity and assertiveness, which may also

result in inter-role conflict. The consequent effort for dealing with this conflict could be

related to the depletion of one’s emotional resources.

For Depersonalisation, the predictor variables accounted for 34.2% of the total

variance. Sociodemographic factors (11.5%) and work stressors (33.6%) both explained a

significant proportion of the variance in Depersonalisation. Work stressors, however, are the

strongest determinants of Depersonalisation. As was the case for Emotional Exhaustion,

younger nurses report higher levels of Depersonalisation than older nurses.

Role Conflict and Patient Care Uncertainty were the main determinants of

Depersonalisation, with Role Conflict explaining more of the variance. Similar findings are

reported in the stress literature. For instance, Cash (1991) and Gil-Monte et al. (1995) also

found Role Conflict to be a significant predictor of Depersonalisation. These researchers

proposed that Depersonalisation is a defence mechanism developed by health care

professionals to cope with conflicting role demands, the result of which is to dehumanise

care to patients (Cash, 1991).

Further, it could be argued that a negative, dehumanising attitude towards one’s

clients may develop as nurses try to cope with unforeseen pressures associated with caring

for patients. Caring for patients is characterised by some degree of uncertainty and

unpredictability. For instance, nurses can never be certain how patients will react to the

medical diagnosis and treatment. Furthermore, nursing staff sometimes fear that a mistake

will be made in the treatment of a patient. It could be inferred that Depersonalisation is one

coping mechanism nurses employ to enable them to deal with the unpredictability and

uncertainty associated with treating patients.

 287

For Personal Accomplishment, role stress variables explain 20.5% of the total

variance. Sociodemographic factors were not significantly associated with Personal

Accomplishment and therefore were excluded from the regression equation. Role Conflict

and Role Ambiguity were the main determinants of reduced Personal Accomplishment, with

Role Conflict explaining most of the variance.

One possible explanation for this finding is that employees associate their ability to

handle many tasks at the one time, and many roles at the one time, with personal competence

(Cash, 1991). When an employee is unable to meet these demands they may feel inadequate

and consequently develop feelings of diminished personal accomplishment (Cash, 1991). In

addition, Cordes and Dougherty (1993, p.646) suggested that “when one feels unsure of what

is expected of one’s performance or if little or no feedback is given in relation to their work,

an individual’s sense of personal accomplishment may be undermined by ambiguity-induced

suboptimal performance.” This ambiguity makes it difficult for employees to perform (or

perceive they are performing) at an optimal level (Cordes & Dougherty, 1993). Not

surprisingly, employees may begin to feel incompetent and start to doubt their ability to cope

with any extra work demands (Jackson et al., 1986; Schwab & Iwanicki, 1982). Other

researchers (Cash, 1991; Holgate & Clegg, 1991) have also found a significant negative

relationship between Personal Accomplishment and Role Ambiguity. Cash (1991) suggested

that although the focus of nursing is on care and not cure, many nurses have come to equate

health care with cure. When the patients are not cured they may see themselves as failures

and experience lower levels of Personal Accomplishment.

Overall, the above findings suggest that younger nurses are especially vulnerable to

burnout. In addition, nurses working full-time are also more susceptible to experiencing
 288

burnout than part-time or casual nurses. The results also demonstrate that work stressors are

the main determinants of burnout. Generic role stressors appear to be better predictors of

burnout than job-specific work stressors. Furthermore, Role Conflict appears to be the

strongest predictor of burnout as it was that only stressor to significantly predicted Emotional

Exhaustion, Depersonalisation and reduced Personal Accomplishment.

In the present study, Emotional Exhaustion is predicted by work stressors responsible

for increasing the amount of effort expended by nurses, namely Role Overload, Job

Conditions, and Role Conflict. Over time, a nurse’s emotional resources become depleted

and Emotional Exhaustion may develop. Depersonalisation appears to be a response to

conflicting role demands and unpredictable events that occur as a result of caring for patients.

In trying to cope with multiple demands and unforeseen pressures, nurses may develop an

excessive detachment from their patients. Personal Accomplishment seems to diminish when

nurses doubt their ability to handle multiple role demands, especially if they do not obtain

feedback about their job performance.

Main Effects of Work Support on Burnout

Evidence for main effects of work support on burnout were limited. For instance, the

work support variables did not explain any additional variance in Emotional Exhaustion once

the work stressors had been accounted. Supervisor Support had a small, but significant main

effect on Depersonalisation (β = -.15, p < .05). Specifically, nurses with low levels of

Supervisor Support report higher levels of Depersonalisation, than nurses with high levels of

Supervisor Support. Overall, however, Supervisor Support explained a slight, but significant

2.7% in the explained variance of Depersonalisation.

 289

Supervisor Support also had a small, but significant main effect on Personal

Accomplishment (β = -.24, p < .01). In particular, nurses with low levels of Supervisor

Support reported lower levels of Personal Accomplishment than nurses with high levels of

Supervisor Support. Overall, however, Supervisor Support explained a slight, but significant,

4% in the explained variance of Personal Accomplishment. Coworker Support did not have a

significant main effect on any of the burnout components. In the present study, the amount

variance explained by work support in the prediction of burnout appears to be much lower

than other studies. For instance, Lee and Ashforth’s (1996) meta-analytic study found that on

average, support from the supervisor explained 14% of the variance in Emotional

Exhaustion, 6% of Depersonalisation, and 2% of Personal Accomplishment. Support from

coworkers explained 5% of the variance in Emotional Exhaustion, 5%, of Depersonalisation

and 2% of Personal Accomplishment. Their results, however, have not been replicated in

longitudinal studies on social support and burnout (Dignam et al., 1986; Gusy, 1995).

In summary, evidence for the main effect model of work support was only apparent

for Supervisor Support. Supervisor Support significantly reduces Depersonalisation and

enhances levels of Personal Accomplishment, independent of nurses’ levels of work stress.

This finding is consistent with the ‘independent distress deterrent’ model by Wheaton (1985).

It could be argued that the supervisor is the most effective person in terms of relieving the

nurses’ burnout. Contrary to expectations, Coworker Support had no significant main effects

on burnout. This may be because coworkers are not in a position of power to alter or change

the working situation at hand.

 290

Buffering Effect of Work Support on Burnout

The buffering effect of work support was tested for each component of burnout. The

interaction terms did not significantly explain any additional variance in the burnout

components. The present study found no evidence of a significant buffering effect of work

support on burnout and thus Hypothesis 12 was rejected.

Although a significant standardised beta coefficient was found for the interaction

between Role Overload and Instrumental Coworker Support on Emotional Exhaustion (β = -

.21, p < .05), the interaction term did not significantly improve the prediction of Emotional

Exhaustion. Visual inspection of the interaction term in graph format confirmed that although

the interaction was very slight, higher levels of Emotional Exhaustion are associated with

low levels of Instrumental Coworker Support when Role Overload is high. Conversely, lower

levels of Emotional Exhaustion are associated with high levels of Instrumental Coworker

Support when Role Overload is high. Most studies (Beehr, Stacy, Murray, & Jex, 1996;

Bourbonnais et al., 1996; El-Bassel et al., 1998) have also failed to find any reliable evidence

for the buffering effect of Coworker or Supervisor Support on health-related strains, such as

burnout. Consequently, there was no significant evidence to support Cutrona and Russell’s

(1990) assertion that for uncontrollable stressful events, emotional support will have a

significant buffering effect on burnout, thus Hypothesis 13 was rejected. Similarly, there was

no significant evidence to suggest that for controllable stressful events, instrumental support

will have a significant buffering effect on burnout, thus Hypothesis 14 was also rejected.

Possible reasons why significant buffering effects have not been found are now discussed.
 291

Limitations

The failure to identify buffering effects may be due to methodological reasons.

According to Aguinis and Stone-Romero (1997), the statistical power to detect interactive

effects is quite low, especially in situations where the predictors are intercorrelated. In the

present study, the work stressor and work support predictors were intercorrelated with

coefficients ranging from .21 to .55. Furthermore, the current study examined only work

support. It may be that for nurses, non-work support (i.e., family and friends) is more

effective than work support in buffering burnout. Finally, studies that measure social support

at a single point in time may provide inadequate tests of the buffering hypothesis because the

effects of support are confounded in a cross-sectional interaction effect (Hutchinson, 1999).

Thoits (1982) advocates a longitudinal design to test the buffering hypothesis.

The failure to find support for the stressor-support matching theory may be attributed

to the measures used to assess Coworker Support and Supervisor Support. A major weakness

of the work support scales in the present research was the high intercorrelations between the

emotional and instrumental support subscales. Factor analysis of the Supervisor Support

Scale in Study 2 revealed that the support items loaded primarily onto one factor. It was

therefore not possible to examine the stressor-support matching hypothesis for Supervisor

Support in Study 3. Although, factor analysis of the Coworker Support identified two

subscales, that is, emotional and instrumental support, the subscales were highly correlated (r

= .72). Significant overlap between the emotional and instrumental dimensions appears to be

a common problem for a number of social support measures (Cutrona, 1990). For instance,

Fenlason and Beehr (1994) used Caplan’s et al. Social Support Instrument to measure

support among nurses and also found that the emotional and instrumental subscales were
 292

highly correlated (r = .74). The inability of the support scales to clearly discriminate between

emotional and instrumental support may explain why supporting evidence for the stressor-

support matching theory is not frequently found.

Summary

Taking into consideration the above findings, three main conclusions can be drawn

from the current study. First, it was demonstrated that work stress is the strongest predictor

of burnout in the nursing profession. Work stress predicts 39.2% of the variance in Emotional

Exhaustion, 32.6% of the variance in Depersonalisation, and 17.7% of the variance in

Personal Accomplishment. Work stressors are differentially related to the burnout syndrome.

The results indicate that Role Overload, Role Conflict, and Job Conditions predict Emotional

Exhaustion. Role Conflict and Patient Care Uncertainty predict Depersonalisation. Role

Conflict and Role Ambiguity predict Personal Accomplishment.

Second, sociodemographic factors explain a small, but significant part of the variance

in burnout. Age and Employment Status explained 3.4% of the variance in Emotional

Exhaustion and Age explained 11.5% of the variance in Depersonalisation. Overall, however,

sociodemographic factors are not major determinants of burnout.

Third, work support is a weak predictor of burnout, explaining 2.7% of the variance

in Depersonalisation and 4.2% of the variance in Personal Accomplishment. Contrary to

expectations, only Supervisor Support has a significant main effect on burnout. Coworker

Support did not significantly influence burnout. Furthermore, there is no evidence to suggest

that work support interacts with work stressors to buffer burnout. Cutrona and Russell’s

(1990) proposition that certain forms of support are most beneficial following specific kinds

of stress was also not supported. It would seem that supporting evidence for the optimal
 293

matching model has primarily come from studies (e.g., Cohen & Wills, 1985; Cutrona &

Russell, 1990) examining stressful life events rather than studies examining stressful work

situations.
 294

CHAPTER 9

Discussion and Conclusion of the Research Program

Chapter 9 presents an overall discussion of the current research program. First,

an overview of the research program is given. Second, the main empirical findings

from the research will be briefly discussed. Next, the theoretical implications of the

research program on burnout are considered. This discussion leads into the fourth

section that focuses on the theoretical and methodological issues associated with

social support. Next, the implications of this research for the health care industry, and

in particular nurses, are discussed. Finally, recommendations for future research,

including suggestions to address the limitations of the current program of research are

made.

Overview of the Research Program

The current research program was designed to identify specific job demands

that relate to high levels of burnout in nurses, and to explore the influence of social

support on the stress-strain relationship. Specifically, the thesis focused on Australian

nurses working in the public hospital system. The research program commenced in

1999. It was at this time that the Australian nursing profession was frequently under

the media’s spotlight. Hospitals were encountering a nursing shortage as staff were

voluntarily leaving the profession or instead, were choosing to work part-time in an

attempt to gain more control over their workload (Iliffe, 2002).

Changes to public hospital management systems over the last decade have had

major implications on the nature of nursing work. The shift to a ‘cost-control’

approach has resulted in a higher patient throughput and a greater acuity of patients

serviced (Lumby, 1996). At the same time, the level of responsibility and the degree

of accountability assigned to nurses is increasing (Lumby, 1996). It is not surprising

 295

that nurses have become increasingly vocal about issues affecting their daily working

lives in the last few years. Nurses in all States and Territories have taken industrial

action, or have threatened to take action over poor wages and unfair job conditions

(Iliffe, 2002).

The timing and value of this research was extremely pertinent and the

researcher seized this opportunity to conduct research on burnout in a highly stressful

occupation. To the researcher’s knowledge, there are no studies to date that have

taken into consideration the influence of work support on Australian nurses’ levels of

burnout.

The program utilised both qualitative and quantitative research methodologies.

First, qualitative data was collected and analysed using a technique known as content

analysis. Second, quantitative data was gathered and subjected to statistical analysis

using SPSS Version 11.0. In Study 1, a focus group methodology provided rich

contextual data on nurses’ perceptions of the types of stressors they are exposed to at

work and the types and sources of support they receive at work. It involved the

facilitation of ten focus groups with 68 nurses from two public hospitals. The decision

to use these two public hospitals was primarily determined by the circumstances

associated with obtaining official permission to conduct the research.

Study 1 provided insight into different types of job-specific stressors and

generic role stressors nurses confront at work. These findings assisted the researcher

in selecting relevant measures of work stress to be used in Study 2. Study 1 also

provided evidence that social support should be conceptualised as a multidimensional

construct. Nurses identified two primary sources of support (i.e., coworkers and the

supervisor) and two major types of support (emotional and instrumental support)
 296

available to them at work. Based on these findings, the researcher developed a

contextually relevant measure of work support for nurses.

Study 2 involved the development of a survey to measure the main variables

of interest to the research program: job-specific stress, role stress, work support

(coworker and supervisor support) and burnout. A total of 273 nurses (38 males, 235

females) from three public hospitals completed the survey. The response rate for the

survey was adequate (approximately 68%). The sample’s sociodemographic

characteristics were similar to those of the larger Australian nursing population.

Although empirically it was not possible to make generalisations about the larger

nursing population, the above finding would seem to suggest that the research

findings are typical of the Australian nursing population.

In Study 2, the factorial validity and internal consistency of the job-specific

stress measure (the HPSI) and the work support measures (i.e., the Supervisor Support

Scale and the Coworker Support Scale) were established. The decision to further

investigate the psychometric properties of the job-specific measure was prompted by

the fact that the HPSI (Wolfgang, 1988a) has not been strongly represented in the

nursing literature and therefore few empirical studies have examined its

appropriateness in measuring nurses’ work stress. Furthermore, the HPSI was

modified for the purpose of the current research. Similarly, the work support scales

were constructed specifically for the purpose of this research and therefore the

soundness of these measures, in terms of their reliability and validity, were not

known.

The factor structure of the HPSI was found to be somewhat different from

previous empirical studies. The emerging factor structure that emerged, however, was

deemed to reflect a more accurate representation of Australian nurses’ perceptions of

 297

major sources of work stress. The new factor composition was therefore retained. The

factor structure of the Coworker Support Scale performed as expected. The items

loaded onto two factors reflecting emotional and instrumental support. This scale was

limited by a degree of overlap between the emotional and instrumental support

subscales (r = .72). However, Tabachnick and Fidell (1996) suggest that statistical

problems created by singularity and multicollinearity occur at much higher

correlations (.90 and higher) and therefore a decision was made to retain the two

subscales for further statistical analyses in Study 3. In contrast, the factor structure of

the Supervisor Support Scale revealed a one-factor solution. Thus, in subsequent

analyses, the items were aggregated to give a unidimensional measure of Supervisor

Support.

The quantitative data collected in Study 2 was subjected to further statistical

analysis in Study 3. In Study 3, the program’s main research hypotheses were

examined. Specifically, referent levels of work stress, burnout, and work support were

established. Second, the relationships between the main research variables were

explored. Third, the primary determinants of burnout were examined. Finally, the

effects of work support on the stress-burnout relationship were investigated.

Overview of the Main Research Findings

In Study 1, the data obtained from the focus group discussions provided

valuable insight into the stressful conditions Australian nurses regularly confront

working in a public hospital setting, thus answering principal research question 1.

Nurses’ work stressors were classified into four broad stress categories. These

included: 1) job conditions; 2) job uncertainty; 3) interpersonal conflict; and 4) lack of

professional recognition and support. A frequency count revealed five key sources of

stress for Australian nurses. These included: 1) work overload; 2) lack of respect and
 298

recognition from doctors; 3) difficult/demanding patients; 4) difficult/demanding

relatives; and 5) lack of support. These findings are consistent with recent Australian

empirical studies (Bryant et al., 2000; Healy & McKay, 2000; Lumby, 1996), as well

as international empirical studies (e.g., Wheeler, 1998; Wheeler & Riding, 1998). The

findings from Study 1 prompted the researcher to utilise Wolfgang’s (1988a) HPSI in

Study 2 to measure nurses’ job-specific stressors.

Study 1 also answered principal research question 2 by clarifying some of the

confusion surrounding the conceptualisation and operationalisation of the social

support construct. In the present research, nurses reported that their coworkers,

followed by their supervisor, were their main sources of support at work.

Furthermore, nurses discussed social support from a multidimensional perspective,

recognising various support components that could be classified broadly as emotional

and instrumental support. Based on these findings, work support was defined as the

emotional and instrumental assistance an individual receives through his or her

interpersonal relationships at work. Due to methodological insufficiencies in previous

nursing support scales, the researcher constructed a work support measure to assess

nurses’ levels of Coworker and Supervisor Support. Items that closely aligned to

nurses’ perceptions of emotional and instrumental support were adapted from

established measures including: Shinn et al.’s (1989) Supervisor Support Scale, Ray

and Miller’s (1994) Supervisor/Coworker Support Scale and King et al.’s (1995)

Family Support Inventory.

In Study 2, supporting evidence to suggest that the HPSI, the Coworker

Support Scale, and the Supervisor Support Scale possess adequate levels of internal

consistency and construct validity was found. Factor analysis of the HPSI revealed a

four-factor structure: Lack of Professional Recognition and Support, Patient Care

 299

Uncertainty, Job Conditions and Interpersonal Conflict. The HPSI subscales yielded

Cronbach coefficient alphas ranging from .62 to .83. Inter-scale correlations were low

to moderate, indicating that none of the four factors were redundant. The HPSI

subscales also demonstrated evidence of convergent validity when the HPSI subscales

were correlated with role stress and burnout.

Factor analysis of the work support scales provided interesting results. As

expected, a two-factor structure emerged for the Coworker Support Scale. These

factors, labelled Emotional Coworker Support and Instrumental Coworker Support,

yielded Cronbach coefficient alphas of .92 and .88 respectively. The emotional and

instrumental subscales shared approximately 52% variance. Since 42% of the

variance remained unique, the subscales were not considered redundant. Factor

analysis of the Supervisor Support Scale, however, revealed a one-factor solution,

yielding a Cronbach coefficient alpha of .96. The work support scales demonstrated

some evidence of construct validity when they were correlated with the work stress

and burnout variables.

The findings from Study 3 indicated that overall Australian nurses report low

to moderate levels of work stress. Although nurses are exposed to a variety of

stressful situations, most job demands are not perceived to be highly stressful. Their

primary source of stress was their excessive workload. It was inferred that for nurses,

individual job demands alone are insufficient to cause high levels of stress. Rather, it

is a combination of work demands that leads to moderate to high levels of stress

among nurses. For instance, nurses perform a number of duties, meet tight deadlines

imposed by others, and contend with multiple interruptions and rationing of

insufficient resources on a daily basis. The qualitative data in Study 1 further

supported this finding. Examination of the wider nurse stress literature (e.g., Bryant et
 300

al., 2000; Hillhouse & Adler, 1997; Healy & McKay, 2000; Riding & Wheeler, 1995;

Taylor et al., 1999; Tyler & Cushway, 1995; Wheeler & Riding, 1994) also reveals

that workload is the key source of stress for nurses. Taken together, these findings

would seem to suggest that workload is the primary source of stress for nurses both

nationally and internationally. It could be concluded that a nurses’ workload is the

main reason why this profession is perceived to be stressful.

In Study 3, nurses reported receiving moderate levels of work support. Nurses’

coworkers were perceived to provide higher levels of support than their supervisors.

The respondents in the focus group discussions in Study 1 also identified their

coworkers as their primary source of support. They indicated that their colleagues

relate well to the pressures and problems they experience at work.

Despite receiving moderate levels of support, the overall sample reported

moderate levels of burnout. Furthermore, a substantial proportion of the sample

population acknowledged experiencing some aspect of burnout. In particular, 40% of

the sample reported high levels of Emotional Exhaustion, while 33% reported high

levels of Depersonalisation. In addition, 42% reported low levels of Personal

Accomplishment. Almost 6% of the study sample could be categorised as having high

levels of burnout (i.e., high Emotional Exhaustion, high Depersonalisation and low

Personal Accomplishment). These results suggest that nurses are indeed vulnerable to

burnout. In comparison to other health care professionals, such as physicians, nurses

reported lower levels of Emotional Exhaustion and Depersonalisation, but higher

levels of diminished Personal Accomplishment.

The findings showed that both job-specific stressors and role stressors are

related to burnout, however role stressors are more strongly related to the burnout

dimensions. Sociodemographic factors (Age and Employment Status) and work

 301

stressors (Role Overload, Role Conflict and Job Conditions) significantly predicted

Emotional Exhaustion among Australian nurses. However, work stressors explained

most of the total variance (39.2%). Role Overload was the key determinant of

Emotional Exhaustion. The study’s variables explained overall 42.2% of the variance

in Emotional Exhaustion.

