5 Factor structure of the Coping Inventory for

Stressful Situations (CISS-21) in adolescents

and young adults with chronic digestive
disorders

Hiske Calsbeek, M.A.¹, Mieke Rijken, Ph.D.¹, Gerard P. van Berge

Henegouwen, M.D., Ph.D.² and Joost Dekker, Ph.D.³

¹ Netherlands Institute for Health Services Research (NIVEL)

² University Medical Centre Utrecht, Department of Gastroenterology
³ VU University Medical Centre Amsterdam, Department of Rehabilitation Medicine,
Institute for Research in Extramural Medicine, Amsterdam, the Netherlands

Submitted

The social position of adolescents and young adults with chronic digestive disorders 83
Abstract

The Coping Inventory for Stressful Situations-21 (CISS-21) is a valid and

reliable measure of generic coping strategies in adult samples with various
chronic diseases. Little is known about application to a younger target group.
In the present study, confirmatory factor analysis on the CISS-21 was
performed in adolescents and young adults with various chronic digestive
disorders (total n=521) and healthy peers (n=274), aged 12-25 years. Results
provide evidence for a satisfactory fit and the invariance of the three-factor
model (task-oriented, emotion-oriented and avoidance coping scales), in
several age groups as well as in patients and controls. In conclusion, the
factor structure of the CISS-21 is being maintained when applied to younger
adolescents with and without chronic digestive disorders. This makes it
possible to compare the use of coping strategies by chronically ill
adolescents and young adults of different ages, as well as between healthy
and chronically ill adolescents and young adults.

Keywords: coping assessment, adolescents, young adults, chronic digestive

disorders

Introduction

Coping can be considered a key-concept in health psychology and related

disciplines, as coping helps to explain the impact of stressors on health and
well-being [1]. Having a chronic disease can be experienced as stressful for
the daily life of patients [2]. Also, being diagnosed with a chronic digestive
disorder, in particular suffering from a chronic liver disease or IBD, can
have a serious impact on the social position of adolescents and young adults
in daily life [3]. It was found that negative consequences occur in education,
leisure activities, labor participation, financial situation, partnership and
sexuality. As coping possibly helps to explain this impact on the social
position, we wished to investigate, and thus assess, the role of coping in
adolescents and young adults with chronic digestive disorders.
Various coping instruments have been presented in the literature [1,2,4,5].
However, little attention has been paid to assessment issues, resulting in

84 The social position of adolescents and young adults with chronic digestive disorders
many coping instruments of poor validity and reliability [1]. According to
De Ridder [1], it is imperative to elaborate on a theoretical rationale of
coping dimensions to make progress in coping assessment.
The current measurement of coping in adolescents and young adults include
both theoretically and empirically derived measuring instruments, whereby
the majority has been empirically derived. In creating scales, researchers
generally rely on exploratory factor analysis, often, however, generating
different results across samples. However, in theoretically based measures,
factors are tested through confirmatory factor analyses, supporting the
hypothesized structure of coping [5].
Compas et al. [5] identified only two measuring instruments in which
confirmatory factor analyses strongly supported the theoretical structure, i.e.
the Children’s Coping Strategies Checklist (CCSC), measuring general
coping styles and tested in age samples of 7-13 years [4] and the Responses
to Stress Questionnaire (RSQ), measuring coping in response to specific
stressors, tested in age samples 11 to 19 years [6]. However, we searched for
a generic measuring instrument that could be applied to adolescents and
young adults in the age of 12 to 25 years, with and without various chronic
digestive disorders, as we wished to compare the use of coping strategies by
adolescents and young adults of different ages and health status. Therefore,
the Coping Inventory for Stressful Situations-21 (CISS-21) [7,8] was chosen.
The CISS-21 is a theoretically derived, generic measuring instrument that
has proven to have good psychometric characteristics in adult samples (as
from 16 years). In the present study, we examined whether the theoretically
assumed factor structure of the CISS-21 is being maintained in younger
adolescents, as from the age of 12 years.
The CISS-21 has been developed to assess three coping strategies: task-
oriented, emotion-oriented and avoidance coping. Task-oriented coping
refers to purposeful task-oriented efforts aimed at solving the problem,
cognitively restructuring the problem or attempts to alter the situation. The
emphasis is on the task or planning, and to attempts to solve the problem.
Emotion-focused coping refers to emotional reactions that are self-oriented.
The aim is to reduce stress. Reactions include emotional responses, self-
preoccupation and fantasizing. Avoidance coping refers to activities and
cognitive changes aimed at avoiding the stressful situation via distracting
oneself with other situations or tasks or via social diversion as a means of
alleviating stress [7]. The development of these three dimensions was