The findings demonstrated that Age, and the work stressors - Role Conflict

and Patient Care Uncertainty significantly predicted Depersonalisation, with work

stress explaining most of the total variance (22.1%). Supervisor Support was also a

significant predictor of Depersonalisation. Supervisor Support, however, did not

significantly buffer the relationship between work stress and Depersonalisation. The

study’s variables explained 34.2% of the total variance in Depersonalisation.

Only role stressors (i.e., Role Conflict and Role Ambiguity) were significant

predictors of Personal Accomplishment. Role stressors explained 17.7% of the

variance in Personal Accomplishment. Again, Supervisor Support significantly

predicted Personal Accomplishment, but it did not significantly buffer the relationship

between role stressors and Personal Accomplishment. The study’s variables explained

20.5% of the total variance in Personal Accomplishment.

The present research program extended our current knowledge of the major

antecedents of burnout in nurses. The findings also seem to suggest that overall,

Emotional Exhaustion is related to quantitative workload rather than qualitative

demands that result from problems in interacting with difficult patients or dealing

with their emotional problems (Schaufeli & Enzmann, 1998). This differs from the

traditional conceptualisation of burnout in which it was assumed that emotionally

charged interactions with clients were its root cause (Schaufeli & Enzmann). In the
 302

section below, the implications of this research on the conceptual model of burnout

are discussed.

Theoretical Implications for Burnout

The current research program extends our understanding of the burnout

phenomenon among nurses. Although the present study is unable to comment on the

sequential ordering of the burnout components, the findings provide further support

for the multidimensionality of burnout and insight into the critical variables that

contribute to its development.

First, the findings demonstrate support for Maslach’s (1993) assertion that

burnout is not a unidimensional concept. Significant modest correlations were found

between Emotional Exhaustion and Depersonalisation. In addition, both Emotional

Exhaustion and Depersonalisation were significantly but weakly related to reduced

Personal Accomplishment. The results indicated that burnout is a three-component

syndrome. The three burnout components were conceptually distinct but were not

empirically uncorrelated.

Second, the value of distinguishing between the three burnout dimensions in

the present research was shown by findings of differential patterns between the

burnout components and the independent variables (sociodemographic factors, work

stressors, social support). Taking into consideration previous models of burnout and

the present study’s findings, a tentative model of burnout for nurses has been devised

(see Figure 9.1). This conceptual model is based on a stress-strain-coping theoretical

framework (Lazarus & Folkman, 1984). Longitudinal studies are required, however,

to confirm the appropriateness of the causal relationships between the variables.

 303

Sociodemographic Factors
• Age
• Employment Status

Emotional Exhaustion

Job/Role Stressors
• Role Overload
• Role Conflict
• Job Conditions

Sociodemographic Factors
• Age

Interpersonal
Coping Resource Depersonalisation
• Supervisor
Support Job/Role Stressors
• Role Conflict
• Patient Care
Uncertainty

Interpersonal
Coping Resource Reduced Personal
• Supervisor Role Stressors Accomplishment
Support • Role Conflict
• Role Ambiguity

Figure 9.1. A conceptual model of burnout among nurses.

It could be inferred from the conceptual model that Emotional Exhaustion is

primarily and strongly associated with work demands that increase the amount of

emotional energy required by nurses to perform their work, for example, Role

Overload, Job Conditions, and Role Conflict. Other researchers have found similar

results (e.g., Cordes & Dougherty, 1997; Jackson et al., 1986). Age and Employment

Status are also weak, but significant contributing factors of Emotional Exhaustion in
 304

nurses. In particular, younger nurses report higher levels of Emotional Exhaustion

than older nurses. Other nursing studies (e.g., Bennett et al., 1991; Koivula et al.,

2000; Van Servallen & Leake, 1993) have also found evidence that a nurses’ age

influences burnout levels, with younger nurses experiencing more burnout than older

nurses. Furthermore, full-time nurses have a greater likelihood of experiencing

Emotional Exhaustion than part-time or casual nurses.

Depersonalisation is argued to be a means (albeit futile) of stanching the flow

of emotional energy, of coping with growing exhaustion (Schaufeli & Enzmann,

1998). It could be inferred that Depersonalisation results from work stressors that

place heavy demands on nurses’ emotional reserves, such as Role Conflict. It could

also be argued that Depersonalisation is strongly related to unpredictable stressful

events, such as Patient Care Uncertainty. Age is also a major determinant of

Depersonalisation. Specifically, younger nurses are more likely to adopt a detached

manner than older nurses. Higher levels of Emotional Exhaustion and

Depersonalisation among younger nursing staff have been attributed to the

discrepancy between their education and the harsh reality of their work (Koivula et

al., 2000).

Finally, it could be assumed that reduced levels of Personal Accomplishment

in nurses is strongly related to factors that suggest that they are unable to competently

manage multiple demands, such as Role Conflict. Low levels of Personal

Accomplishment may also result from being unsure about what is expected, such as

Role Ambiguity. It could be argued that it is difficult to develop strong feelings of

efficacy when role demands are incongruent or incompatible or when one is unsure of

what is expected of one’s performance.

 305

While Supervisor Support does not significantly predict Emotional

Exhaustion, it is significantly associated with Depersonalisation and reduced Personal

Accomplishment among nurses. This finding may provide some support for Leiter’s

(1993) model of burnout in which work demands and resources (e.g. coping) are

differentially associated with the three dimensions. Specifically, it is posited that

demands are more strongly related to Emotional Exhaustion, and resources are more

strongly related to either Depersonalisation or reduced Personal Accomplishment.

Hobfoll and Freedy (1993) suggested that job demands trigger strain in the form of

physical and emotional exhaustion, whereas resources help to overcome the need for

defensive coping and enhance one’s self-efficacy. In the present research, it was

found that Supervisor Support enhanced a nurse’s ability to cope with burnout.

Theoretical and Methodological Implications for Social Support

The current research program extended our knowledge of the social support

construct and its effects on stress and burnout among nurses. First, content analysis of

the qualitative data in Study 1 enabled the research to offer a concise

conceptualisation of the social support construct. In the present research program,

work support was defined as the emotional and instrumental assistance an individual

receives through his or her interpersonal relationships at work.

It was observed by the researcher that when nurses were asked about their

main sources of work support, their first thoughts were of their nursing colleagues. A

frequency count identified nursing colleagues as their primary source of support,

followed by their supervisor. The quantitative findings in Study 3 also indicated that

nurses reported receiving higher levels of Coworker Support than Supervisor Support.

In Study 1, the types of assistance nurses receive at work were broadly

classified as emotional support and instrumental support. For the purpose of this
 306

research program, Emotional support was defined as the emotional comfort an

individual receives during a stressful situation that leads the person to believe that

they are cared for and valued by others. Instrumental support was defined as the

instrumental assistance that a person receives as a result of being given the necessary

resources (e.g., physical help with a task) to cope with the stressful situation, or

guidance or advice to help solve a problem. The researcher noted during the focus

group discussions that nurses predominantly discussed examples of emotional support

when they discussed the supportive actions provided by their coworkers. In contrast,

nurses primarily discussed examples of instrumental support when they discussed the

assistance provided by their supervisors.

Interestingly, factor analysis of the work support scales in Study 2 revealed

that although nurses were able to recognise the difference between emotional and

instrumental support when offered by their coworkers, they were unable to clearly

distinguish between emotional and instrumental support when it was provided by their

supervisor. Perhaps instrumental acts of support provided by a supervisor (e.g., giving

advice or practical assistance) are also perceived by nurses as evidence that their

supervisor cares and understands the stressful situation, which is also an indication of

emotional support. The relationship between different types of social support needs to

be further investigated (Scheck et al., 1997). Although the present thesis does not test

this possibility, it seems that there may be a reciprocal relationship between emotional

and instrumental support. A review of the social support literature suggests that there

is a lack of theory focusing on the specific relations among various support

components (Scheck et al.). These findings may also suggest that for nurses, the

person who is providing the support (the source of support) is more important than the

actual support given (type of support). This raises questions about the relative value of
 307

measuring different types of support. Perhaps future research should focus on testing

the unique contributions of different sources of support rather than the different types

of support.

The research demonstrated that work support assists nurses in one of two

ways. First, work support mitigated perceived work stressors. Perhaps support helps

nurses to redefine the potential harm in a stressful situation (Cordes & Dougherty,

1993). Study 3 revealed that Coworker Support and Supervisor Support were

significantly negatively related to certain job-specific stressors and role stressors.

More specifically, high levels of work support were associated with lower levels of

work stress. Alternatively, low levels of work support were associated with higher

levels of work stress. This finding may point to the need for further investigation of

the indirect effect model in which social support is hypothesised to have an indirect

effect on strains by directly reducing the impact of work stressors.

Second, work support was found to reduce the level of strain experienced by

nurses, regardless of their levels of work stress. Examination of the effects of work

support on burnout revealed that only Supervisor Support had a main or direct effect

on burnout among nurses. Specifically, nurses reporting higher levels of Supervisor

Support report lower levels of Depersonalisation than nurses reporting lower levels of

Supervisor Support. Similarly, nurses reporting higher levels of Supervisor Support

report higher levels of Personal Accomplishment than nurses reporting lower levels of

Supervisor Support. These findings are inconsistent with previous studies (Barrera,

1986; Sandler & Barrera, 1984; Sarason, Shearin, Pierce, & Sarason, 1987; Stokes,

1983) documenting that the receipt of support is associated with poor adjustment.

While some researchers (e.g., Barrera, 1986; Fisher, Nadler, & Whitcher-Alagna,

1982) believe that the receipt of support entails a cost to self-esteem, the present
 308

research clearly demonstrated that the receipt of work support, specifically Supervisor

Support, promotes well-being in nurses.

The results demonstrated that neither Emotional nor Instrumental Coworker

Support had a main effect on burnout. This may suggest that coworkers are not in a

position to provide the appropriate resources that enhance a nurse’s ability to cope

with the different aspects of burnout. There is some evidence in the research

literature, that supervisors are more effective in reducing stressors than coworkers

(Beehr et al., 1990; Ganster et al., 1986; Jackson et al., 1986; O’Driscoll & Beehr,

1994). It is proposed that because supervisors are in a higher position than their

nursing subordinates, they also have more power to alter stressors in the work

environment. Alternatively, it could be that a supervisor is able to provide the

necessary resources (or instrumental support) to assist nurses to cope with the

situation. In addition, the nursing supervisor may encourage subordinates to engage in

problem-focused coping, rather than emotion-focused coping. According to Lazarus

and Folkman’s (1984) coping theory, when coping actions change the person-

environment relationship for the better, they eliminate or reduce the psychological

grounds for harm or threat. The resulting changed appraisal may in turn lead to

changes in the psychological response to the stressful situation. Overall, however,

work support explained between 1 to 4% of the variance in burnout. This finding

suggests that the receipt of work support plays a small, but significant, role in

reducing burnout among nurses.

The current research program found no evidence to suggest that work support

significantly buffers burnout in nurses. Although the results in Study 3 indicated that

Instrumental Coworker Support interacts with Role Overload to reduce Emotional

Exhaustion, it did not significantly improve the prediction of Emotional Exhaustion.

 309

Several other studies have also found no evidence of the buffering effect of social

support (e.g., Bourbonnais et al., 1998; El-Bassel, et al., 1998). According to Cohen

and Wills (1985) and Cutrona and Russell (1990), buffering is most likely to occur

when there is a reasonable match between the needs elicited by the stressful event and

the functions of support. If this proposition is correct, it could be assumed from the

present study’s findings that there is an inadequate match between nurses’ needs and

the types of support they receive at work. Alternatively, it may be that support from

outside the workplace has a greater influence of nurse’s health and well-being. Some

studies (e.g., Munro, Rodwell, & Harding, 1998; Peeters & LeBlanc, 2001) have

found that non-work support (i.e., family, friends) is more significantly related to

employee’s health and satisfaction than work support. For instance, Peeters and

LeBlanc (2001) found that family support moderated the relationship between

quantitative demands and Depersonalisation in nurses. Furthermore, a study by

Roxburgh (1999) also showed a gender difference with regard to the relative

importance of work and non-work support. Women seemed to profit more from non-

work support, whereas men benefited more from work support. Since nursing is

predominantly a female profession, it could be inferred that overall, non-work support

may be more beneficial to nurses than work support. Further research is necessary,

however, to confirm this hypothesis. Thoits (1982) argued that the modest support for

a significant buffering effect is probably overestimated as the majority of studies have

used cross-sectional data. When longitudinal studies (e.g., Lin & Easel, 1984; Thoits,

1982) have been conducted, support for the buffering hypothesis has been weak and

probably not greater than would be expected by chance (Dignam & West, 1988;

Payne & Jones, 1987).

 310

The failure to find any evidence of the significant buffering effects of social

support in the present research, however, may also be attributed to methodological

issues. The present research took the perspective that social support is effective in

promoting coping and reducing the effects of stress when the type of assistance

provided matches the demands of the stressor. The emphasis was on the actual

assistance given by others during a stressful situation. Thus, a measure of received

support was used to investigate the stressor-support matching hypothesis. There has

been, however, a lack of evidence supporting the stress-buffering effect when

measures of received support have been used (Cohen et al., 1984; Cohen &

Hoberman, 1983; Sandler & Lakey, 1982; Wethington & Kessler, 1986).

One possible explanation may be that it is not the actual receipt of support, but

the belief that support is available if required that protects persons against the adverse

effects of stress. From this perspective, the emphasis is on an individual’s perception

that support is available if needed. According to Lazarus and Folkman (1984), how

people interpret situations is very important in determining the stressfulness of an

event. Cohen and Hoberman (1983) hypothesised the belief that support is available

reduces the effects of stress by contributing to less negative appraisals. Because the

subjects’ perceptions of the availability of support are central to the appraisal of

stress, a measure of perceived availability of functional support is considered optimal

when investigating the stressor-support matching theory. It could be argued therefore

that in the present study, a measure of perceived support may have been more

appropriate than a measure of received support to test the stressor-support matching

model. However, it was the intention of the current research program to assess

supportive transactions between nurses and their primary sources of support by

examining the actual level of support nurses’ receive. The present research heeded
 311

Stewart’s (1993) advice and went beyond individual self-reports of perceived support

to examine how actual work relationships influence nurses’ adaptation to work stress

and burnout.

Another limitation of the current research program was that the Coworker and

Supervisor Support Scales did not possess optimal psychometric properties. Testing

of the buffering model requires sophisticated assessment techniques. In the current

research program, a measure of support was developed to assess nurses’ perceptions

of the main types of support they receive at work. There was, however, a degree of

overlap between the emotional and instrumental subscales on the Coworker Support

Scale. In addition, the Supervisor Support Scale was unable to discriminate between

emotional and instrumental support, thereby making it impossible to tests the

differential effects of the two types of supportive functions. Furthermore, the work

support scales were designed to measure two broad types of support emotional and

instrumental support. Perhaps a measure of work support that taps into multiple

independent support functions would be more effective in determining the particular

support functions that interact with work stressors to affect burnout, thereby shedding

light on the mechanisms linking social support to health. Researchers need to focus

their attention on designing a work support measure that does not suffer from high

inter-scale correlations. To date, most multidimensional measures of social support

are limited by the significant overlap between the social support dimensions.

Furthermore, it is proposed that quantitative surveys should be used in

conjunction with qualitative methods to research the buffering effect of social support.

For instance, a standardised interview (verbal or written) could be used to determine

what types of support are optimally effective in reducing different types of stressful

situations. Respondents could be asked to discuss what types of support they use
 312

from different sources of support to combat specific types of stress. At the same time,

the researcher could obtain information about the nurses’ degree of control over each

stressful situation to establish whether there is a link between the types of support

used and the controllability of the stressful situation. It is proposed that a qualitative

approach may be a more valid and in-depth way of determining how support

functions to enhance coping and reduce stress-related outcomes.

Implications for the Health Care Industry

The findings from the current research program have implications for the

health care industry, and more specifically for nurses. The current research has shown

that burnout in nurses is strongly related to organisational demands and that a nurse’s

supervisor plays a small, but important role in reducing burnout. In addition, younger

nurses and nurses working full-time are the most vulnerable to burnout.

Previous studies (e.g., Constantini et al., 1997; Duquette et al., 1995; Kilfedder

et al., 2001) have demonstrated that high levels of burnout among employees have

detrimental consequences for both the individual and the organisation. For

individuals, consequences may include a variety of physical and psychological

problems such as irritability, tension, anxiety, headaches and depression (Constantini

et al., 1997; Duquette et al., 1995; Kilfedder et al., 2001). From an organisational

perspective, high levels of burnout are associated with increased absenteeism,

intention to leave the organisation, low job satisfaction, and low levels of employee

commitment to the organisation (Constantini et al., 1997; Duquette et al., 1995;

Kilfedder et al., 2001). It is therefore important that hospital managers acknowledge

that nurses are stressed and are vulnerable to burnout and accept that interventions are

required to combat this problem.

 313

The most effective way to manage burnout among nurses is for management

to take a proactive stance and target nurses who are potentially vulnerable to burnout.

Burnout may be tackled at an individual level and organisational level (Edwards &

Burnard, 2003). In the section below, interventions designed to reduce burnout are

discussed. Interventions aimed at an organisational level will be outlined briefly,

before examining interventions targeting individuals.

Organisational Interventions

Organisational interventions generally focus on strategies to minimise stressful

work conditions. There is some evidence that organisational level interventions are

more beneficial to nursing staff than individual interventions (Akroyd et al., 2002;

Healy & McKay, 2000). The present thesis identified a number of key environmental

stressors experienced by Australian nurses working in public hospitals. These

included: (1) work overload pressures due to insufficient time and resources, (2) role

conflict pressures due to multiple task demands, (3) pressure due to lack of

involvement in planning and decision-making, (4) pressure related to insufficient pay,

(5) pressure due to insufficient performance feedback, and (6) pressures associated

with difficult/demanding patients. In the section below, potential strategies to reduce

Australian nurses’ main sources of work stress are outlined.

It would seem that management’s first responsibility might be to try and

reduce nurses’ workload. This is a complex issue that will require more than one

method and further funding from the Australian government to help make a

difference. A nurse’s workload often varies and it is difficult to control. Their

workload is dependent on their daily patient load and adequacy of staffing numbers.

Their workload is unevenly distributed throughout the day and may alternate between

slow and fast paced work, but generally the day consists of multiple demands on their
 314

time (Gaudine, 2000). In order to reduce nurses’ perceptions of workload,

management may need to consider putting in place some strategies designed to

increase nurses’ control over their workload. For example, if nurses are given some

control over staffing on their ward, they may feel less stressed. Nurses could be given

access to a list of available staff (e.g., agency staff, non–rostered nurses) that they

could call upon during peak periods if it was absolutely necessary. This would be

particularly useful for nurses doing night duty. It is important, however, that nurses on

the availability list are actually available. A nurse would feel the availability list is

worthless if they had to waste time on the telephone trying to find someone who is

willing to come into work at short notice. According to Gaudine (2002), hospitals that

value quality of patient care and their human resources should give some of the

responsibility for managing workload to nurses.

Other strategies to assist nurses manage their workload may be the

introduction of team nursing where nurses work in small groups to complete daily

tasks (e.g., answering the phone, administering medications, and doing patient

observations). Alternatively, a system could be put in place where staff are re-

allocated to different patients or different tasks every so often so that they are not

exposed to the same stressor for long periods of time. Another strategy may be the

introduction of nurse-patient ratio limits in which a nurse’s patient load is dependent

on patients’ level of acuity.

To reduce role conflict, nurses need to be given a clear definition of the

boundaries of their work. It is not acceptable that other health care professionals and

other hospital staff off-load additional work and responsibilities to nurses without

their consent. The division between nurses and other members of the health care

profession may need to be made more explicit. Supervisors may need to re-educate
 315

nurses about their role boundaries. This way, nurses will feel less obliged to take on

extra work if they are already busy. They will also feel more in control of their

workload if they are able to choose to accept additional tasks or responsibilities that

are requested from other staff.

In the present research, there is some evidence that nurses do not feel involved

in decisions concerning the successful functioning of the hospital. To allow nurses

some input into larger hospital issues, a nursing representative could be chosen from

each ward to attend management meetings in which pertinent hospital matters are

discussed. The ward representative could be responsible for speaking on behalf of

their nursing colleagues at the meetings. This will also ensure nurses are informed

about current hospital issues. In turn, this may enhance nurses’ organisational

commitment.

Overall, the sample of nurses in the current study indicated that they were

inadequately paid in relation to other health care professionals. Nurses also believed

that their skills, qualifications and experience were not rewarded appropriately in

relation to other human service professionals. Since conducting this study in 2000,

nurses’ wages have been reviewed by the Australian Industrial Relations

Commission. In May 2003, all Level One nurses (i.e., general ward nurses) were

awarded a 16.3% pay rise to be rolled out over three years (Queensland Nurses Union,

2003). This may serve to enhance nurses’ perceptions that they are valued members of

the health care industry.

Nurses also indicated that they do not receive regular feedback about their job

performance. In order to reduce the level of uncertainty and ambiguity experienced by

nurses as a result of their work environment, and to avoid role conflict, it is important

supervisors give adequate and frequent feedback to nurses about their performance on
 316

the job. Formal performance reviews conducted every six to twelve months may serve

to reduce nurses’ levels of role ambiguity and role conflict, and improve their levels

of personal accomplishment.

Finally, it is evident that nurses require further training in constructive and

appropriate ways of handling challenging (e.g., anger, violence, verbal abuse) or

demanding patient behaviour (e.g., highly dependent). To enhance awareness, recent

nursing graduates should also be lectured on the different types of patient behaviours

they may encounter at work. This will help prepare younger nurses for what to expect

in the ‘real’ world.

Individual Interventions

Hospitals may also offer individual-level interventions to reduce stress and to

prevent burnout. Most individual-level interventions focus on managing burnout

through stress-inoculation training. The training begins with strategies designed to

increase nurses’ awareness of stress and burnout, such as self-assessment and self-

monitoring (Akroyd et al., 2002). Monitoring one’s levels of stress could involve

keeping track of stressful events over a short period of time and relating them to one’s

overall feeling of well-being. Nurses could also complete self-report instruments

designed to establish individual levels of stress and burnout. It is recommended that

nurses be given feedback about what their scores on the tests mean and how their

score compares to a representative norm sample. Those who score highly on a stress

or burnout measure should be encouraged to attend stress management workshops on

adopting a healthy lifestyle or learning relaxation techniques.

A healthy lifestyle workshop may promote stress-reducing activities such as

physical exercise, proper nutrition, and weight control (Akroyd et al., 2002). Research

has demonstrated that physical exercise may be one of the more effective strategies to
 317

prevent or lessen the effects of depression and anxiety. Relaxation techniques are at

the core of many stress management programs. These programs attempt to provide the

individual with positive physiological and psychological responses to replace the

negative responses accumulated from exposure to various stressors.

The current research suggests that supervisors play a key role in creating a

healthy work environment for nurses. Supervisors need to be aware of their potential

capacity to affect levels of Depersonalisation and Personal Accomplishment in their

staff. It is proposed that supervisors be trained in different ways to provide support to

nurses. For instance, instrumental support may be effective during task execution and

emotional support may be useful during conflict management. Supervisors should

also be trained in providing feedback to nurses on their job performance. Performance

feedback will serve to enhance role clarity and reduce the ambiguity and uncertainty

associated with the clinical nursing role. These interventions may contribute to the

prevention and reduction of burnout among nurses.

The above interventions may be part of a broader development strategy of the

health care institution aimed at the promotion of a healthy and productive working

environment for nurses. Improving the working conditions of nurses may not only

decrease their levels of stress but it might also encourage them to continue nursing.

Genuine efforts to alleviate burnout among nurses should have positive effects,

including increased quality of patient care, improved quality of work life, higher

levels of job satisfaction and commitment, and lower staff turnover (Akroyd et al.,

2002). Interventions designed to identify and lessen specific aspects of workplace

stress at an organisational level may be more effective than individual-level

interventions because they focus on the cause of the problem, not predominantly on

the outcomes of the problem. However, the effectiveness of the individual and
 318

organisational interventions will only be known if the interventions are systematically

evaluated over a period of time.

In summary, healthcare organisations will ultimately pay the price for high

levels of employee burnout via the quality of services they offer (Akroyd et al., 2002).

Organisational improvements involving the individual in the workplace may have a

more pervasive and long-lasting impact on burnout than individual interventions.

Providing employees with techniques to reduce job stress is helpful, but such methods

will be more effective if the organisation also seeks to make the job less stressful. An

organisation must recognise that burnout is a legitimate workplace problem before

effective interventions can be put in place.

Limitations

It has to be recognised that the current research program had some limitations.

First, the program employed cross-sectional data to identify significant relationships

between the research variables. Consequently, no firm conclusions can be made

regarding the exact magnitude of the causal effects. Longitudinal designs, although

much more difficult to achieve, are crucial for furthering our understanding of the

nature of burnout among nurses and its development over time. A second weakness

was that the measures used to assess the main variables of interest to the research

program were all self-report instruments. Self-report measures are open to biases in

reporting. For example, the nursing sample may have under-estimated or over-

estimated their levels of stress, burnout and work support. While reliance on self-

report data always has its criticisms, some researchers argue that stress is an

experience based on the perception of a mismatch between demands and resources,

and therefore subjective reporting is paramount (Kilfedder et al., 2001). It was also

evident that the work support measures developed for the purpose of this research
 319

require further revising in order to improve their psychometric properties. In the

future, investigators should attempt to identify a social support instrument that uses a

format that has been shown to convenient and reliable, that reflects the multifaceted

nature of social support, demonstrates adequate construct validity, that can be adapted

to obtain responses that are relevant to the aims of the study and is applicable in a

variety of clinical nurse settings (Hutchinson, 1999; Stewart, 1993). Finally, some

caution should be taken when interpreting the research findings in relation to the

Australian nursing population. This is because participants were recruited from a

small catchment area (i.e., the Brisbane metropolitan district), respondents providing

the measure of the predictor and criterion variables were the same people, the

response rate was lower than desired, and some nursing divisions comprised a small

number of respondents. In addition, the management of public hospitals is not

consistent across States and Territories of Australia. Furthermore, the majority of the

sample population were female (approximately 86%) and therefore one has to be

careful with regard to generalisation of these findings to male nurses.

Future Research Recommendations

The research program extended current nursing knowledge in the areas of

occupational stress, social support, and burnout. Traditionally in nursing research,

work stress and burnout have been examined as separate and static states. The present

research viewed the development of burnout as a process, in which antecedent

conditions such as work stressors, and interpersonal coping resources such as social

support play an important role.

Most of our knowledge of burnout, particularly its antecedents and

consequences, has come from empirical research. More recently, burnout research has

been integrated into a wider psychological theoretical framework. For instance, in the
 320

present research, the relationship between job stressors, burnout, and work support,

were examined using a stress-strain-coping theoretical framework. Furthermore,

there has been a renewed interest in the process of burnout and its sequential

relationships. A limitation of the present study, however, is its use of cross-sectional

data to investigate the links between job demands, coping resources, and burnout. A

limitation of cross-sectional research is that no causal inferences can be made. Testing

the causal factors of burnout requires longitudinal research designs. In addition to

providing better evidence for causal relationships, longitudinal studies would yield

valuable information about the development and successive phases of burnout.

Although time-consuming and costly, such studies are critical to our understanding of

burnout.

It is recommended that further research be conducted to address the lack of

health care studies investigating levels of burnout. To date, reliable estimates of the

prevalence of burnout in the health care setting do not exist. On the empirical front,

more research into burnout rates is required. Researchers must replicate this study to

confirm burnout levels in nurses working in the public hospital system. A replication

will also address previously discussed concerns about the sample’s representativeness

which arose from the lower than desired response rate.

Finally, very little is known about the success of interventions designed to

reduce burnout. To date, there is no concrete evidence that interventions actually

work. A major goal of burnout research is to determine effective strategies for dealing

with it and therefore evaluation research is absolutely critical. Baseline measures of

burnout should be collected prior to the intervention to help determine the

effectiveness of the intervention over time. Future studies could extend this program
 321

of research by systematically and periodically monitoring nurse burnout in public

hospitals on a six-monthly basis.

While great strides are being made in our understanding of burnout, our

knowledge of how social support facilitates coping and adjustment among employees

is still very poor. The investigation into the effects of social support on nurses’ levels

of burnout is still in its early stages and several issues remain unclear. Until the role of

social support in preventing burnout is understood, the benefits of social support

cannot be used to its full potential (Ganster et al., 1986). In the attempt to demystify

social support, more questions have been raised. For instance, does received support

promote the initiation and maintenance of coping efforts? Do people change the way

they cope in response to received support? What types of received support are most

effective in positively influencing coping efforts?

While the current research provides evidence that the supervisor plays a

significant role in reducing burnout, it is not known what types of support provided by

the supervisor promote coping. This was because the measure developed to assess

supervisor support did not adequately distinguish between the different support

functions.

It is also not apparent from the present investigation why support provided to

nurses by their coworkers did not significantly reduce burnout. Since very little

research has been conducted in the empirical literature concerning sources of support,

it seems appropriate to extend the stressor-support matching hypothesis by including

the match between sources of support and stressful events. It is proposed that

supervisors and coworkers assist nurses in different ways and their effect on nurses’

health and well-being is poorly understood. For example, Supervisor Support may be

useful in resolving work-related problems, whereas Coworker Support may provide

 322

nurses with a sense of social integration. It is imperative that research into received

support continues and that a more refined measure of received support is used that

takes into consideration multiple support functions.

It is also important to keep in mind that the study of social support is complex

and necessitates the consideration of many factors when designing studies to measure

this concept (Hutchinson, 1999). Mixed findings regarding the role of social support

in the process of burnout are attributed to the lack of consensus on how to measure

support due to the variability in definitions and conceptualisations. Future research

should therefore adopt a definition that is based on a theoretical framework and that

can be operationalised in accordance with the aims of the study.

The present study employed a measure of received support to examine the

stressor-support matching theory (Cohen & Wills, 1985; Cutrona & Russell, 1990). It

was proposed that received support is more likely to predict outcomes when support is

optimally matched to the demands of the stressor. However, Cohen and Wills (1985)

posited that it is the perception that appropriate support is available that protects

persons against the adverse effects of stressors. According to Lazarus and his

colleagues (Lazarus, 1966; Lazarus & Folkman, 1984), how people interpret

situations is very important in determining an event’s stressfulness. There is some

evidence that perceptions of support availability are more effective in altering

appraisals if they counter the specific needs elicited by the stressful event (Cohen &

Hoberman, 1983; Cohen & McKay, 1984). It is therefore recommended that further

research examining the stressor-support matching theory use a measure of perceived

support that adequately distinguishes between different support functions (Lakey &

Cohen, 2000). It is also recommended that the buffering effects of both work and non-

work support on the stress-burnout relationship among nurses be investigated. Finally,

 323

both qualitative and quantitative approaches should be used to measure social support,

although qualitative measures of social support are considered by some to be superior

because of their greater predictive power regarding personal adjustment (e.g., Griffith,

1985). The value of collecting qualitative data is in its ability to provide rich data on

the support needs of this particular occupational group. It is paramount that the

mechanisms of social support are further researched before promoting the

implementation of formal social support interventions within the nursing profession.

Conclusion

In conclusion, the present research program further contributed to our

understanding of stress, burnout, and work support among a sample of Australian

nurses. The findings suggest that burnout can be understood within a stress-strain-

coping context. According to Lazarus and Folkman’s (1984) transactional coping

theory, strain (burnout) occurs when demands (e.g., work stressors) exceed coping

resources (e.g., social support).

The present study demonstrated that nursing in a public hospital is stressful.

The most immediate concern for nurses is their work overload. Overall, the sample

reported moderate levels of burnout. However, in comparison to other health care

professionals, the current sample of nurses generally experienced lower levels of

Emotional Exhaustion and Depersonalisation, but higher levels of reduced Personal

Accomplishment. Investigation of the major determinants of burnout revealed that

role stressors explain more of the total variance in burnout than job-specific stressors.

The current program has shown that the effects of social support in

ameliorating work stress and reducing burnout deserves more attention in the nursing

literature. This cross-sectional research program found evidence that social support

can influence the severity of work stressors and reduce the level of strain experienced
 324

by individuals. The results indicated that both Supervisor and Coworker Support

reduced the likelihood that job demands would be perceived as stressful. Furthermore,

support received from supervisors significantly reduced Depersonalisation and

increased Personal Accomplishment in nurses, irrespective of their level of work

stress (i.e., main effect). There was no significant evidence that support received from

coworkers or the nursing supervisor moderated the effects of work stress on burnout

(i.e., buffering effect). Future studies should try to examine whether specific

combinations of sources and types of support significantly add to our understanding

of the mechanisms of social support.

 381

APPENDIX A

Information Sheet and Survey Used in Study 2

A Study on Nurses’ Work Stressors, Work Support, and

Psychological Health and Well-being

Dear Respondents,

Before filling out the following questionnaire, it is important that as a participant,

you understand the importance of this study and what it hopes to achieve. The
present study is designed to examine how work support influences perceptions of
work stress. A key unanswered question in the social support literature is
whether certain forms of social support are most beneficial following specific
kinds of stress. Cutrona and Russell (1990) propose that by matching support
components to specific types of stress, better social support-based interventions
could be designed. Theoretically, the discovery of optimal stress-support
combinations may help us better understand how stressful events threaten and
how social support protects or enhances individual well-being.

The current research involves two studies. Study 1 has been completed. It
consisted of a series of 30-minute discussion groups in which nurses from the
Princess Alexandra Hospital and Redland Hospital provided valuable feedback
about events in their daily working life that cause them the most stress; and the
types and sources of support they receive from within the workplace. The
information gained from Study 1 was used to design this questionnaire. Study 2
involves the completion of a questionnaire. It is designed to measure the types of
work stressors you experience in your ward, your health and well-being, and
finally the level and types of support you receive from within your ward.

If you would like to be involved in Study2, it should be noted that your responses
to the questionnaire will remain confidential and anonymous. All response forms
are coded and therefore your responses will not at any time be able to identify
you. Responses to this survey will be viewed only by the researcher. If you have
any queries about the study itself or how to fill in the questionnaires, please do
not hesitate to call me on [Insert Details] or my supervisor on [Insert Details].

If you have any complaints concerning the manner in which this research project
is conducted and an independent person is preferred, you make contact the:

University’s Research Ethics Officer

[Insert Details].

Thank you for supporting this research initiative.

 382

A Study on Nurses’ Work Stressors, Work Support, and

Psychological Health and Well-being

The information from this questionnaire will be used to determine how workplace
supports can be used more effectively to reduce work stress and to enhance
overall well-being. Please provide the information as carefully as you can.

N.B. The completion and submission of this questionnaire to the

researcher will be understood to be informed consent to participate in the
this study.
 383

Please read the instructions carefully before completing each section. In order for this
research to be valid, each question should be answered. Please note there are no right
or wrong answers.

Demographic Information

1. Please circle the appropriate responses below.

Gender Male Female

Age < 25 yrs 25-29 yrs 30-39yrs 40-49yrs > 50 yrs
Employment Full-time Permanent Casual Graduate Undergraduat
Status Part-Time Placement e Placement
Years practicing < 1 yr 1-2yrs 2-5 yrs 5-10yrs > 10yrs
in the nursing
field?
Length of time 1-6 6-12 months 1-2 yrs 2-5yrs > 5 yrs
working within months
this particular
hospital?
Length of time 1-6 6-12 months 1-2 yrs 2-5yrs > 5 yrs
working in this months
particular ward?

2. Please circle the hospital you work at and the ward that you currently work in.

Hospital Royal Brisbane Princess Alexandra Redland Hospital

Hospital Hospital

If Other hospital please state………………………………………………………………

Ward Orthopaedics Emergency Renal Unit Surgical Operating

Department Ward Theatre
Children’s Ward ICU Rehabilitation Adult Ward Maternity Other
Unit Ward

If Other, please state………………………………………………………...…

 384

Health Professions Stress Inventory (Adapted from Wolfgang, 1998)

The following survey consists of 31 job situations health professionals might be expected
to encounter in the performance of their duties.

Circle the number (from 0 to 4) that best describes how often you find each work
situation to be stressful.

Never/ Occasionally Often Usually Very Often

Rarely

1. Having so much work to do that 0 1 2 3 4

everything cannot be done well.

2. Experiencing conflicts with supervisors 0 1 2 3 4

and/or administrators.

3. Feeling ultimately responsible for patient 0 1 2 3 4

outcomes.

4. Not receiving the respect or recognition 0 1 2 3 4

that you deserve from physicians.

5. Being uncertain about what to tell a 0 1 2 3 4

patient or family about the patient’s
condition and/or treatment.

6. Caring for the emotional needs of 0 1 2 3 4

patients.

7. Disagreeing with other health 0 1 2 3 4

professionals concerning the treatment of a
patient.

8. Not having opportunities to share feelings 0 1 2 3 4

and experiences with colleagues.

9. Experiencing conflicts with coworkers. 0 1 2 3 4

10. Having job duties which conflict with 0 1 2 3 4

family responsibilities.

11. Allowing personal feelings/emotions to 0 1 2 3 4

interfere with the care of patients.

12. Keeping up with new developments in 0 1 2 3 4

order to maintain professional competence.

13. Feeling that opportunities for 0 1 2 3 4

advancement on the job are poor.

14. Trying to meet society’s expectations for 0 1 2 3 4

high-quality medical care.

15. Supervising the performance of 0 1 2 3 4

coworkers.

16. Dealing with ‘difficult’ patients. 0 1 2 3 4

17. Not being recognised or accepted as a 0 1 2 3 4

true health professional by other health
professionals.
 385

Health Professions Stress Inventory (continued)

Never/ Occasionally Often Usually Very Often

Rarely

18. Being inadequately prepared to meet 0 1 2 3 4

the needs of patients.

19. Possessing inadequate information 0 1 2 3 4

regarding a patient’s medical condition.

20. Not receiving adequate feedback on 0 1 2 3 4

your job performance.

21. Not having enough staff to adequately 0 1 2 3 4

provide necessary services.

22. Having non-health professionals 0 1 2 3 4

determine the way you must practice your
profession.

23. Not knowing what type of job 0 1 2 3 4

performance is expected.

24. Being interrupted by phone calls or 0 1 2 3 4

people while performing job duties.

25. Not being allowed to participate in 0 1 2 3 4

making decisions about your job.

26. Not being challenged by your work. 0 1 2 3 4

27. Feeling that you are inadequately paid 0 1 2 3 4

as a health professional.

28. Caring for terminally ill patients. 0 1 2 3 4

29. Not being able to use your abilities to 0 1 2 3 4

the fullest extent on the job.

30. Fearing that a mistake will be made in 0 1 2 3 4

the treatment of a patient.

31. Dealing with difficult/demanding 0 1 2 3 4

relatives.
 386

Maslach Burnout Inventory (Maslach, Jackson, & Leiter, 1996)

 387

Maslach Burnout Inventory (continued)

 388

Occupational Roles Questionnaire (Adapted from Osipow & Spokane,1987)

Below are 30 statements related to your role at work. Read each statement carefully. For
each statement, circle the number (from 1 to 5) that best describes how often you feel
this way about your role at work.

Never/ Occasionally Often Usually Most of the

Rarely time

1. At work I am expected to do too many different 1 2 3 4 5

tasks in too little time.

2. I feel that my job responsibilities are increasing. 1 2 3 4 5

3. I am expected to perform tasks on my job for which 1 2 3 4 5

I have never been trained.

4. I have to take work home with me. 1 2 3 4 5

5. I have the resources I need to get my job done. 1 2 3 4 5

6. I feel competent in what I do. 1 2 3 4 5

7. I work under tight deadlines. 1 2 3 4 5

8. I wish I had more help to deal with the demands 1 2 3 4 5

placed upon me at work.

9. My job requires me to work in several equally 1 2 3 4 5

important areas at once.

10. I am expected to do more work than is 1 2 3 4 5

reasonable.

11. I feel conflict between what my employer expects 1 2 3 4 5

me to do and what I think is right or proper.

12. I feel caught between factions at work. 1 2 3 4 5

13. I have more than one person telling me what to 1 2 3 4 5

do.

14. I feel I have a stake in the success of this hospital. 1 2 3 4 5

15. I feel good about the work I do. 1 2 3 4 5

16. My supervisors have conflicting ideas about what I 1 2 3 4 5

should be doing.

17. I am proud of what I do for a living. 1 2 3 4 5

18. It is clear who really runs things where I work. 1 2 3 4 5

19. I have divided loyalties on my job. 1 2 3 4 5

20. The work I do has as much pay-off for me as my 1 2 3 4 5

employer.

21. My supervisor provides me with useful feedback 1 2 3 4 5

about my performance.
 389

Occupational Roles Questionnaire (continued)

Never/ Occasionally Often Usually Most of the

Rarely time

22. It is clear to me what I have to do to get ahead. 1 2 3 4 5

23. I am uncertain about what I am supposed to 1 2 3 4 5

accomplish in my work.

24. When faced with several tasks I know which 1 2 3 4 5

should be done first.

25. I know where to begin a new project when it is 1 2 3 4 5

assigned to me.

26. My supervisor asks for one thing, but really wants 1 2 3 4 5

another.

27. I understand what is acceptable personal 1 2 3 4 5

behaviour on my job (e.g. dress, interpersonal
relations).

28. The priorities of my job are clear to me. 1 2 3 4 5

29. I have a clear understanding of how my supervisor 1 2 3 4 5

wants me to spend my work time.

30. I know the basis on which I am evaluated. 1 2 3 4 5

Coworker Support Scale (Adapted from Shinn, Wong, Simko, and Ortiz-Torres’
(1989) Supervisor Support Scale, Ray and Miller’s (1994) Supervisor/Coworker
Support Scale, and King, Mattimore, King, and Adams’ (1995) Family Support
Inventory for Workers)

The following 12 statements are designed to examine the level of support you receive at
work from your coworkers. Indicate the degree to which your coworkers support you at
work by circling a response from 1 (strongly disagree) to 5 (strongly agree).

Strongly Somewhat Neither Somewhat Strongly Agree

Disagree Disagree Agree or Agree
Disagree

1. My coworkers listen to my problems. 1 2 3 4 5

2. My coworkers are understanding and sympathetic. 1 2 3 4 5

3. My coworkers respect me. 1 2 3 4 5

4. My coworkers appreciate the work I do. 1 2 3 4 5

5. My coworkers seem to make time for me if I need to 1 2 3 4 5

discuss my work.

6. I feel comfortable asking my coworkers for help if I 1 2 3 4 5

have a problem.

7. When I’m frustrated by some aspect of my work, 1 2 3 4 5

my coworkers try to understand.
 390

Coworker Support Scale (continued)

Strongly Somewhat Neither Somewhat Strongly Agree

Disagree Disagree Agree or Agree
Disagree

8. My coworkers will help me figure out a work 1 2 3 4 5

problem.

9. My coworkers cooperate with me to get things done 1 2 3 4 5

at work.

10. If my job duties become very demanding, my 1 2 3 4 5

coworkers will take on extra work responsibilities to
help me.

11. My coworkers can be relied on to help when 1 2 3 4 5

things get tough at work.

12. My coworkers share useful ideas or advice with 1 2 3 4 5

me.

Supervisor Support Scale (Adapted from Shinn, Wong, Simko, and Ortiz-Torres’
(1989) Supervisor Support Scale, Ray and Miller’s (1994) Supervisor/Coworker
Support Scale, and King, Mattimore, King, and Adams’ (1995) Family Support
Inventory for Workers)

The following 12 statements are designed to examine the level of support you receive at
work from your supervisor (e.g., your NPC/CNC). Indicate the degree to which your
supervisor supports you at work by circling a response from 1 (strongly disagree) to 5
(strongly agree).

Strongly Somewhat Neither Somewhat Strongly Agree

Disagree Disagree Agree or Agree
Disagree

1. My supervisor listens to my problems. 1 2 3 4 5

2. My supervisor is understanding and 1 2 3 4 5

sympathetic.

3. My supervisor respects me. 1 2 3 4 5

4. My supervisor appreciates the work I do. 1 2 3 4 5

5. My supervisor always seems to make time for 1 2 3 4 5

me if I need to discuss my work.

6. I feel comfortable asking my supervisor for help 1 2 3 4 5

if I have a problem.

7. When I’m frustrated by some aspect of my 1 2 3 4 5

work, my supervisor tries to understand.

8. My supervisor will help me figure out a work 1 2 3 4 5

problem.

9. My supervisor cooperates with me to get things 1 2 3 4 5

done at work.
 391

Supervisor Support Scale (continued)

Strongly Somewhat Neither Somewhat Strongly Agree

Disagree Disagree Agree or Agree
Disagree

10. If my job duties become very demanding, my 1 2 3 4 5

supervisor will take on extra work responsibilities
to help me.

11. My supervisor can be relied on to help when 1 2 3 4 5

things get tough at work.

12. My supervisor shares ideas or advice with me. 1 2 3 4 5

You have now come to end of the questionnaire.

Thank you for taking the time to complete this questionnaire.
 392

APPENDIX B

The information sheet and consent form provided to nurses in Study 1 is presented below.
In addition, the coding instructions and coding sheet used to assist the independent
researcher in analysing the focus group data is provided.

Information Sheet

A Study on Nurses’ Work Stressors, Work Support, and

Psychological Health and Well-Being in Nurses

Before participating in the focus group, it is important you understand what this study hopes to
achieve. I am interested in examining nurses in order to discover which particular types of
workplace support are particularly effective in enabling nursing staff to cope with particular types
of workplace stress. By doing so we can answer a key unanswered question in the social support
literature that is, whether certain forms of workplace support are most beneficial following
specific kinds of stress.

The current research involves two studies. In Study 1, a 30-minute focus group will be
conducted. Study 2 will involve the completion of a questionnaire measuring work stress, work
support and your overall well-being. In the focus groups, you will be asked to discuss:

1. Which particular events within your daily working life cause you the most stress?
2. Which people within your work environment serve as a support to you when you are
experiencing a stressful day at work?
3. What do these people do that makes you feel supported?

The information gained from your participation in this discussion will assist the researcher’s
understanding of the most salient work stressors experienced by nurses and the types of support
your receive at work.

It is important that you understand that what is expressed in the focus group is strictly
confidential and therefore any information disclosed during the focus group is not to be taken
outside of the room. It is essential that you understand the importance of maintaining
confidentiality and anonymity before commencing. If at anytime you feel uncomfortable in
responding to a question, feel free to decline. Alternatively, you may withdraw from the focus
group discussion at any time. If you have any queries about the study itself, please do not hesitate
to call me on [Insert Contact Number] or my supervisor on [Insert Details].

The focus group will be held:

[Insert Details]

You are most welcome to attend this focus group.

Yours sincerely,

Rebecca Spooner-Lane
 393

Consent Form

A Study on Nurses’ Work Stressors, Work Support, and

Psychological Health and Well-Being

I understand that:
1) there are no risks associated with taking part in this research;

2) my participation is voluntary and I have the right to withdraw from

the research at any time;

3) my responses will only be available to the researcher;

4) the information that I may provide may be published in professional

publications, but never in such a way that I
can be identified. My anonymity is assured.

Thank you for taking the time to consider this request

I give my consent to participate in this focus group.

Participant’s Signature ................................................. Date........................................

 394

Coding Sheet Used to Analyse the Focus Group Data

Background
A series of focus group discussions were conducted with nurses from the Princess
Alexandra Hospital and Redland Hospital. Nurses were required to discuss three
questions:

1. Which particular events within your daily working life cause you the most stress?
2. Which people within your work environment serve as a support to you when you are
experiencing a stressful day at work?
3. What do these people do that makes you feel supported?

Content Analysis Procedure

1. The researcher listened to the ten tapes. The tapes were transcribed verbatim and
observational data was added to the transcripts (Refer to the shaded information in
brackets).
2. The main issues discussed for each question were recorded.
3. Broad overarching themes were established.
4. Definitions for these broad over arching themes were created.
5. The recording units (i.e., key words, sentences and phrases) were generated.

It should be noted that when analysing the work stress factors, sub-categories were also
identified.

The aim of this exercise is to check the accuracy/reliability of the coding criteria used to
analyse the focus group data. A set of steps for you to follow are provided below.

1. On each tape you will notice that the ward, hospital and date have been provided.
Listen to each tape one at a time. When listening to the tape, try to get an overall
‘feel’ of the focus group discussion. For example, listen to the tone of voice used
by each speaker. Take note of whether the participant is speaking positively or
negatively about a particular topic.
2. Read the corresponding transcript for the tape.
3. Code all the text. A coding table has been provided for you. Make a tally in the
‘frequency’ column to indicate the number of times a recording unit (i.e., key
words, sentences, and phrases) is mentioned.
4. Note in the ‘ward(s)’ column which ward discussed the recording unit.
5. Re-read the transcript and check for accuracy. Ensure that all the main categories
(and sub-categories have been identified).
6. Make a note of any inconsistencies, or difficulties that you had in assigning a
recording unit to a particular category.
7. Make a note of work stressors that are unique to a specific ward.
 Coding Sheet

Ward: Hospital: Date:

Q1. WHICH PARTICULAR EVENTS WITHIN YOUR DAILY WORKING LIFE CAUSE YOU THE MOST STRESS?
Work Stress Category
Job Conditions:
Demands associated with being in a nursing role, as well as the conditions associated with working as a nurse in a hospital.
Work Stress Subcategory Examples Frequency Ward(s)
(i.e., the no. of times a recording unit is mentioned)
Work Overload: Being Staff shortages, too much
assigned too many tasks or documentation, interruptions,
insufficient time to heavy patient load, too many
accomplish assigned tasks. responsibilities. Workload
increasing.
Role Conflict: Being required Being responsible for the running
to assume two or more of the department as a well as the
conflict roles or when patient’s well-being. Doing the
expectations of these roles work that clerks should be doing.
differ. Carrying a full workload and
teaching as well.
Doing other people’s jobs.
Role Ambiguity: A lack of Not being sure about what is
clarity about the expectations required of nurses.
associated with a role, the Not having access to written
way in which the role should protocol.
be fulfilled, or the Not knowing whether it is our role
consequences of role to enforce things (e.g., visiting
performance. hours).
Lack of Resources: Lack of beds, space, medical
Insufficient resources to carry equipment, experienced staff.
out nursing role successfully.

395
Work Stress Subcategory Examples Frequency Ward(s)
(i.e., the no. of times a recording unit is
mentioned)
Other Noise

396
Job Uncertainty:
Unexpected stressful events that often beyond a nurse’s control.
Work Stress Subcategory Examples Frequency Ward(s)
(i.e., the no. of times a recording unit is mentioned)
Unpredictability: The Rostering schedule, patient load,
unpredictability associated with interruptions, mistakes in treating a
the nursing role. patient.
Unfamiliar Problems: Being Treating medical conditions that
expected to perform tasks that you have not experienced before.
you have never done before. Using new medical equipment.
Doctor Availability: Doctors that Waiting for doctors after-hours.
are not available when needed. We ring the doctor up so many
times and they don’t answer. They
hold us up for a couple of hours.
They disappear.
Lack of communication: The There’s a lack of communication
lack of communication between between staff. There’s not enough
health care professionals and communication between
between nursing wards. departments.

397
Work Stress Subcategory Examples Frequency Ward(s)
(i.e., the no. of times a recording unit is mentioned)
Other

398
Interpersonal Conflict:
Conflicts that arise from working so closely with patients and their relatives, as well as doctors, all in times of acute stress conditions.
Work Stress Subcategory Examples Frequency Ward(s)
(i.e., the no. of times a recording unit is mentioned)
Difficult/Demanding Patients Demanding, attention-seeking,
psychiatric, acutely ill, abusive,
violent, intoxicated patients.
Stay too long, don’t take
responsibility for themselves.
Patients that expect too much of
nurses.
Difficult/Demanding Relatives that blame the nurse for a
Relatives patient’s condition. Dysfunctional
families. Relatives that expect too
much of nurses.
Demanding Doctors Different personalities, different
philosophies about how tasks
should be done, different
expectations. Demanding doctors
who want you to stop everything to
do what they want.
Other

399
Lack of Professional Recognition and Support:
The undervaluing of nurses’ skills, experience, and qualifications by other health professionals, especially doctors, and the insufficient provision of support.
Work Stress Subcategory Examples Frequency Ward(s)
(i.e., the no. of times a recording unit is mentioned)
Disrespectful Doctors: The They blame us for their mistakes.
lack of respect demonstrated Their tone of voice suggests that
by doctors towards nurses. we’re beneath them. They don’t
value our experience or skills.
They don’t recognise our
qualifications.
Lack of Support: The lack of We don’t receive any feedback or
support available to nurses at recognition for what we do. We
work. don’t get the support we need.
There is a lack of medical
supervision and guidance.
Insufficient wages: We don’t get paid enough to do
what we do.
Not adequately paid.
Other

400
 2. WHICH PEOPLE WITHIN YOUR WORK ENVIRONMENT SERVE AS A SUPPORT TO YOU WHEN YOU ARE EXPERIENCING A
STRESSFUL DAY AT WORK?
Sources of Support Examples Frequency Ward(s)
(i.e., the no. of times a recording unit is mentioned)
Nursing colleagues Nursing co-workers, other nursing
staff.
Nursing Supervisor CNC, NPC, head nurse, senior
nurse. Level 2 nurse. Bay person.
Health Professionals Doctors, staff counsellors, staff
educators, physiotherapists, allied
health, after-hours co-ordinator.
Non-Health Professionals: Ward persons

Other Patients, relatives.

401
 3. WHAT DO THESE PEOPLE DO THAT MAKES YOU FEEL SUPPORTED?
Types of Support Examples Frequency Ward(s)
(i.e., the no. of times a recording unit is mentioned)
Emotional Support: The Listening, identifying with the
emotional comfort an problem, understands the problem,
individual receives during a emphathises with the situation,
stressful situation that leads a shows respect, appreciates the
person to believe that they are work we do, reassurance, laughing
cared for and valued by together.
others.
Instrumental Support: The Getting some help or assistance
instrumental assistance that a with jobs, getting the advice to be
person receives as a result of able to do our jobs, sharing
being given the necessary experiences with others, problem
resources to cope with a solving.
stressful situation, or the
guidance, or advice to help
solve a problem.
Other Social gatherings outside of work.

402
 403

APPENDIX C

In the section below, the rating system used by experienced nurses to classify nurses’
work stressors into controllable and uncontrollable events is delineated.

Are the Stressful Events Commonly Experienced by Nurses Predominantly Controllable

or Uncontrollable?

The aim of this rating exercise is to examine stressful events commonly experienced by
nurses working in a hospital setting in order to identify which stressful events are
predominantly controllable (i.e., the occurrence or consequences of the stressful event
can be prevented) or uncontrollable (i.e., there is nothing that can be done to prevent the
event from occurring or lessen its consequences).

In Part A, you will examine job stressors that are thought to be specifically related to
nurses. These items were taken from Wolfgang’s (1988) Health Professions Stress
Inventory. In Part B, you will examine role stressors that are thought to be common
among most occupations. Items were taken from Osipow and Spokane’s (1987)
Occupational Roles Questionnaire.

Part A
Job-Specific Stress

Instructions
The stressful events outlined below have been broadly categorised using four broad work
stress factors: 1) Lack Of Professional Recognition; 2) Patient Care Uncertainty; 3) Job
Conditions; and 4) Interpersonal Conflict.

Briefly examine the definition provided for each broad work stress factor. Identify
whether each factor is a source of work stress that hospital nurses can generally control or
whether it is a source of work stress that is generally uncontrollable.

To assist you in making your decision, the types of stressful events that align to the broad
work stress category are provided. Look at each stressful event independently and
determine whether it is a controllable or uncontrollable event.

For each work stress factor, total the number of controllable events and total the number
of uncontrollable events.

If the total number of controllable events is greater than the total number of
uncontrollable events, this would indicate that the work stress factor is predominantly
controllable.

Alternatively, if the total number of uncontrollable events is greater than the total number
of controllable events, this would indicate that the work stress factor is predominantly
uncontrollable.
 404

In the column beside the work stress factor, place a ‘C’ for controllable if you believe the
source of stress is predominantly preventable. Place a ‘U’ for uncontrollable if you
perceive the source of work stress to be unpreventable.

Points to Remember.
1. There are no right or wrong answers.
2. It is important that you do not spend too much time in determining whether the
source of work stress or associated work stressors are controllable or
uncontrollable. Your initial reaction to the statement is probably your best
indication of your perception of the controllability of the event.
3. In some cases, you will feel that the work stressor could be controllable or
uncontrollable depending on the situation. Although this may be true, it is
necessary to identify whether overall, the stressor is predominantly controllable or
predominantly uncontrollable. You will need to make a choice one way or
another.

A controllable stressful event is one in which the occurrence or consequences of the

stressful event can be prevented. An uncontrollable stressful event is one in which there
is nothing that can be done to prevent the event from occurring or lessen its
consequences.

C = Controllable
U = Uncontrollable
Lack of Professional Recognition
The absence of opportunities for nurses to use their abilities fully and to
contribute to important decisions concerning their job, and the lack of
recognition and support provided to nurses by other health professionals.

Survey Items C = Controllable

U = Uncontrollable
Not being allowed to participate in making decisions about your job.
Not being able to use your abilities to the fullest extent on the job.
Not receiving adequate feedback on your job performance.
Not being challenged by your work.
Feeling that opportunities for advancement on the job are poor.
Not being recognised or accepted as a true health professional by other health
professionals.
Feeling that you are inadequately paid as a health professional.
Not receiving the respect or recognition that you deserve from physicians.
Not having opportunities to share feelings and experiences with colleagues.
Having non-health professionals determine the way you must practice your
profession.

Total Number of Controllable Events = ∋ Total Number of Uncontrollable Events = ∋

 405

C = Controllable
U = Uncontrollable
Patient Care Uncertainty
The unpredictability and uncertainty associated with caring for patients
and the patient’s family.

Survey Items C = Controllable

U = Uncontrollable
Being uncertain about what to tell a patient or their family about the patient’s
condition and/or treatment.
Dealing with difficult/demanding relatives.
Dealing with difficult patients.
Caring for terminally ill patients.
Allowing personal feelings or emotions to interfere with the care of patients.
Being inadequately prepared to meet the needs of patients.
Possessing inadequate information regarding a patient’s medical condition.
Fearing that a mistake will be made in the treatment of a patient.
Disagreeing with other health professionals concerning the treatment of a patient.

Total Number of Controllable Events = ∋ Total Number of Uncontrollable Events = ∋

C = Controllable
U = Uncontrollable
Job Conditions
The tasks and responsibilities that nurses’ undertake in addition to caring
for the physical needs of patients.

Survey Items C = Controllable

U = Uncontrollable
Being interrupted by phone calls or people while performing job duties.
Trying to meet society’s expectations for high quality medical care.
Feeling ultimately responsible for patient outcomes.
Supervising the performance of less experienced workers.
Caring for the emotional needs of patients.
Having so much work to do so that everything cannot be done well.
Not having enough staff to adequately provide necessary services.
Keeping up with new developments in order to maintain professional
competence.

Total Number of Controllable Events = ∋ Total Number of Uncontrollable Events = ∋

 406

C = Controllable
U = Uncontrollable
Interpersonal Conflict
The conflict that may arise as a result of working closely with other
nurses.

Survey Items C = Controllable

U = Uncontrollable
Experiencing conflicts with supervisors or administrators.
Experiencing conflicts with co-workers

Total Number of Controllable Events = ∋ Total Number of Uncontrollable Events = ∋

Part B
Generic Role Stressors

Instructions
The stressful events outlined below have been broadly categorised using three broad role
stress factors: 1) Role Overload, 2) Role Conflict, and 3) Role Ambiguity.

Briefly examine the definition provided for each broad role stress factor. Identify whether
each factor is a source of stress that hospital nurses can generally control or whether it is
a source of stress that is generally uncontrollable.

To assist you in making your decision, the types of stressful events that align to the broad
role stress category are provided. Look at each stressful event independently and
determine whether it is a controllable or uncontrollable event.

For each role stress factor, total the number of controllable events and total the number of
uncontrollable events.

If the total number of controllable events is greater than the total number of
uncontrollable events, this would indicate that the role stress factor is predominantly
controllable.

Alternatively, if the total number of uncontrollable events is greater than the total number
of controllable events, this would indicate that the role stress factor is predominantly
uncontrollable.

In the column beside the role stress factor, place a ‘C’ for controllable if you believe the
source of stress is predominantly preventable. Place a ‘U’ for uncontrollable if you
perceive the source of role stress to be unpreventable.
 407

Points to Remember.
1. There are no right or wrong answers.
2. It is important that you do not spend too much time in determining whether the
source of role stress or associated stressors are controllable or uncontrollable.
Your initial reaction to the statement is probably your best indication of your
perception of the controllability of the event.
3. In some cases, you will feel that the role stressor could be controllable or
uncontrollable depending on the situation. Although this may be true, it is
necessary to identify whether overall, the role stressor is predominantly
controllable or predominantly uncontrollable. You will need to make a choice one
way or another.

A controllable stressful event is one in which the occurrence or consequences of the

stressful event can be prevented. An uncontrollable stressful event is one in which there
is nothing that can be done to prevent the event from occurring or lessen its
consequences.

C = Controllable
U = Uncontrollable
Role Overload
This occurs when too many behaviours are expected of an individual, or
the behaviour expected is too complicated or difficult for the individual to
execute.

Survey Items C = Controllable

U = Uncontrollable
At work I am expected to do too many different tasks in too little time.
I feel that my job responsibilities are increasing.
I am expected to perform tasks on my job for which I have never been trained.
I have to take work home with me.
I have the resources I need to get my job done.
I feel competent in what I do.
I work under tight deadlines.
I wish I had more help to deal with demands placed upon me at work.
My job requires me to work in several equally important areas at once.
I am expected to do more work than is reasonable.

Total Number of Controllable Events = ∋ Total Number of Uncontrollable Events = ∋

 408

C = Controllable
U = Uncontrollable
Role Conflict
This occurs for an individual when a person in the work environment
communicates an expectation about how he or she should behave and
this expectation makes it difficult or impossible to fulfil another
behavioural expectation or set of expectations.

Survey Items C = Controllable

U = Uncontrollable
I feel conflict between what my employer expects me to do and what I think is
right or proper.
I feel caught between factions at work.
I have more than one person telling me what to do at work.
I feel I have a stake in the success of this hospital.
I feel good about the work I do.
My supervisors have conflicting ideas about what I should be doing.
I am proud of what I do for a living.
It is clear who really runs things where I work.
I have divided loyalties on my job.
The work I do has as much pay-off for me as for my employer.

Total Number of Controllable Events = ∋ Total Number of Uncontrollable Events = ∋

C = Controllable
U = Uncontrollable
Role Ambiguity
This results when there is inadequate, unclear, or confusing information
about expected role behaviours; unclear or confusing information about
what behaviours may enable the incumbent to fulfil role expectations;
or uncertainty about the consequences of certain role behaviours.

Survey Items C = Controllable

U = Uncontrollable
My supervisor provides me with useful feedback about my performance.
It is clear to me what I have to do to get ahead.
I am uncertain about what I am supposed to accomplish in my work.
When faced with several tasks, I know which should be done first.
I know where to begin a new project when it is assigned to me.
My supervisor asks for one thing, but really wants another.
I understand what is acceptable personal behaviour on my job (e.g., dress,
interpersonal relations).
The priorities of my job are clear to me.
I have a clear understanding of how my supervisor wants me to spend my work
time.
I know the basis on which I am evaluated.

Total Number of Controllable Events = ∋ Total Number of Uncontrollable Events = ∋

409
 325

REFERENCES

Ackerly, G. D., Burnell, J., Holder, D. C., & Kurdek, L. A. (1988). Burnout among

licensed psychologists. Professional Psychology: Research and Practice, 18, 624-

631.

Adali, E. & Priami, M. (2002). Burnout among nurses in intensive care units, internal

medicine wards and emergency departments in Greek hospitals. ICUs and

Nursing Web Journal, 11, 1-19.

Adomat, R., & Killingworth, A. (1994). Care of the critically ill patient: The impact of

stress on touch in intensive therapy units. Journal of Advanced Nursing, 19, 912-

922.

Aguinis, H. & Stone-Romero, E. F. (1997). Methodological artefacts in moderated

multiple regression and their effects on statistical power. Journal of Applied

Psychology, 82, 192-206.

Agüir, V. E., Pons, R. M., Echegaray, M. C., Ramos, D. B., & Sänchez, R. G. (2000).

Estresores laborales y bienestar psicologico: Impacto en la Enfermería

hospitalaria. [The impact of psychological welfare of nurses in hospitals caused

by work stress]. Revista de Emfermerâia, 23, 506-611.

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting

interactions. Newbury Park, CA: Sage.

Akhtar, S. & Lee, J. S. (2002). Confirmatory factor analysis and job burnout correlates of

the Health Professions Stress Inventory. Psychological Reports, 90, 243-250.

 326

Akroyd, D., Caison, A. & Adams, R. (2002). Patterns of burnout among U.S.

radiographers. Radiologic Technology, 73, 215-223.

Albrecht, T. L., & Halsey, J. (1991). Supporting the staff nurse under stress. Nursing

Management, 22 (7), 60-63.

Aldwin, C. M. (1994). Stress, coping and development: An integrative perspective. New

York: Guildford Press.

Alexander, D. E. (1983). Occupational stress, strain and coping among family physicians

and teaching in military and civilian settings. Unpublished doctoral dissertation.

Columbus: Ohio State University.

Altun, I. (2002). Burnout and nurses’ personal and professional values. Nursing Ethics,

9(3), 269-278.

Antonucci, T. C. (1990). Social support and social relationships. In R. H. Binstock & L.

K. George (Eds.). Handbook of aging and the social sciences (3rd ed., pp. 205-

226). Orlando, FL: Academic Press.

Antonucci, T. C., & Depner, C. S. (1982). Social support and informal helping

relationships. In T. Willis (Ed.), Basic processes in helping relationships (pp.

233-254). New York: Academic Press.

Argyle, M. (1992). Benefits produced by supportive social relationships. In H. O. F.

Veiel & U. Baumann (Eds.), The meaning and measurement of social support (pp.

13-32). New York: Hemisphere.

Armstrong, N. (1987). Coping with Diabetes Mellitus. Nursing Clinics of North America,

22, 559-568.
 327

Ashforth, B.E., & Lee, R. T. (1997). Burnout as a process: Commentary on Cordes,

Dougherty and Blum. Journal of Organizational Behavior, 18, 703-708.

Australian Institute of Health and Welfare (1999). Nursing Labour Force 1999. (National

Health Labour Force Series No. 20). Canberra: Australian Institute of Health and

Welfare.

Australian Institute of Health and Welfare (2002). Australia’s health 2002: The eighth

biennial health report of the Australian Institute of Health and Welfare. Canberra:

AIHW.

Australian Nursing Council Incorporated (2002, February). Federal Government Pre-

Budget Submission (ANF Publication No. 962, 19). In Senate Community Affairs

References Committee. The patient profession, time for action report on the

inquiry into nursing. Canberra: Senate Community Affairs References Committee

Secretariat.

Baba, V. V., Galperin, B. L., & Lituchy, T. R. (1999). Occupational mental health: A

study of work-related depression among nurses in the Caribbean. International

Journal of Nursing Studies, 36, 163-169.

Babbie, E. (1995). The practice of social research. Belmont, CA: Wadsworth Publishing

Company.

Bailey, R., & Clarke, M. (1989). Stress and coping in nursing. London: Chapman and

Hall.

Bakker, A. B., Killmer, C. H., Siegrist, J., & Schaufeli, W. B. (2000). Effort-reward

imbalance and burnout among nurses. Journal of Advanced Nursing, 31(4), 884-

891.
 328

Baldwin, P. J. (1999). Stress and the nurse. Nursing Standard, 13(16), 22-25.

Barrera, M., Jr. (1986). Distinctions between social support concepts, measures and

models. American Journal of Community Psychology, 14(4), 413-446.

Barrera, M., Jr., Sandler, I. N., & Ramsay, T. B. (1981). Preliminary development of a

scale of social support: Studies on college students. Journal of Community

Psychology, 9, 435-447.

Bartlett (1998). Stress: Perspectives and processes. Philadelphia: Open University Press.

Beck, A. (1967). Depression: Clinical, experimental and theoretical aspects. New York:

Harper & Row.

Beehr, T. A. (1985). The role of social support in coping with organisational stress. In T.

A. Beehr & R. S. Bhagat (Eds.), Human stress and cognition in organizations:

An integrated perspective (pp. 375-398). New York: Wiley.

Beehr, T. A. (1995). Psychological stress in the workplace. London: Routledge.

Beehr, T. A., Jex, S. A., Stacy, B. A., & Murray, M. A. (2000). Work stressors and

coworker support as predictors of individual strain and job performance. Journal

of Organizational Behavior, 21, 391-405.

Beehr, T. A., King, L. A., King, D. W. (1990). Social support and occupational stress:

Talking to supervisors. Journal of Vocational Behavior, 36, 61-81.

Beehr, T. A., & McGrath, J. E. (1992). Social support, occupational stress, and

anxiety. Anxiety, Stress, and Coping, 5, 7-19.

Beehr, T. A., Stacy, B. A., Murray, M. A., & Jex, S. M. (1996). Work stress and

coworker support as predictors of individual strain and job performance. Paper

presented at the annual meeting of the Academy of Management, Cincinnati, OH.

 329

Bennett, L., & Michie, P., & Kippax, S. (1991). Quantitative analysis of burnout and its

associated factors in AIDS nursing. AIDS Care 3, 181-192.

Berg, B. L. (1995). Qualitative research methods for the social sciences. Sydney: Allyn

and Bacon.

Berkman, L. F., & Syme, S. L. (1979). Social networks, host resistance, and mortality: A

nine-year follow-up study of Alameda County residents. American Journal of

Epidemiology, 109, 186-204.

Blau, G. (1981). An empirical investigation of job stress, social support, service length,

and job strain. Organizational Behavior and Human Performance, 27, 279-

302.

Bourbonnais, R., Comeau, M., Vézina, M., & Dion, G. (1998). Job strain, psychological

distress, and burnout in nurses. American Journal of Industrial Medicine, 34, 20-

28.

Boyd, B. J., & Pasley, B. K. (1989). Role stress as a contributor to burnout in child

care professionals. Child and Youth Care Quarterly, 18, 243-258.

Brandt, P., & Weinert, C. (1981). The PRQ: A social support measure. Nursing Research,

30(5), 277-280.

Brief, A. P., & George, J. M. (1994). Psychological stress and the workplace: A brief

comment on Lazarus’ outlook. In R. Crandall & P. L. Perrewe (Eds.),

Occupational Stress: A Handbook (pp. 15-19). Washington DC: Taylor and

Francis.

Brill, P. L. (1984). The need for an operational definition of burnout. Family and

Community Health, 6, 12-24.

 330

Brookings, J. B., & Bolton, B. (1988). Confirmatory factor analysis of the interpersonal

support evaluation list. American Journal of Community Psychology, 16, 137-147.

Bruhn, J. G. (1973). Obtaining and interpreting psychosocial data in studies and coronary

heart disease. In J. P. Naughton & H. K. Hellerstein (Eds.), Exercise testing and

exercise training in coronary heart disease (pp. 263-273). New York: Academic

Press.

Bruhn, J. G. (1991). People need people: Perspectives on the meaning and measurement

of social support. Integrative Physiological and Behavioral Science, 26, 325-329.

Bryant, C., Fairbrother, G., & Fenton, P. (2000). The relative influence of personal and

workplace descriptors on stress. British Journal of Nursing, 9(13), 876-880.

Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models: Applications and

data analysis methods. Newbury Park, CA: Sage.

Burckhardt, C. S. (1987). Coping strategies of the chronically ill. Nursing Clinics of

North America, 22, 543-549.

Burke, R. J. & Greenglass, E. (1993). Work stress, role conflict, social support and

psychological burnout among teachers. Psychological Reports, 73, 371-380.

Burke, R. J. & Greenglass, E. (1995). A longitudinal examination of the Cherniss model

of psychological burnout. Social Science Medicine, 40(10), 1357-1363.

Burke, R. J., & Greenglass, E. (2001). Hospital restructuring and psychological

burnout in nursing staff. Equal Opportunities International, 20, 61-71.

Burke, R. J., Shearer, J., & Deszca, E. (1984). Burnout among men and women in

police work: An examination of the Cherniss model. Journal of Health and

Human Resources Administration, 7, 162-188.

 331

Büssing, A., & Glaser, J. (1999). Work stressors in nursing in the course of redesign:

Implications for burnout and interactional stress. European Journal of Work and

Organizational Psychology, 8(3), 401-426.

Butterworth, T., Carson, J., Jeacock, J., White, E., & Clements, A. (1999). Stress, coping,

burnout and job satisfaction in British nurses: Findings from the Clinical

Supervision Evaluation Project. Stress Medicine, 15, 27-33.

Buunk, B. P., & Schaufeli, W. B. (1993). Burnout: A perspective from social comparison

theory. In W. B. Schaufeli, C. Maslach, & T. Marek (Eds.), Professional burnout:

Recent developments in theory and research (pp. 53-69). Washington, DC: Taylor

& Francis.

Buunk, B. P., Van Yperen, N. W., Taylor, S. E., & Collins, R. B. (1991). The drive

upward in social comparison theory revisited. European Journal of Social

Psychology, 21, 529-546.

Byrne, B. M. (1993). The Maslach Burnout Inventory: Testing for factorial validity and

invariance across elementary, intermediate and secondary teachers. Journal of

Occupational and Organizational Psychology, 66, 197-212.

Byrne, B. M. (1994). Burnout: Testing for the validity, replication, and invariance of

causal structure across elementary, intermediate and secondary teachers.

American Educational Research Journal, 31, 645-673.

Callaghan, P., & Field, M. (1991). Organization and stress among mental nurses. Nursing

Times, 87, 50.

 332

Capel, S. A., Sisley, B. L., & Desertain, G. S. (1987). The relationship of role conflict

and role ambiguity to burnout in high school basketball coachers. Journal of

Sport Psychology, 9, 106-117.

Caplan, G. (1974). Support systems and community mental health. New York:

Behavioral Publications.

Caplan, R. (1983). Person-environment fit: Past, present and future. In C. Cooper (Ed.).

Stress research: Issues for the eighties. Chichester: Wiley.

Caplan, R. D., Cobb, S., French, J. R. P., Van Harrison, R., & Pinneau, S. R. (1975). Job

demands and worker health. Ann Arbor: Institute for Social Research.

Capner, M., & Caltabiano, M. L. (1993). Factors affecting the progression towards

burnout: A comparison of professional and volunteer counsellors. Psychological

Reports, 73, 555-561.

Carayon, P. (1995). Chronic effect of job control, supervisor social support, and work

pressure on officer worker stress. In S. L. Sauter & L. R. Murphy (Eds.),

Organizational risk factors for job stress (pp. 357-370). Washington: American

Psychological Association.

Carr, A., Roseingrave, O., & Fitzgerald, M. X. (1996). Factors associated with stress

responses in caring for patients with cystic fibrosis. Irish Journal of Psychology,

17, 241-250.

Carveth, W. B., & Gottlieb, B. H. (1979). The measurement of social support and its

relation to stress. Canadian Journal of Behavioral Science, 11, 179-187.

Cash, D. (1991). A study of the relationship of demographics, personality, and role

stress in burnout in intensive care unit nurses. Ann Arbor, MI: UMI.
 333

Cassam, N. H., & Hackett, T. P. (1972). Sources of tension for the CCU nurse. American

Journal of Nursing, 72, 1426-1430.

Ceslowitz, S. B. (1989). Burnout and coping strategies among hospital staff nurses.

Journal of Advanced Nursing, 14, 553-557.

Chang, E., & Hancock, K. (2003). Role stress and role ambiguity in new nursing

graduates in Australia. Nursing and Health Sciences, 5, 155-163.

Chapman, J. (1993). Collegial support linked to reduction of job stress. Nurse

Management, 24, 52-54.

Chay, Y. W. (1993). Social support, individual differences and well-being: A study of

small business entrepreneurs and employees. Journal of Occupational and

Organizational Psychology, 66, 285-302.

Cheng, Y., Kawachi, I., Coakley, E. H., Schwartz, J., & Colditz, G. (2000). Association

between psychosocial work characteristics and health functioning in American

women: Prospective study. British Medical Journal, 320, 1432-1436.

Cherniss, C. (1980). Professional burnout in the human service organizations. New

York: Praeger.

Cherniss, C. (1995). Beyond burnout: Helping teachers, nurses, therapists, and lawyers

recover from stress and disillusionment. Routledge: New York.

Cheuk, W. H., & Wong, K. S. (1995). Stress, social support and teacher burnout in

Macau. Current Psychology: Developmental, Learning, Personality, Social, 14,

42-46.
 334

Chung, M. C., & Corbett, J. (1998). The burnout of nursing staff working with

challenging behaviour clients in hospital-based bungalows and a community unit.

International Journal of Nursing Studies, 35, 56-64.

Cobb, S. (1976). Social support as a moderator of life stress. Psychosomatic Medicine, 4,

300-314.

Cobb, S. (1979). Social support and health through the life course. In M. W. Riley (Ed.),

Aging from birth to death: Interdisciplinary perspectives (pp. 93-106). Boulder,

CO: Westview.

Cohen, S. (1988). Psychosocial models of the role of social support in the etiology of

physical disease. Health Psychology, 7, 269-297.

Cohen-Mansfield, J. (1989). Sources of satisfaction and stress in nursing home

caregivers: Preliminary results. Journal of Advanced Nursing, 72, 1426-1430.

Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the

behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum Associates.

Cohen, L. H., McGowan, J., Fooskas, S., & Rose, S. (1984). Positive life events and

social support and the relationship between life stress and psychological disorder.

American Journal of Community Psychology, 12, 564-587.

Cohen, S., & Hoberman, H. (1983). Positive events and social support as buffers of life

change stress. Journal of Applied Social Psychology, 13, 99-125.

Cohen, S., & McKay, G. (1984). Social support, stress, and the buffering hypothesis: A

theoretical analysis. In A. Baum, J. E. Singer, & S. E. Taylor (Eds.), Handbook of

psychology and health (vol 4) (pp.253-267). Hillsdale, NJ: Erlbaum.

 335

Cohen, S., Mermelstein, R., Kamarck, T., & Hoberman, H. (1985). Measuring the

functional components of social support. In I. G. Sarason & B. R. Sarason (Eds.).

Social support: Theory, research and application (pp. 73-94). The Hague:

Martinus Nijhoff.

Cohen, S., Underwood, L. G., & Gottlieb, B. H. (2000). Social support measurement

and intervention. Oxford: Oxford University Press.

Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis.

Psychological Bulletin, 98 (2), 310-357.

Cole, F. L., Slocumb, E. M., & Mastey, J. M. (2001). A measure of critical care nurses’

post code stress. Journal of Advanced Nursing, 34(3), 281-294.

Collins, M. A. (1996). The relation of work stress, hardiness, and burnout among full-

time hospital staff. Journal of Nursing Staff Development, 12(2), 81-85.

Comrey, A. L., & Lee, H. B. (1992). A first course in factor analysis. (2nd ed.) Hillsdale,

NJ: Erlbaum.

Considine, G., & Buchanan, J. (1999). The hidden costs of understaffing: An analysis of

contemporary nurses working conditions in Victoria. Victoria: Australian Nursing

Federation.

Constantini, A., Solano, l., Di Napoli, R., & Bosco, A. (1997). Relationship between

hardiness and risk of burnout in a sample of 92 nurses working in oncology and

AIDS wards. Psychotherapy and Psychosomatics, 66, 78-82.

Cordes, C. L., & Dougherty, T. W. (1993). A review and integration of research on job

burnout. Academy of Management Review, 18, 621-656.

 336

Cordes, C. L., & Dougherty, T. W., & Blum, M. (1997). Patterns of burnout among

managers and professionals: A comparison of models. Journal of Organizational

Behavior, 18, 685-701.

Corrigan, P.W., Homes, E. P., & Luchins, D. (1995). Burnout and collegial support in

state psychiatric hospital staff. Journal of Clinical Psychology, 51, 703-710.

Corrigan, P.W., Homes, E. P., Luchins, D., Buican, B., Basit, A., & Parks, J. J. (1994).

Staff burnout in a psychiatric hospital: A cross-lagged panel design. Journal of

Organizational Behavior, 15, 65-74.

Coyne, J. C., & Lazarus, R. S. (1980). Cognitive style, stress perception, and coping. In I.

L. Kutash, L. B. Schlesinger, & associates (Eds.), Handbook of stress and anxiety

(pp. 144-158). San Francisco: Jossey-Bass.

Cox, T. (1978). Stress. London: MacMillan Press.

Cox, T., Kuk, G., & Leiter, M. P. (1993). Burnout, health, work stress, and organizational

healthiness. In W. B. Schaufeli, C. Maslach, & T. Marek (Eds.), Professional

burnout: Recent developments in theory and research (pp. 177-197). Washington,

DC: Taylor & Francis.

Crane, S. J., & Iwanicki, E. F. (1986). Perceived role conflict, role ambiguity, and

burnout among special education teachers. Rase, 7, 24-31.

Craven, R. F., & Hirnle, C. J. (2000). Fundamentals of nursing. New York: Lippincott,

Williams, and Wilkins.

Cronin-Stubbs, D., & Brophy, E. B. (1985). Burnout: Can social support save the

psychiatric nurse? Journal of Psychosocial Nursing and Mental Health Services,

23, 8-13.
 337

Cutrona, C. E. (1990). Stress and social support – In search of optimal matching.

Journal of Social and Clinical Psychology, 9(1), 3-14.

Cutrona, C. E. & Russell, D.W. (1987). The provision of social relationships and

adaptation to stress. In W. H. Jones & D. Perlman (Eds.), Advances in personal

relationships (Vol. 1, pp. 37-67). Greenwich, CT: JAI Press.

Cutrona, C. E. & Russell, D.W. (1990). Type of social support and specific stress:

Toward a theory of optimal matching. In B. R. Sarason, I. G. Sarason, & G. R.

Pierce (Eds.), Social support: An interactional view (pp. 319-365). New York:

Wiley.

Dallender, J., Nolan, P., Soares, J., Thomsen, S., & Arnetz, B. (1999). A comparative

study of the perceptions of British mental health nurses and psychiatrists of their

work environment. Journal of Advanced Nursing, 29, 36-43.

Deaux, K. (1984). From individual differences to social categories: Analysis of a

decade’s research on gender. American Psychologist, 39, 105-116.

Decker, P. J., & Borgen, F. H. (1993). Dimensions of work appraisal: Stress, strain, and

coping, job satisfaction, and negative affectivity. Journal of Counseling

Psychology, 40, 470-478.

De Jonge, J., Janssen, P. P. M., & Van Breukelen, G. J. P. (1996). Testing the demand-

control-support model among health care professionals. A structural equation

model. Work and Stress, 10, 209-224.

De Jonge, J. & Schaufeli, W. B. (1998). Job characteristics and employee well-being: A

test of Warr’s Vitamin model in health care workers using structural equation

modelling. Journal of Organizational Behavior, 19, 387-407.

 338

Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2000). A model of

burnout and life satisfaction amongst nurses. Journal of Advanced Nursing, 32(2),

454-494.

DePew, C. L., Gordon, M., Yoder, L. H., & Goodwin, C. W. (1999). The relationship of

burnout, stress, and hardiness in nurses in a military medical center: A replicated

descriptive study. Journal of Burn Care & Rehabilitation, 20, 515-522.

De Rijk, A. E., LeBlanc, P. M., Schaufeli, W. B., & De Jonge, J. (1998). Active coping

and need for control as moderators of the demand-control model: Effects on

burnout. Journal of Occupational and Organizational Psychology, 71, 1-18.

Descamp, K. & Thomas, C. (1993). Buffering nursing stress through play at work.

Western Journal of Nursing Research, 15, 619-627.

Dewe, P. (1989). Developing stress management programs: What can we learn from

recent research? Journal of Occupational Health Safety – Australia and New

Zealand, 5, 493-499.

Dey, I. (1993). Qualitative data analysis: A user-friendly guide for social scientist.

London: Routledge.

Dignam, J. T., Barerra, M. Jr, & West, S. G. (1986). Occupational stress, social support,

and burnout among correctional officers. American Journal of Community

Psychology, 14, 177-193.

Dignam, J. T., & West, S. G. (1988). Social support in the workplace: Tests of six

theoretical models. American Journal of Community Psychology, 16, 701-724.

 339

Dolan, S. L., & Renaud, S., (1992). Individual, organizational and social determinants of

managerial burnout: A multivariate approach. Journal of Social Behavior and

Personality, 7, 95-110.

Dunkel-Schetter, C., & Bennett, T. L. (1990). Differentiating the cognitive and

behavioral aspects of social support. In B. R. Sarason, I. G. Sarason, & G. R.

Pierce (Eds.), Social support: An interactional view (pp. 267-296). Canada: John

Wiley & Sons.

Duquette, A., Kerouac, S., Sandhu, B. K., & Beaudet, L. (1994). Burnout: A review

of empirical knowledge. Issues in Mental Health Nursing, 15, 337-358.

Duquette, A., Kerouac, S., Sandhu, B. K., Ducharme, F., & Saulnier, P. (1995).

Psychosocial determinants of burnout in geriatric nursing. International Journal

of Nursing Studies, 32(5), 443-456.

Eastburg, M. C., Williamson, M., Gorsuch, R., & Ridley, C. (1994). Social support,

personality, and burnout in nurses. Journal of Applied Psychology, 24, 1233-

1250.

Ebbinghaus, M. (1996). Conceptualising burnout: Development and validation of a

measurement instrument for applications in various work domains. Unpublished

manuscript, University of Oldenberg at Oldenberg, Germany.

Eckenrode, J. (1984). Impact of chronic and acute stressors on daily reports of mood.

Journal of Personality and Social Psychology, 46(4), 907-918.

Edwards, D., & Burnard, P. (2003). A systematic review of stress and stress management

interventions for mental health nurses. Journal of Advanced Nursing, 42, 169-200.
 340

Edwards, D., Burnard, P., Coyle, D, Fothergill, A., & Hannigan, B. (2000a). Stress and

burnout in community mental health nursing: A review of the literature. Journal

of Psychiatric and Mental Health Nursing, 7, 7-14.

Edwards, D., Burnard, P., Coyle, D, Fothergill, A., & Hannigan, B. (2000b). Stressors,

moderators and stress outcomes: Findings from the All-Wales Community Mental

Health Nurse Study. Journal of Psychiatric and Mental Health Nursing, 7, 529-

537.

Eells, T. D., Lacefield, P., & Maxey, J. (1994). Symptom correlates and factor structure

of the Health Professions Stress Inventory. Psychological Reports, 75, 1563-1568.

El-Bassel, N., Guterman, N., Bargal, D., & Su, K. (1998). Main and buffering effects of

emotional support on job- and health-related strains: A national survey of Israeli

Social Workers. Employee Assistance Quarterly, 13(3), 1-18.

Ellis, B. H., & Miller, K. I. (1994). Supportive communication among nurses: Effects on

commitment, burnout and retention. Health Communication, 6(2), 77-96.

Emslie, C., Hunt, K., & Macintyre, S. (1999). Problematizing gender, work and health:

The relationship between gender, occupational grade, working conditions and

minor morbidity in full-time bank employees. Social Science and Medicine, 48,

33-48.

Escot, C., Artero, S., & Gandubert, C., Boulenger, J. P., & Ritchie, K. (2001). Stress

levels in nursing staff working in oncology. Stress and Health, 17, 273-279.

Etsion, D., & Westman, M. (1994). Social support and sense of control as moderators of

the stress-burnout relationship in military careers. Journal of Social Behavior and

Personality, 9, 639-656.
 341

Etzion, D. (1987). Burnout: The hidden agenda of human distress (Working Paper No.

930/987). Tel-Aviv: Tel-Aviv University.

Fagin, L., Carson, J., Leary, J., De Villiers, N., Bartlett, H., O’Mally, M. W.,

McElfatrick, S., & Brown, D. (1996). Stress, coping, and burnout in mental health

nurses: Findings from three research studies. International Journal of Social

Psychiatry, 42(2), 102-111.

Farrell, G. (1999). Aggression in clinical settings: nurses’ views – A follow-up study.

Journal of Advanced Nursing, 29, 532-541.

Farrington, A. (1995). Stress and nursing. British Journal of Nursing, 4 (10), 574-578.

Fenlason, K. J., & Beehr, T. A. (1994). Social support and occupational stress: Effects of

talking to others. Journal of Organizational Behavior, 15, 157-175.

Fimian, M. J. (1984). Organizational variables related to stress and burnout in

community-based programs. Education and Training of the Mentally Retarded,

19, 201-209.

Firth, H., McKeown, P., McIntee, J., & Britton, P. (1987). Professional depression,

“burnout” and personality in long stay nursing. International Journal of Nursing

Studies, 24, 227-237.

Fisher, J. D., Nadler, A., & Whitcher-Alagna, S. (1982). Recipient reactions to aid.

Psychological Bulletin, 91, 27-54.

Folkman, S., & Lazarus, R. (1980). An analysis of coping in a middle aged community

sample. Journal of Health and Social Behavior, 21, 219-239.

 342

Fong, C. A. (1993) A longitudinal study of the relationships between overload, social

support and burnout among nurse educators. Journal of Nursing Education, 32,

24-29.

Foxall, M. J., Zimmerman, L., Standle, R., & Bene, B. (1990). A comparison of

frequency and sources of nursing job stress perceived by intensive care, hospice

and medical-surgical nurses. Journal of Advanced Nursing, 15, 577-584.

French, S. E., Lenton, R., Walters, V., & Eyles, J. (2000). An empirical evaluation of an

expanded nursing stress scale. Journal of Nursing Measurement, 8(2), 161-178.

Freudenberger, H. J. (1974). Staff burn-out. Journal of Social Issues, 30, 159-165.

Freudenberger, H. J. (1975). The staff burnout syndrome in alternative institutions.

Psychotherapy: Theory, Research and Practice, 12, 72-83.

Freudenberger, H. J. (1977a). Burnout: Occupational hazard of the child care worker.

Child Care Quarterly, 56, 90-99.

Freudenberger, H. J. (1977b). Burnout-out: The organizational menace. Training and

Development Journal, 31, 26-27.

Frisch, S., Dembeck, P., & Shannon, V. (1991). The head nurse: Perceptions of stress and

ways of coping. Canadian Journal of Nursing Administration, 4(4), 6-13.

Frone, M. R., Russell, M., & Cooper, M. L. (1995). Job stress, job involvement and

employee health: A test of identity theory. Journal of Occupational and

Organizational Psychology, 68, 1-11.

Gaines, J., & Jermier, J. M. (1983). Emotional exhaustion in a high stress organization.

Academy of Management Journal, 26, 567-586.

 343

Ganster, D. C., Fusilier, M. R., & Mayes, B. T., (1986). Role of social support in the

experience of stress at work. Journal of Applied Psychology, 71, 102-110.

Ganster, D. C., & Schaubroeck, J. (1991). Work, stress and employee health. Journal of

Management, 17, 235-271.

Garden, A. M. (1989). Burnout: The effect of psychological type on research findings.

Journal of Occupational Psychology, 62, 223-234.

Garrett, D., & McDaniel, A. A. (2001). A new look at nurse burnout: The effects of

environmental uncertainty and social climate. Journal of Nursing Administration,

31, 91-96.

Gaudine, A. P. (2000). What do nurses mean by workload and work overload? Canadian

Journal of Nursing Leadership, 13, 22-27.

Gil-Monte, P. R., & Peiro, J. M. (1999). Testing the factorial validity of the MBI: Results

for a multi-occupational sample. Psicothema, 11, 679-689.

Gil-Monte, P. R., Valcarcel, P., & Zornova, A. (1995). Role stress: Burnout antecedents

in nursing professionals. In J. M. Peiro, & F. Prierto (Eds.), Work and

organizational psychology (pp. 77-87). London: Taylor and Francis.

Glass, D., McKnight, J. D., & Validmarsdottir, H. (1993). Depression, burnout and

perceptions of control in hospital nurses. Journal of Consulting Clinical

Psychology, 61, 147-155.

Gmelch, W. H., & Gates, G. (1998). The impact of personal, professional, and

organizational characteristics on administrator burnout. Journal of Educational

Administration, 36(2), 146-159.

 344

Goddard, R., & O’Brien, P. (in press). Beginning teachers’ perceptions of their work,

well-being, and intention to leave. Asia-Pacific Journal of Teacher Education and

Development.

Goddard, R., Patton, W., & Creed, P. (2000). Case manager burnout in the Australian Job

Network. Journal of Applied Health Behaviour, 2(2) 1-6.

Golec, R. R. (1983). Occupational role stress in groups of stable and recently transferred

airline employees. Unpublished master’s thesis, University of Maryland, College

Park.

Golembiewski, R. (1989). Highlighting two models of burnout. Group and Organization

Studies, 14(1), 5-14.

Golembiewski, R., & Munzenrider, R., (1981). Efficacy of three versions of one burnout

measure: MBI as total score, sub-scale scores, or phases? Journal of Health and

Human Resources Administration, 7, 290-323.

Golembiewski, R., & Munzenrider, R., (1984). Phases of psychological burn-out and

organizational covariants. Organizational Development Journal, 4, 25-30.

Golembiewski, R., & Munzenrider, R., (1988). Phases of burnout: Developments in

concepts and applications. New York: Praeger.

Golembiewski, R., Munzenrider, R., & Stevenson, R. (1986). Stress in organizations:

Towards a phase model of burnout. New York: Praeger.

Goodwin, L. D., & Goodwin, W. L. (1985). Statistical techniques in AERJ articles, 1979-

1983. The preparation of graduate students to read the educational research

literature. Educational Researcher, 14, 5-11.

 345

Gore, S. (1985). Social support and styles of coping with stress. In S. Cohen & S. L.

Syme (Eds.), Social support and health (pp. 263-278). Orlando, FL: Academic

Press.

Gottlieb, B. H. (1983). Social support strategies. Beverly Hills, CA: Sage.

Gottlieb, B. H. (1985). Social networks and social support: An overview of research,

practice, and policy implications. Health Education Quarterly, 12, 5-22.

Gray-Toft, P., & Anderson, J. G. (1981a). The Nursing Stress Scale: Development of an

instrument. Journal of Behavioral Assessment, 3, 11-23.

Gray-Toft, P., & Anderson, J. G. (1981b). Stress among hospital nursing staff: Its causes

and effects. Social Science and Medicine, 15A, 639-647.

Gray-Toft, P., & Anderson, J. G. (1985). Organizational stress in the hospital:

Development of a model for diagnosis and prediction. Health Services Research,

19, 753-774.

Green, D., & Walkey, F. H. (1988). Confirmation of the three factor structuring of the

Maslach Burnout Inventory. Educational and Psychological Measurement, 48,

579-585.

Greenglass, E. R. (1982). A world of difference: Gender roles in perspective. New York:

Wiley.

Greenglass, E. R. (1991). Burnout and gender: Theoretical implications. Canadian

Psychology, 32, 562-572.

Greenglass, E., Burke, R. J., & Fiksenbaum, L. (2001). Workload and burnout in nurses.

Journal of Community and Applied Social Psychology, 11, 211-215.

 346

Greenglass, E., Burke, R. J., & Konarski, R. (1997). The impact of social support on the

development of burnout in teachers: Examination of a model. Work and Stress, 11

(3), 267-278.

Greenglass, E., Burke, R. J., & Ondrack, M. (1990). A gender-role perspective of coping

and burnout. Applied Psychology: An International Review, 39(1), 5-27.

Greenglass, E., Fiksenbaum, L., & Burke, R. J. (1996). Components of social support,

buffering effects and burnout: Implications for psychological functioning.

Anxiety, Stress, and Coping, 9, 185-197.

Griffith, J. (1985). Social support providers: Who are they? Where are they met? and

the relationship of network characteristics to psychological distress. Basic and

Applied Psychology, 6,(1), 41-60.

Gross, N., Mason, W. S., & McEachern, A. W. (1958). Exploration in role analysis. New

York: Wiley.

Grout, J. W., Steffen, S. M., & Bailey, J. T. (1981). The stresses and the satisfiers of

the intensive care unit: A survey. Critical Care Quarterly, 3, 35-45.

Gulick, E. E. (1995). Coping among spouses or significant others of persons with

multiple sclerosis. Nursing Research, 44, 220-226.

Gupchup, G. V., & Wolfgang, A. P. (1994). The Health Professions Stress Inventory:

Factor structures for pharmacists. Perceptual and Motor Skills, 79, 515-519.

Guppy, A., & Gutteridge, T. (1991). Job satisfaction and occupational stress in UK

general hospital nursing staff. Work and Stress, 5, 315-323.

Gusy, B. (1995). Job stress, social support, and burnout. A causal analysis. Munich:

Profil.
 347

Hackman, J. R. (1986). The psychology of self-management in organizations. In M. S.

Pallak & R. Perloff (Eds.), Psychology and work: Productivity, change, and

employment (pp. 123-161). Washington, DC: American Psychological

Association.

Hague, C. (1987). Caring can damage your health. Intensive Care Nursing, 3, 38-33.

Hall, A., & Wellman, B. (1985). Social networks and social support. In S.Cohen & S. L.

Syme (Eds.), Social Support and health (pp. 23-41). New York: Academic Press.

Handy, J. (1990). Occupational stress in a caring profession. London: Gower.

Happell, B., Pinikahana, J., & Martin, T. (2003). Stress and burnout in forensic

psychiatric nursing. Stress and Health, 19, 63-68.

Hardy, M. E., & Hardy, W. L. (1988). Role stress and role strain. In M. E. Hardy & M.

E. Conway (Eds.), Role theory: Perspectives for health professionals (pp. 159-

239). Norwalk: Appelton Thayer.

Hare, J., Pratt, C. C., & Andrews, D. (1988). Predictors of burnout in professional and

paraprofessional nurses working in hospitals and nursing homes. International

Journal of Nursing Studies, 25, 105-115.

Hartrick, G., & Hills, A. (1993). Staff nurse perceptions of stressors and support needs in

the workplace. The Canadian Journal of Nursing Research, 25, 23-30.

Hatcher, S, & Laschinger, H. (1996). Staff nurses’ perceptions of job empowerment and

level of burnout: A test of Kanter’s theory of structural power in organizations.

Canadian Journal of Nursing Administration, 9, 74-94.

 348

Healy, C. M., & McKay, M. F. (2000). Nursing stress: The effects of coping strategies

and job satisfaction in a sample of Australian nurses. Journal of Advanced

Nursing, 31(3), 681-688.

Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.

Heller, K. (1979). The effects of social support: Prevention and treatment implications. In

A. P. Goldstein & F. H. Kanfer (Eds.), Maximising treatment gains: Transfer

enhancement in psychotherapy. New York: Academic Press.

Heller, K., & Swindle, R. W. (1983). Social networks, perceived social support, and

coping with stress. In R. D. Felner, L. A. Jason, J. N. Moritsugu, & S. S. Farber

(Eds.), Preventative psychology: Theory, research and practice (pp. 87-103).

New York: Pergamon Press.

Helps, S. (1996). Experiences of stress in accident and emergency nurses. Accident and

Emergency Nursing, 5, 48-53.

Hendel, T., Fish, M., & Aboudi, S. (2000). Strategies used by hospital nurses to cope

with a national crisis: A manager’s perspective. International Council of Nurses,

47, 224-231.

Henderson, S., Duncan-Jones, P., Byrne, D. G., & Scott, R. (1980). Measuring social

relationships. The Interview Schedule for Social Interaction. Psychological

Medicine, 10, 723-734.

Hillhouse, J. J., & Adler, C. (1996). Evaluating a simple model of work, stress, burnout,

affective, and physical symptoms in hospital nurses. Psychology, Health, and

Medicine, 1, 297-306.
 349

Hillhouse, J. J., & Adler, C. (1997). Investigating stress effect patterns in hospital staff

nurses: Results of a cluster analysis. Social Science Medicine, 45 (12), 1781-1788.

Himle, D., Jayaratne, S., & Thyness, P. (1991). Buffering effects of four social support

types on burnout among social workers. Social Work Research and Abstracts,

27, 22-27.

Hinds, P. S., Quargnenti, A. G., Hickey, S. S., & Magnum, H. (1994). A comparison of

the stress-response sequence in new and experienced pediatric oncology nurses.

Cancer Nursing, 17, 61-71.

Hingley, P., & Cooper, C. L. (1986). Stress and the nurse manager. Chichester: Wiley

& Sons.

Hipwell, A. E., Tyler, P. A., & Wilson, C. M. (1989). Sources of stress and

dissatisfaction among nurses in four hospital environments. British Journal of

Medical Psychology, 62, 71-79.

Hobfell, S. E., & Freedy, J. (1993). Conservation of resources: A general stress theory

applied to burnout. In W. B. Schaufeli, C. Maslach, & T. Marek (Eds.),

Professional burnout: Recent developments in theory and research (pp. 115-129).

Washington, DC: Taylor & Francis.

Holohan, C. J., Moos, R. H., & Bonin, L. (1997). Social support, coping, and

psychological adjustment: A resources model. In G. R. Pierce, B. Lakey, I. G.

Sarason, & B. R. Sarason (Eds.), Sourcebook of social support and personality

(pp. 169-186). New York: Plenum Press.

Holgate, A. M., & Clegg, I. J. (1991). The path to probation officer burnout: New dogs,

old tricks. Journal of Criminal Justice, 19, 325-337.

 350

Holloway, D., & Wallinga, C. R. (1990). Burnout in child life specialists: The relation

of role stress. Children’s Health Care, 19, 10-18.

House, J. S. (1981). Work stress and social support. Reading, MA: Addison-Wesley.

House, J. S., Kahn, R. L., McLeod, J. D., & Williams, D. (1985). Measures and concepts

of social support. In S. Cohen & S. L. Syme (Eds.), Social support and health (pp.

83-108). Orlando, FL: Academic Press.

House, J. S., Robbins, C., & Metzner, H. L. (1982). The association of social

relationships and activities with mortality: Prospective evidence from the

Tecumseh Community Health Study. American Journal of Epidemiology, 116,

123-140.

House, J. S., & Wells, J. A. (1978). Occupational stress, social support and health. In A.

McLean, G. Black, & M. Colligan (Eds.), Reducing occupational stress:

Proceedings of a conference. U.S. Department of Health, Education and Welfare,

HEW (NIOSH) Publication No. 78-140.

Hutchinson, C. (1999). Social support: Factors to consider when designing studies that

measure social support. Journal of Advanced Nursing, 26, 1520-1526.

Iacovides, A., Fountoulakis, K., Moysidou, C., & Ierodiakonou, C. (1997). Burnout in

nursing staff: A clinical syndrome rather than a psychological reaction? General

Hospital Psychiatry, 19, 419-428.

Iliffe, J. (2000). Campaigning for quality health care. [Electronic Version]. Australian

Nursing Journal, May, 1-2.

Iwanicki, E. F., & Schwab, R. L. (1981). A cross validation study of the MBI.

Educational and Psychological Measurement, 41, 1167-1174.

 351

Iwata, N., & Suzuki, K. (1997). Role stress - Mental health relations in Japanese bank

workers: A moderating effect of social support. International Association of

Applied Psychology, 81, 210-217.

Jaccard, J., Turrisi, R., & Wan, C. K (1990). Interaction effects in multiple regression.

Newbury Park, CA: Sage.

Jackson, S., & Schuler, R. (1985). A meta-analysis and conceptual critique of research

on role ambiguity and role conflict in work settings. Organizational

Behavior and Human Decision Processes, 36, 16-78.

Jackson, S. E., Schwab, R. L., & Schuler, R. S. (1986). Toward an understanding of

the burnout phenomenon. Journal of Applied Psychology, 71, 630-640.

Jackson, S. E., Turner, J., & Brief, A. P. (1987). Correlates of burnout among public

service lawyers. Journal of Occupational Behaviour, 8, 339-349.

Jacobson, S. (1978). Stressful situations for neonatal intensive care nurses. MCN: The

American Journal of Maternal Child Nursing, 3, 144-150.

Janis, I. L. (1965). The problem of validating content analysis. In H. D. Lasswell & N.

Leites (Eds.), Language of politics. Cambridge: MIT Press.

Janssen, P. P. M., De Jonge, J., & Bakker, A. B. (1999). Specific determinants of intrinsic

work motivation, burnout and turnover intentions: A study among nurses. Journal

of Advanced Nursing, 29(6), 1360-1369.

Janssen, P. P. M., Schaufeli, W. B., & Houkes, I. (1999). Work-related and individual

determinants of three burnout dimensions. Work and Stress, 13(1), 74-86.

 352

Jex, S. M., & Beehr, T. A. (1991). Emerging theoretical and methodological issues in the

study of work-related stress. Research in Personnel and Human Resources

Management, 9, 311-365.

Johnson, D., & Preston, B. (2001). An overview of issues in nursing education. (DETYA

No.6761HERCO1A). Retrieved May 22, 2003, from

http://www.dest.gov.au/highered/eippubs/eip01_12/fullreport.htm

Jones, J. G. (1987). Stress among psychiatric nurses. In R. Payne & J. Firth-Cozens

(Eds.), Stress in health professionals (pp. 189-210). New York: Wiley.

Jones, J. W. (1980). Preliminary manual: The Staff Burnout Scale for Health

Professionals. Park Ridge, IL: London House.

Jones, S. R. (2002). (Re)writing the word: Methodological strategies and issues in

qualitative research. Journal of College Student Development, 43, 461-473.

Kahn, R. L. (1979). Aging and social support. In M. W. Riley (Ed.), Aging from birth to

death: Interdisciplinary perspectives (pp. 77-91). Coulder, CO: Westview Press.

Kahn, R. L., & Antonucci, T. C. (1980). Convoys over the life course: Attachments, roles

and social support. In P. B. Baltes & O. Brim (Eds.), Lifespan development and

behavior (Vol. 3, pp. 253-286). New York: Academic Press.

Kahn, R. L., & Byosiere, P. (1992). Stress in Organisations. In M. D. Dunnette & L.

M. Hough (Eds.), Handbook of industrial and organizational psychology (pp.

571-650). California: Consulting Psychologists Press.

Kahn, R. L., Wethington, E., & Ingersoll-Dayton, B. N. (1987). Social support and social

networks: Determinants, effects, and interactions. In P. R. Abeles (Ed.), Life-span

perspectives and social psychology (pp. 139-165). New York: Academic Press.
 353

Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D. & Rosenthal, R. A. (1964).

Organizational stress: Studies in role conflict and ambiguity. New York: Wiley.

Kammer, C. H. (1994). Stress and coping of family members responsible for nursing

home placement. Research in Nursing Health, 17(2), 89-98.

Kandolin, I. (1993). Burnout of female and male nurses in shiftwork. Ergonomics, 36,

141-147.

Kanner, A. D., Coyne, J. C., Schaefer, C., & Lazarus, R. S. (1981). Comparison of two

models of stress measurement: Daily hassles and uplifts versus major life events.

Journal of Behavioral Medicine, 4, 4-39.

Kaufmann, G. R., & Beehr, T. A. (1986). Interactions between job stressors and social

support. Journal of Applied Psychology, 71, 522-526.

Kaufmann, G. R., & Beehr, T. A. (1989). Occupational stressors, individual strains, and

social supports among police officers. Human Relations, 42, 185-197

Keane, A., Ducette, J., & Adler, D. C. (1985). Stress in ICU and non-ICU nurses.

Nursing Research, 34, 231-236.

Kennedy, B. R. (1999). Stress and burnout of nursing staff with geriatric patients in long-

term care. Dissertation Abstracts International: Section B: The Sciences and

Engineering, 59(7-B), 3347.

Kennedy, P., & Grey, N. (1997). High pressure areas. Nursing Times, 93(29), 26-27.

Kessler, R. C., & McLeod, J. D. (1985). Social support and mental health in community

samples. In S. Cohen & S. L. Syme (Eds.), Social support and health (pp. 219-

240). Florida: Academic Press.

 354

Ketola, O., & Kevätsalo, K. (1994). A profession or a calling? Social welfare and health

care work and organizations in transition. Helinski: Painatuskeskus.

Kilfedder, C. J., Power, K. G., & Wells, T. J. (2001). Burnout in psychiatric nursing.

Journal of Advanced Nursing, 34(3), 383-396.

King, L. A., & Beehr, T. A. (1983). Therapist burnout: Reliability and validity of the

Maslach Burnout Inventory. Paper presented at the annual meeting of the

Midwestern Psychological Association, Chicago.

King, L. A., Mattimore, L. K., King, D. W., & Adams, G. A. (1995). Family Support

Inventory for Workers: A new measure of perceived social support from family

members. Journal of Organizational Behavior, 16, 235 - 258.

Kirkaldy, B. D., & Martin, T. (2000). Job stress and satisfaction among nurses: Individual

differences. Stress Medicine, 16, 77-89.

Kitzinger, J. (1994). The methodology of focus groups: The importance of interactions

between research participants. Sociology of Health and Illness, 16, 103-121.

Kitzinger, J. (1996). Introducing focus groups in qualitative research. In N. Mays & C.

Pope (Eds.), Health care (pp. 36-45). London: BMJ Publishing Group.

Koeske, G. F., & Kelly, T. (1995). The impact of overinvolvement on burnout and job

satisfaction. American Journal of Orthopsychiatry, 65(2), 282-292.

Koeske, G. F., & Koeske, R.D. (1989). Construct validity of the Maslach Burnout

Inventory: A critical review and reconceptualization. Journal of Applied

Behavioral Science, 25, 131-132.

 355

Koeske, G. F., & Koeske, R. D. (1993). A preliminary test of a stress-strain-outcome

model for reconceptualizing the burnout phenomenon. Journal of Social Service

Research, 17, 107-135.

Koivula, M., Paunonen, M., & Laippala, P. (2000). Burnout among nursing staff in two

Finnish hospitals. Journal of Nursing Management, 8, 149-158.

Kong, S. X., & Wertheimer, A. I. (1994). Social support: Concepts, theories, and

implications for pharmacy research. Journal of Pharmaceutical Marketing and

Management, 9(1), 63-90.

Kong, S. X., Wertheimer, A. I., & Myers, M. J. (1994). Effect of social support on role

stress and organizational commitment: Test of three theoretical models. Journal

of Pharmaceutical Marketing and Management, 9, 93-116.

Koniarek, J., & Dudek, B. (1996). Social support: A buffer in the stress-burnout

relationship. International Journal of Stress Management, 3(2), 99-106.

Kottkamp, R. B., & Mansfield, J. R. (1985). Role conflict, role ambiguity, powerlessness

and burnout among high school supervisors. Journal of Research and

Development in Education, 18, 29-38.

Kottkamp, R. B., & Travlos, A. L. (1986). Selected job stressors, emotional

exhaustion, job satisfaction, and thrust behaviour of the high school

principal. Alberta Journal of Educational Research, 32, 234-248.

Kreuger, R. A. (1993). Quality control in focus group research. In D. L. Morgan (Ed.),

Successful focus groups: Advancing the state of the art (pp. 65-85). Newbury

Park, CA: Sage.

 356

Krippendorf, K. (1980). Content analysis: An introduction to its methodology. Beverly

Hills, CA: Sage.

Lakey, B., & Cohen, S. (2000). Social support theory and measurement. In S. Cohen, L.

G. Underwood, & B. H. Gottlieb (Eds.), Social support measurement and

intervention (pp.29-51). Oxford: Oxford University Press.

Lambert, V. A., & Lambert, C. E. (1987). Coping with Rheumatoid Arthritis. Nursing

Clinics of North America, 22 (3), 551-558.

Lambert, V. A., & Lambert, C. E. (2001). Literature review of role stress/strain on

nurses: An international perspective. Nursing and Health Sciences, 3, 161-172.

Landau, K. (1992). Psycho-physical strain and the burn-out phenomenon amongst health

care professionals. In M. Estryn-Behar, C. Gadbois, & M. Pottier (Eds.),

Ergonomie al’ hospital (Hospital Ergonomics), International Symposium Paris

1991, Editions Octares: Toulouse.

Landbergis, P. A. (1988). Occupational stress among nurses: New developments in

theory and prevention. In J. H. Humphrey (Ed.), Human stress: Current selected

research (Vol. 3, pp. 173-195). New York: AMS Press.

Langner, T. (1962). A twenty-two item screening score of psychiatric symptoms

indicating impairment. Journal of Health and Social Behavior, 3, 269-276.

LaRocco, J. M., House, J. S., & French, J. R. P., Jr. (1980). Social support,

occupational stress, and health. Journal of Health and Social Behavior, 21,

202-218.

Larson, D. (1987). Interval stressors in nursing: Helper secrets. Journal of Psychosocial

Nursing, 3, 161-172.
 357

Laschinger, H. K. S., Finegan, J., Shamian, J., & Almost, J. (2001). Testing Karasek’s

demand-control model in restructured healthcare settings: Effects of job strain on

staff nurses’ quality of work life. Journal of Nursing Administration, 31(5), 233-

243.

Laschinger, H. K. S., Finegan, J., Shamian, J., & Wilk, P. (2001). Impact of structural

and psychological settings: Expanding Kanter’s model. Journal of Nursing

Administration, 31, 260-272.

Lauver, D., & Tak, Y. (1995). Optimism and coping with breast cancer symptoms.

Nursing Research, 44, 202-207.

Lazarus, R. S. (1966). Psychological stress and the coping process. New York: McGraw-

Hill.

Lazarus, R. S. (1993). From psychological stress to the emotions: A history of changing

outlooks. Annual Review of Psychology, 44, 1-21.

Lazarus, R. S. (1995). Psychological stress in the workplace. In P. L. Perrewe & R.

Crandall (Eds.), Occupational stress: A handbook (pp. 3-14). Washington, DC:

Taylor & Francis.

Lazarus, R. S., & Cohen, J. B. (1977). Environmental stress. In I. Altman & J. F.

Wohlwill (Eds.), Human behavior and the environment: Current theory and

research (pp.89-127). New York: Spectrum.

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York:

Springer.

Leavy, R. L. (1983). Social support and psychological disorder: A review. Journal of

Community Psychology, 11, 3-21.

 358

LeCroy, C. W., & Rank, M. R. (1986). Factors associated with burnout in the social

services: An exploratory study. [Special Issue]. Journal of Social Service

Research, 10, 23-39.

Lee, J. K. L. (2003). Job stress, coping and health perceptions of Hong Kong primary

care nurses. International Journal of Nursing Practice, 9, 86-91.

Lee, R. T., & Ashforth, B. E. (1990). On the meaning of Maslach’s three dimensions of

burnout. Journal of Applied Psychology, 75, 743-747.

Lee, R. T., & Ashforth, B. E. (1993). A further examination of managerial burnout:

Toward an integrated model. Journal of Organizational Behavior, 14, 3-20.

Lee, R. T., & Ashforth, B. E. (1996). A meta-analytic examination of the correlates of the

three dimensions of job burnout. Journal of Applied Psychology, 81(2), 123-133.

Lefcourt, H. M., Martin, R. A., & Saleh, W. E. (1994). Locus of control and social

support: Interactive moderators of stress. Journal of Personality and Social

Psychology, 47, 378-389.

Leiter, M. P. (1990). The impact of family resources, control coping, and skill utilization

on the development of burnout: A longitudinal study. Human Relations, 43, 1067-

1083.

Leiter, M. P. (1991). Coping patterns as predictors of burnout: The function of control

and escapist coping patterns. Journal of Organizational Behavior, 12, 123-144.

Leiter, M. P. (1993). Burnout as a developmental process. Consideration of models. In

W. B. Schaufeli, C. Maslach, & T. Marek (Eds.), Professional burnout: Recent

developments in theory and research (pp. 1-16). Washington, DC: Taylor &

Francis.
 359

Leiter, M. P., & Harvie, P. L. (1996). Burnout among mental health workers: A review

and research agenda. International Journal of Social Psychiatry, 42(2), 90-101.

Leiter, M. P., & Maslach, C. (1988). The impact of interpersonal environment on burnout

and organizational commitment. Journal of Organizational Behavior, 9, 297-308.

Lewinson, N. J., Conley, J., & Blessing-Moore, J. (1981). Measurement of hypothetical

burnout in cystic fibrosis caregivers. Acta Pediatrica Scandinavica, 70, 935-939.

Lin, N., & Ensel, W. M. (1984). Depression-mobility and its social etiology: The role of

life events and social support. Journal of Health and Social Behavior, 25, 176-

188.

Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Newbury Park, CA: Sage.

Linder-Pelz, S., Pierce, J. P., & Minslow, M. (1987). Nurses’ work stressors in an

Australian metropolitan teaching hospital. Journal of Occupational Health Safety

– Australia and New Zealand, 3, 624-630.

Linquist, C. A., & Whitehead, J. T. (1986). Correctional officers as parole officers: An

examination of community supervision sanction. Criminal Justice Behaviour,

13, 197-122.

Lumby, J. (1996). Stressors in the life of a contemporary nurse. International Journal of

Nursing Practice, 2, 45-47.

Lyons, T. F. (1971). Role clarity, need for clarity, satisfaction, tension, and withdrawal.

Organizational Behavior and Human Performance, 6, 99-110.

MacNeil, J., & Weisz, G. (1987). Critical care nursing stress: Another look. Heart Lung,

16, 274-277.
 360

Mallet, K. L., Price, J. H., Jurs, S. G., & Slenker, S. (1991). Relationships among

burnout, death anxiety, and social support in hospice and critical care nurses.

Psychological Reports, 68, 1347-1359.

Manlove, E. E. (1994). Conflict and ambiguity over work roles: The impact on child

care worker burnout. Early Education and Development, 5, 41-55.

Martocchio, J. J., & O’Leary, A. M. (1989). Sex differences in occupational stress: A

meta-analytic review. Journal of Applied Psychology, 74, 495-501.

Maslach, C. (1976). Burned-out. Human Behavior, 5, 16-22.

Maslach, C. (1982). Burnout: The cost of caring. Englewood Cliffs, NJ: Prentice-Hall.

Maslach, C. (1993). Burnout: A multidimensional perspective. In W. B. Schaufeli, C.

Maslach, & T. Marek (Eds.), Professional burnout: Recent developments in

theory and research (pp. 19-32). Washington, DC: Taylor & Francis.

Maslach, C. (1998). A multidimensional theory of burnout. In C. L. Cooper (Ed.),

Theories of organizational stress (pp. 68-85). New York: Oxford University

Press.

Maslach, C., & Jackson, S. E. (1981). The measurement of experienced burnout.

Journal of Occupational Behaviour, 2, 99-113.

Maslach, C., & Jackson, S. E. (1982). Burnout in health professionals: A social

psychological analysis. In G. Sanders & J. Suls (Eds.), Social psychology of

health and illness. Hillsdale, NJ: Erlbaum.

Maslach, C., & Jackson, S. E. (1985). The role of sex and family variables in burnout.

Sex Roles, 12, 837-851.

 361

Maslach, C., & Jackson, S. E. (1986). Maslach Burnout Inventory Manual (2nd ed.). Palo

Alto, CA: Consulting Psychologists Press.

Maslach, C., & Jackson, S. E ., & Leiter, M. (1996). Maslach Burnout Inventory Manual

(3rd ed.). Pala Alto, CA: Consulting Psychologists Press.

Maslach, C., & Leiter, M. P. (1999). Take this job…and love it! Psychology Today, 32,

50-53.

Maslach, C., & Pines, A. (1977). The burnout syndrome in the dare care setting. Child

Care Quarterly, 6, 100-113.

Maslach, C., & Schaufeli, W. B. (1993). Historical and conceptual development of

burnout. In W. B. Schaufeli, C. Maslach, & T. Marek (Eds.), Professional

burnout: Recent developments in theory and research (pp.1-16). Washington, DC:

Taylor & Francis.

Maslanka, H. (1996). Burnout, social support and AIDS volunteers. AIDS Care, 8, 195-

206.

Mattimore, L. K. (1990). Family Support Inventory: A measure of perceived family social

support for workers. Unpublished master’s thesis, Central Michigan University,

Michigan.

Matuszek, P. A. C., Nelson, D. L., & Quick, J. C. (1995). Gender differences in distress:

Are we asking all the right questions? Journal of Social Behavior and

Personality, 10, 99-120.

Mayo, P. R. (1983). Personality traits and the retrieval of positive and negative memories.

Personality and Individual Differences, 4, 465-471.

 362

McCranie, E. W., Lambert, V. A., & Lambert, C. E. (1987). Work stress, hardiness, and

burnout among hospital staff nurses. Nursing Research, 36(6), 374-378.

McGrath, A., Reid, N., & Boore, J. (1989). Occupational stress in nursing. International

Journal of Nursing Studies, 26(4), 343-358.

McGuire, W. J. (1966). The current status of cognitive consistency theories. In S.

Feldman (Ed.), Cognitive consistency (pp.1-46). New York: Academic Press.

McNally, S. T., & Newman, S. (1999). Objective and subjective conceptualizations of

social support. Journal of Psychosomatic Research, 46(4), 309-314.

Miles, M. B., & Huberman, M. A. (1984). Qualitative data analysis: A source book of

new methods. London: Sage.

Miller, K. I., Ellis, B. H., Zook, E. G., & Lyles, J. S. (1990). An integrated model of

communication, stress, and burnout in the workplace. Communication Research,

17(3), 300-326.

Miller, L., Reesor, K., McCarrey, M., & Leiken, L. (1995). Nursing burnout. Employee

Assistance Quarterly, 10(4), 29-52.

Molleman, E., Pruyn, J., & Van Knippenberg, A. (1986). Social comparison processes

among cancer patients. British Journal of Social Psychology, 25, 1-13.

Monat, A., & Lazarus, R. S. (1977). Stress and coping: An anthology. New York:

Colombia University Press.

Mugo, F. W. (2003). Sampling in research. Retrieved October 1, 2003, from

http://trochim.human.cornell.edu./tutorial/mugo/tutorial.htm

Munn, E. K., Barber, C. E., & Fritz, J. J. (1996). Factors affecting the professional well-

being of child life specialists. Children’s Health Care, 25, 71-91.

 363

Munro, L., Rodwell, J., & Harding, L. (1998). Assessing occupational stress in

psychiatric nurses using the full job strain model: The value of social support to

nurses. International Journal of Nursing Studies, 35, 339-345.

Murray, T. (1998). From outside the walls: A qualitative study of nurses who recently

changed from hospital-based practice to home health care nursing. Journal of

Continuing Education Nursing, 29, 55-60.

Naisberg-Fennig, S., Fennig, S., Keinan, G., & Elizur, A. (1991). Personality

characteristics and proneness to burnout: A study among psychiatrists. Stress

Medicine, 7(4), 201-205.

Nandram, S. S., & Klandermans, B. (1993). Stress experienced by active members of

trade unions. Journal of Organizational Behavior, 14, 415-431.

Nätterlund, B., & Ahlström, G. (1999). Experience of social support in rehabilitation: A

phenomological study. Journal of Advanced Nursing, 30(6), 1332-1340.

Neuman, W. L. (2000). Social research methods (4th ed.). Sydney: Allyn and Bacon.

Norbeck, J. S. (1984). The Norbeck Social Support Questionnaire. Birth Defects:

Original Article Series, 20(5), 45-55.

Norbeck, J. S. (1988). Social support. Annual Review of Nursing Research, 6, 85-109.

Norbeck, J. S., Lindsey, A. M., & Carrieri, V. L. (1981). The development of an

instrument to measure social support. Nursing Research, 30, 264-269.

Norbeck, J. S., Lindsey, A. M., & Carrieri, V. L. (1983). Further development of the

Norbeck Social Support Questionnaire: Normative data and validity testing.

Nursing Research, 32, 4-9.

 364

Numerof, R. E., & Abrams, M. N. A. (1984). Sources of stress among nurses: An

empirical investigation. Journal of Human Stress, 10(2), 88-100.

O’Driscoll, M. P., & Beehr, T. A. (1994). Supervisor behaviors, role stressors and

uncertainty as predictors of personal outcomes for subordinates. Journal of

Organizational Behavior, 15, 141-155.

O’Driscoll, M. P., & Cooper, C. L. (1996). Sources and management of excessive job

stress and burnout. In P. Warr (Ed.), Psychology at work (4th ed., pp. 188-233).

London: Penguin.

Ogus, E. D. (1990). Burnout and social support among ward nurses. Issues in Mental

Health Nursing, 11, 267-281.

Ogus, E. D. (1992). Burnout and coping strategies: A comparative study of ward nurses.

Journal of Social Behavior and Personality, 7, 11-124.

Ogus, E. D., Greenglass, E. R., & Burke, R. J. (1990). Gender role differences, work

stress and depersonalisation. Journal of Social Behavior and Personality, 5, 387-

398.

Osgood, C. E., & Tannenbaum, P. H. (1955). The principle of congruity in the prediction

of attitude change. Psychological Review, 62, 42-55.

Oskins, S. L. (1979). Identification of situational stressors and coping methods by

intensive care nurses. Heart Lung, 8(2), 953-960.

Osipow, S. H., & Spokane, A. R. (1981). Measures of Occupational Stress, Strain, and

Coping (Form-E1). Columbus, OH: Marathon Consulting and Press.

Osipow, S. H., & Spokane, A. R. (1983). Manual for Measures of Occupational Stress,

Strain, and Coping (Form E-2). Columbus, OH: Marathon Consulting and Press.
 365

Osipow, S. H., & Spokane, A. R. (1987). Manual for the Occupational Stress Inventory.

Odessa, FL: Psychological Assessment Resources.

Pagel, I. S., & Wittmann, M. E. (1986). Relationship of burnout to personal and job-

related variables in acute-care paediatric settings. Issues in Comprehensive

Paediatric Nursing, 9, 131-143.

Parasuraman, S., Greenhaus, J. H., & Granose, C. S. (1992). Role stressors, social

support, and well-being among two career couples. Journal of Organizational

Behavior, 13, 339-356.

Parker, L. E., & Price, R. H. (1994). Empowered managers and empowered workers: The

effects of managerial support and managerial perceived control on worker’s sense

of control over decision making. Human Relations, 47(8), 911-928.

Parker, P. A., & Kulik, J. A. (1995). Burnout, self- and supervisor-rated job

performance, and absenteeism among nurses. Journal of Behavioral Medicine,

18, 581-599.

Patterson, G. T. (2003). Examining the effects of coping and social support on work and

life stress among police officers. Journal of Criminal Justice, 31, 215-226.

Patton, M. Q. (1990). Qualitative evaluation and research methods. Newbury Park, CA:

Sage.

Patton, W., & Goddard, R. (2003). Psychological distress and burnout in Australian

employment service workers: Two years on. Journal of Employment Counseling,

40, 2-16.

Payne, N. (2001). Occupational stressors and coping as determinants of burnout in female

hospice nurses. Journal of Advanced Nursing, 33(3), 396-405.

 366

Payne, R. L., & Jones, J. G. (1987). Measurement and methodological issues in social

support. In S. V. Kasl & C. L. Cooper (Eds.), Stress and health: Issues in

research methodology (pp. 167-205). New York: John Wiley & Sons.

Pearce, J. (1981). Bringing some clarity to role ambiguity research. Academy of

Management Review, 6, 665-674.

Peeters, M., C. W., & Le Blanc, P. M. (2001). Towards a match between job demands

and sources of social support: A study among oncology care providers. European

Journal of Work and Organizational Psychology, 10(1), 53-72.

Peiro, J. M., & Zurriaga, R. (1985, September). Role ambiguity and occupational burnout

stress in hospitals in Spain. Paper presented at the workshop on work and

organizational problems in hospitals, E. N. O. P., University of Wales, Wales.

Perlman, B., & Hartman, E. A. (1982). Burnout: Summary and future research. Human

Relations, 35, 283-305.

Pettigrew, A., & Whipp, R. (1991). Managing change for competitive success. Oxford:

Blackwell.

Pick, D., & Leiter, M. (1991). Nurses’ perceptions of the nature and causes of burnout: A

comparison of self reports and standardised measures. Canadian Journal of

Nursing Research, 23, 33-48.

Pierce, C. M., & Molloy, G. N. (1989). The construct validity of the Maslach Burnout

Inventory: Some data from down under. Psychological Reports, 65, 1340-1342.
 367

Pierce, G. R., Sarason, B. R., & Sarason, I. G. (1990). Integrating social support

perspectives: Working models, personal relationships, and situational factors. In

S. Duck (Ed.), Personal relationships and social support (pp.173-189). London:

Sage.

Pikó, B. (1999). Work-related stress among nurses: A challenge for health care

institutions. The Journal of Royal Society for the Promotion of Health, 119(3),

156-162.

Pines, A. M. (1993). Burnout: An existential perspective. In W. B. Schaufeli, C.

Maslach, & T. Marek (Eds.), Professional burnout: Recent developments in

theory and research (pp. 33-51). Washington, DC: Taylor & Francis.

Pines, A. M., Aronson, E. (1988). Career burnout: Causes and cures. New York: Free

Press.

Pines, A. M., Aronson, E., & Kafrey, D. (1981). Burnout: From tedium to personal

growth. New York: Free Press.

Pines, A., & Maslach, C. (1978). Characteristics of staff burnout in mental health

settings. Hospital and Community Psychiatry, 29, 233-237.

Pines, A., & Maslach, C. (1980). Combating staff burn-out in a day care center: A case

study. Child Care Quarterly, 9(1), 5-16.

Plante, A., & Bouchard, L. (1995). Occupational stress, burnout and professional support

in nurses working with dying patients. Journal of Death and Dying, 32, 93-109.

Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2003). Common method biases

in behavioral research: A critical review of the literature and recommended

remedies. Journal of Applied Psychology, 88(5), 879-903.

 368

Power, G., & Sharp, G. R. (1988). A comparison of sources of nursing stress

and job satisfaction among mental handicap and hospice staff. Journal of

Advanced Nursing, 13, 726-732.

Pruessner, J. C., Hellhammer, D. K., & Kirschbaum, C. (1999). Burnout, perceived

stress, and cortisol responses to awakening. Psychosomatic Medicine, 61, 197-

205.

Queensland Health (2001). Workforce characteristics. Retrieved October 23, 2003, from

http://www.health.gov.au/hic/nurses.pdf

Queensland Nurses’ Union (2003, May). Nurses: Worth looking after campaign [Press

Release]. Retrieved from http://www.qnu.org.au/news_and_notices/press_releases

/may_2003.

Quick, J. C., Quick, J. D., Nelson, D. L., & Hurrell, J. J., Jr. (1997). Preventative stress

management in organizations. WA: American Psychological Association.

Raquepaw, J. M., & Miller, R. S. (1989). Psychotherapist burnout: A componential

analysis. Professional Psychology: Research and Practice, 20, 32-36.

Ray, E. B. (1987). Support relationships and occupational stress in the workplace. In T.

L. Albrecht, & M. B. Adelman (Eds.), Communicating social support (pp.179-

191). Newbury Park, CA: Sage.

Ray, E. B., & Miller, K. I. (1994). Social support, home/work stress, and burnout: Who

can help? Journal of Applied Psychology, 30(3), 357-373.

Rees, D., & Cooper, C. (1992). The occupational stress indicator locus of control scale:

Should this be regarded as a state rather than a trait measure? Work and Stress, 6,

45-48.
 369

Reich, J. W., & Zautra, A. (1981). Life events and personal causation: Some relationships

with satisfaction and distress. Journal of Personality and Social Psychology, 41,

1002-1112.

Revicki, D. A., & May, H. J. (1989). Organizational characteristics, occupational stress,

and mental health in nurses. Behavioral Medicine, 15(1), 30-36.

Richards, D. (2000, May). Dealing with increasing workloads and overtime. Australian

Nursing Federation Industrial News. Retrieved February 6, 2003, from

http://www.anf.org.au/news_industrial/news_industrial_0005.html

Riding, R. J., & Wheeler, H. H. (1995a). Occupational stress and cognitive style in

nurses. British Journal of Nursing, 4(2), 103-107.

Riding, R. J., & Wheeler, H. H. (1995b). Occupational stress and cognitive style in

nurses. British Journal of Nursing, 4(3), 160-168.

Robinson, N. (1999). The use of focus group methodology – with selected examples from

sexual health research. Journal of Advanced Nursing, 29(4), 905-913.

Robinson, S. E., Roth, S. L., Keim, J., Levenson, M., Flenje, J. R., & Bashor, K. (1991).

Nurse burnout: Work related and demographic factors as culprits. Research in

Nursing and Health, 14, 223-228.

Rook, K. S. (1984). The negative side of social interactions: Impact on psychological

well-being. Journal of Personality and Social Psychology, 46, 1097-1108.

Rook, K. S., & Dooley, D. (1985). Applying social support research: Theoretical

problems and future directions. Journal of Social Issues, 41, 5-28.

Ross, R. R., & Altmaier, E. M. (1994). Intervention in occupational stress: A handbook

of counselling for stress at work. California: Sage.

 370

Roxburgh, S. (1999). Exploring the work and family relationship: Gender differences in

the influence of parenthood and social support on job satisfaction. Journal of

Family Issues, 20, 771-788.

Runciman, P. J. (1983). Ward sister at work. Edinburgh: Livingstone.

Russell, D., & Cutrona, C. E. (1984, August). The provisions of social relationships and

adaptation to stress. Paper presented at the annual meeting of the American

Psychological Association, Toronto.

Saint-Arnaud, L., Gingras, S., Boulard, R., Vezina, M., & Lee-Gosselin, H. (1992). Les

symptoms psychologiques en milieu hospitalier (Psychological symptoms in

hospitals). In M. Estryn-Behar, C. Gadbois, & M. Pottier (Eds.), Ergonomie al’

hospital (Hospital Ergonomics), International Symposium Paris 1991, Editions

Octares: Toulouse.

Sandler, I. N., & Barrera, M., Jr. (1984). Toward a multimethod approach to assessing the

effects of social support. American Journal of Community Psychology, 12, 37-52.

Sandler, I. N., & Lakey, B. (1982). Locus of control as a stress moderator: The role of

control perceptions and social support. American Journal of Community

Psychology, 10, 65-78.

Sandler, I. N., Miller, P., Short, J., & Wolchick, S. A. (1989). Social support as a

protective factor for children in stress. In D. Belle (Ed.). Children’s social

networks and social supports (pp. 277-307). Oxford: John Wiley & Sons.

Sarason, B. R., Sarason, I. G., & Pierce, G. R. (1990). Social support: An interactional

view. New York: Wiley.

 371

Sarason, I. G., Sarason, B. R., & Pierce, G. R. (1992). Three contexts of social support. In

H. O. F. Veiel, & V. Bauman (Eds.), The meaning and measurement of social

support (pp. 143-155). Washington D.C.: Hemisphere Publishing.

Sarason, I. G., Sarason, B. R., & Pierce, G. R. (1994). Relationship-specific social

support: Toward a model for the analysis of social interactions. In B. R. Burleson,

T. L. Albrecht, & I. G. Sarason (Eds.), Communication of social support:

Messages, interactions, relationships, and community (pp. 91-112). Thousand

Oaks, CA: Sage.

Sarason, B. R., Shearin, E. N., Pierce, G. R., & Sarason, I. G. (1987). Interrelations of

social support measures: Theoretical and practical implications. Journal of

Personality and Social Psychology, 52 (4), 813-832.

Schaefer, J. A., & Moos, R. H. (1993). Work stressors in health care: Context and

outcomes. Journal of Community and Applied Social Psychology, 3, 235-242.

Schaefer, C., Coyne, J. C., & Lazarus, R. S. (1981). The health-related functions of social

support. Journal of Behavioral Medicine, 4(4), 381-406.

Schaufeli, W. B. (1990). Burnout. Rotterdam: Donker.

Schaufeli, W. B. (1999). Burnout. In J. Firth-Cozens & R. L. Payne (Eds.), Stress in

health professionals (pp. 19-32). NY: John Wiley & Sons.

Schaufeli, W. B. & Enzmann, D. (1998). The burnout companion to study and practice:

A critical analysis. London: Taylor & Francis.

Schaufeli, W. B., & Janczur, B. (1994). Burnout among nurses: A Polish-Dutch

comparison. Journal of Cross-Cultural Psychology, 25(1), 95-113.

 372

Schaufeli, W. B., Maslach, C., & Marek, T. (1993). Professional burnout. Recent

developments in theory and research. Washington, D.C.: Taylor & Francis.

Schaufeli, W. B., & Van Dierendonck, D. (1993). The construct validity of two burnout

measures. Journal of Organizational Behavior, 14, 631-647.

Scheck, C. L., Kinicki, A. J., & Davy, J. (1997). Testing the mediating processes between

work stressors and subjective well-being. Journal of Vocational Behavior, 50, 96-

123.

Schmieder, R. A., & Smith, C. S. (1996). Moderating effects of social support in

shiftworking and non-shiftworking nurses. Work and Stress, 10(2), 128-140.

Schmitz, N., Neumann, W., & Oppermann, R. (2000). Stress, burnout and locus of

control in German nurses. International Journal of Nursing Studies, 37, 95-99.

Schwab, R. L., & Iwanicki, E. F. (1982). Who are burned out teachers? Educational

Research Quarterly, 7, 5-16.

Senate Community Affairs References Committee (2002). The patient

profession, time for action report on the inquiry into nursing. Canberra: Senate

Community Affairs References Committee Secretariat.

Seyle, H. (1967). Stress in health and disease. Boston: Butterworth.

Shinn, M., Rosario, M., Morch, H., & Chestnut, D. E (1984). Coping with job stress and

burnout in the human services. Journal of Personality and Social Psychology, 46,

864-876.

Shinn, M., Wong, N. W., Simko, P. A., & Ortiz-Torres, B. (1989). Promoting the well-

being of working parents: Coping, social support, and flexible job schedules.

American Journal of Community Psychology, 17(1), 31-55.

 373

Shirom, A. (1989). Burnout in work organizations. In C. L. Cooper & I. T. Robertson

(Eds.), International Review of Industrial and Organizational Psychology (pp. 25-

48). Chichester: John Wiley & Sons.

Shumaker, S., & Brownell, A. (1984). Toward a theory of social support: Closing

conceptual gaps. Journal of Social Issues, 40(4), 11-36.

Siegrist, J. (1996). Adverse health effects of high effort-low reward conditions. Journal

of Occupational Health Psychology, 1, 27-41.

Silverman, D. (1985). Qualitative methodology and sociology: Describing the social

world. Aldershot: Gower.

Simoni, P. S., & Paterson, J. J. (1997). Hardiness, coping, and burnout in the nursing

workplace. Journal of Professional Nursing, 13, 178-185.

Snape, J., & Cavanagh, S. (1993). Occupational stress in neurosurgical nursing. Intensive

Critical Nursing, 9, 62-70.

Stehle, J. L. (1981). Critical care nursing stress: The findings revisited. Nursing

Research, 30, 182-186.

Stewart, D. W., & Shamdasani, P. M. (1990). Focus groups: Theory and practice.

Newbury: CA: Sage.

Stewart, M. J. (1989). Social support instruments created by nurse investigators. Nursing

Research, 38(5), 268-275.

Stewart, M. J. (1993). Integrating social support in nursing. CA: Sage.

Stokes, J. P. (1983). Predicting satisfaction with social support from social network

structure. American Journal of Community Psychology, 11, 141-152.

 374

Stordeur, S., D’Hoore, W., & Vandenberghe, C. (2001). Leadership, organizational

stress, and emotional exhaustion among hospital nursing staff. Journal of

Advanced Nursing, 35(4), 533-542.

Sullivan, P. (1993). Stress and burnout in psychiatric nursing. Nursing Standard, 8, 36-

39.

Sullivan, S.E., & Bhagat, R. S. (1992). Organizational stress, job satisfaction, and job

performance: Where do we go from here? Journal of Management, 18, 353-374.

Tabachnick, B. G. & Fidell, L. S. (1996). Using multivariate statistics (3rd ed.). New

York: HarperCollins.

Tardy, C. H. (1985). Social support measurement. American Journal of Community

Psychology, 13(2), 187-203.

Taris, T. W., Schreurs, P. J. G., & Schaufeli, W. B. (1999). Construct validity of the

Maslach Burnout Inventory - General Survey: A two sample examination of its

factor structure and correlates. Work and Stress, 13, 223-227.

Taylor, S., White, B., & Muncer, S. (1999). Nurses’ cognitive structural models of work-

based stress. Journal of Advanced Nursing, 29(4), 974-983.

Terry, D. J., Nielsen, M., & Perchard, L. (1993). Effects of work stress on psychological

well-being and job satisfaction: The stress-buffering role of social support.

Australian Journal of Psychology, 45, 168-175.

Tetzloff, C. E., & Barrera, M. (1987). Divorcing mothers and social support: Testing the

specificity of the buffering effect. American Journal of Community Psychology,

15(4), 419-434.
 375

Tierney, W. G. (1996). Leadership and postmodernism: On voice and the qualitative

method. Leadership Quarterly, 7(3), 371-383.

Thoits, P. A. (1982). Conceptual, methodological and theoretical problems in studying

social support as a buffer against life stress. Journal of Health and Social

Behavior, 23, 145-159.

Thoits, P. A. (1985). Social support and psychological well-being: Theoretical

possibilities. In I. G. Sarason & B. R. Sarason (Eds.), Social support: Theory,

research, and applications. Dordrecht, Netherlands: Martinus Nijhoff.

Thoits, P. A. (1995). Stress, coping, and social support processes: Where are we? What

next? Journal of Health and Social Behavior [Extra Issue], 53-79.

Tilden, V. P., Nelson, C. A., & May, B. A. (1990). The IPR Inventory: Development and

psychometric characteristics. Nursing Research, 39(6), 337-343.

Tolsdorf, C. (1976). Social networks, support and coping: An exploratory study. Family

Process, 15, 407-417.

Topf, M. (1989). Personality, hardiness, occupational stress and burnout in critical care

nurses. Research in Nursing and Health, 12, 179-186.

Tracy, E. M. (1990). Identifying social support resources of at-risk families. Social Work,

35, 252-258.

Tummers, G., Janssen, P., Landerweed, A., & Houkes, I. (2001). A comparative study of

work characteristics and reactions between general and mental health nurses: A

multi-sample analysis. Journal of Advanced Nursing, 36(1), 151-163.

 376

Turner, R. J. (1992). Measuring social support: Issues of concept and method. In H. O. F.

Veiel & U. Baumann (Eds.), The meaning and measurement of social support (pp.

217-234). New York: Hemisphere.

Turner, R. J., Frankel, B. G., & Levin, D. M. (1983). Social support: Conceptualization,

measurement, and implications for mental health. In J. R. Greenley (Ed.),

Research in community and mental health (pp. 67-11). Greenwich, CT: JAI.

Turnipseed, D. L., & Turnipseed, P. H. (1997). A bicultural analysis of the cost of caring:

nursing burnout in the United States and the Philippines. Career Development,

2(4), 180-188.

Tyler, P. A., & Cushway, D. (1995). Stress in nurses – The effects of coping and social

support. Stress Medicine, 11, 243-251.

Tyler, P. A., & Ellison, R. N. (1994). Sources of stress and psychological well-being in

high-dependency nursing. Journal of Advanced Nursing, 19, 469-476.

Vandenberghe, R., & Huberman, A. M. (1999). Understanding and preventing teacher

burnout: A sourcebook of international research and practice. New York:

Cambridge University Press.

Van Sell, M., Brief, A. P., & Schuler, R. S. (1981). Role conflict and role ambiguity:

Integration of the literature and directions for future research. Human Relations,

34, 43-71.

Van Servallen, G., & Leake, B. (1993). Burnout in hospital nurses: a comparison of

acquired immunodeficiency syndrome, oncology, general medical, and intensive

care unit nurse samples. Journal of Professional Nursing, 9(3), 169-177.

 377

Van Yperen, N. W., Buunk, A. P., & Schaufeli, W. B. (1992). Communal orientation and

the burnout syndrome among nurses. Journal of Applied Social Psychology, 22,

173-189.

Vaux, A. (1988). Social support: Theory, research, and intervention. New York:

Praeger.

Vaux, A., Phillips, J., Holly, L., Thompson, B., Williams, D., & Steward, D. (1986). The

Social Support Appraisals Scale: Studies of reliability and validity. American

Journal of Community Psychology, 14(2), 217-233.

Vaux, A., Riedel, S., & Stewart, D. (1987). Modes of social support: The Social Support

Behaviors (SSB) Scale. American Journal of Community Psychology, 15, 209-

237.

Veiel, H. O. F., & Baumann, U. (1992). The meaning and measurement of social

support. New York: Hemisphere.

Viswesvaran, C., Sánchez, J. I., & Fisher, J. (1999). The role of social support in the

process of work stress: A meta-analysis. Journal of Vocational Behavior, 54, 314-

334.

Walters, V., & Hams, I. (1989). Workload and occupational stress in nursing. Canadian

Journal of Nursing Research, 21(3), 49-58.

Warr, P. (1996). Psychology at work (4th ed.). London: Penguin.

Weber, R. P. (1990). Basic content analysis. Beverly Hills, CA: Sage.

Webster, L., & Hackett, R. K. (1999). Burnout and leadership in community mental

health systems. Administration Policy Mental Health, 26, 387-399.

 378

Weinert C. (1987). A social support measure: PRQ 85. Nursing Research, 36(5), 273-

277.

Weinert C., & Brandt, P. (1987). Measuring social support with the Personal Roles

Questionnaire. Western Journal of Nursing Research, 9(4), 589-602.

Weinert C., & Tilden, V. P. (1990). Measures of social support: Assessment of validity.

Nursing Research, 2(4), 212-216.

Weiss, R. S. (1974). The provisions of social relationships. In Z. Rubin (Ed.), Doing

unto others (pp.17-26). Englewood Cliffs, NJ: Prentice-Hall.

Wellman, B. (1981). Applying network analysis to the study of support. In B. H.

Gottlieb (Ed.), Social networks in social support (pp. 171-200), Beverly Hills,

Sage.

Wethington, E., & Kessler, R. C. (1986). Perceived support, received support, and

adjustment to stressful life events. Journal of Health and Social Behavior, 27, 78-

89.

Wheaton, B. (1985). Models for the stress buffering functions of coping resources.

Journal of Health and Social Behavior, 26, 352-364.

Wheeler, H. H. (1994a). Occupational stress in nurses and midwives and the influence of

cognitive style on nurses’ perception of work stress. Unpublished doctoral

dissertation. University of Birmingham, Birmingham.

Wheeler, H. H. (1994b). Occupational stress in general nurses and midwives. British

Journal of Nursing, 3(10): 527-534.

Wheeler, H. H. (1997a). Nurse occupational stress research 3: A model of stress for

research. British Journal of Nursing, 6(16), 944-949.

 379

Wheeler, H. H. (1997b). Nurse occupational stress research 4: The prevalence of stress.

British Journal of Nursing, 6, (21), 1256-1260.

Wheeler, H. H. (1998). Nurse occupational stress research 5: Sources and determinants of

stress. British Journal of Nursing, 7, 40-43.

Wheeler, H. H., & Riding, R. (1994). Occupational stress in general nurses and

midwives. British Journal of Nursing, 3(10), 527-534.

Wills, T. A., & Shinar, O. (2000). Measuring perceived and received support. In S.

Cohen, L. G. Underwood, & B. H. Gottlieb (Eds.), Social support measurement

and intervention (p. 95). Oxford: Oxford University Press.

Winnubust, J., Marcelissen, F. H. G., & Kleber, R. J. (1982). Effects of social support in

the stressor-strain relationship: A Dutch sample. Social Science and Medicine, 16,

475-482.

Wolfgang, A. P. (1988a). The Health Professions Stress Inventory. Psychological

Reports, 62, 220-222.

Wolfgang, A. P. (1988b). Job stress in the health professions: A study of physicians,

nurses, and pharmacists. Behavioral Medicine, 14, 43-47.

Wright, T. F., Blache, C. F., Ralph, J., & Luterman, A. (1993). Hardiness, stress, and

burnout among intensive care nurses. Journal of Burn Care and Rehabilitation,

14, 376-381.

Yu, L. C., Mansfield, P. K., Packard, J. S., Vicary, J., McCool, W. (1989). Occupational

stress among nurses in hospital settings. American Association of Occupational

Health Nurses Journal, 37(4), 121-128.

 380

Zani, B., & Pietrantoni, L. (2001). Gender differences in burnout, empowerment and

somatic symptoms among health professional: Moderators and mediators. Equal

Opportunities International, 20, 39-48.