The social position of adolescents and young adults with chronic digestive disorders 85
 grounded on consensus in the coping literature that there should be a basic
distinction between emotion-focused and problem-focused coping, whereby
emotion-focused coping strategies refer to a person orientation and problem-
focused coping strategies to a task orientation [7]. On the basis of empirical
research Endler and Parker [9] also suggested a third basic coping strategy,
namely avoidance, which included either person-oriented (social diversion)
and task-oriented (distraction) strategies.
As aforementioned, the CISS originally was developed for use in adults. The
aim of this study was to investigate whether a valid and reliable measure of
coping strategies could be adopted for use in a younger target group with
and without chronic digestive disorders. For this purpose, the theoretically
assumed three-factor structure was examined by means of confirmatory
factor analysis.

Method

Study population
This study was carried out in 521 adolescents and young adults with chronic
digestive disorders and 274 healthy controls. Digestive disorders were
categorized in five diagnostic groups, including inflammatory bowel
diseases (IBD) (n=190), chronic liver diseases (n=51), congenital digestive
disorders (n=122), coeliac disease (n=61) and food allergy (n=97). Patients
were recruited via their medical specialists in academic and specialized
hospitals using the following criteria: being diagnosed within one of the five
diagnostic categories stated by a certified medical specialist, illness duration
of at least six months, age from 12 to 25, being non-institutionalized, being
aware of diagnosis, not being terminally ill, being mentally capable to
participate and mastering the Dutch language sufficiently. Extra patients
with coeliac disease were recruited through a patient organization, using the
same criteria (diagnosis had to be confirmed by at least one small-bowel
biopsy). There was no selection on illness activity or severity of complaints
and disability.
Healthy controls were randomly recruited from the patient files of general
practitioners of participating patients. Except being diagnosed within one of
the five previously named diagnostic categories, the same criteria were used
as for the recruitment of patients. For the recruitment procedure, general

86 The social position of adolescents and young adults with chronic digestive disorders
practitioners were provided with a random set of three different letters of the
alphabet on the basis of which they were requested to select three controls
with surnames starting with the indicated letters and fitting the specified
criteria.

The study was approved by all participating hospitals' ethical committees.

All subjects gave written informed consent. Only after informed consent was
received, a written questionnaire was sent to the respondents.

In Table 1 the characteristics of the research group are summarized. The

mean age in the different groups, diagnostic groups and control group, varied
from 18.4 years in adolescents and young adults with congenital digestive
disorders to 22.1 years in the IBD group. The percentage female patients
varied from 45.9% (congenital digestive disorders group) to 70.1% (food
allergy group). The mean score on socio-economic status varied from 3.8
(chronic liver disease group) to 4.7 (coeliac disease group).

The social position of adolescents and young adults with chronic digestive disorders 87
Table 1 Characteristics of the total sample, five separate diagnostic groups and control group (total n=795)

I II III IV V
Inflammatory Chronic Congenital Coeliac Food Control group

bowel disease liver digestive disease allergy

disease disorders
n=190 n=51 n=122 n=61 n=97 n=274
Age (Mean,SD) 22.1 (2.3) 22.0 (2.7) 18.4 (3.7) 18.4 (3.7) 18.4 (3.8) 18.5 (3.8)
% 12-14 0.5 2.0 21.3 18.0 21.6 17.9
% 15-17 4.2 3.9 18.0 21.3 17.5 26.6
% 18-20 17.4 23.5 24.6 23.0 27.8 19.0
% 21-24 77.9 70.6 36.1 37.7 33.0 36.5

Gender (% women) 58.9 51.0 45.9 67.2 70.1 55.5

1
Socio-economic status (Mean, SD) 4.2 (1.7) 3.8 (2.0) 4.3 (1.7) 4.7 (1.6) 4.6 (1.6) 4.5 (1.6)

¹ Operationalized by the highest education of one of the parents or caregivers on a 7-point-scale: (1) no education or only primary school to
(7) a completed university training.
CISS-21
The CISS-21 [7] is assumed to assess coping by three basic coping
strategies: emotion-oriented, task-oriented and avoidance coping. We used
the Dutch translation as proposed and validated by De Ridder and Maes [8].
Each scale of the CISS-21 consists of 7 items, randomly distributed within
the form to control for order effects (see Appendix). Respondents were
asked to rate each item on a five point scale ranging from (1) “not at all” to
(5) “very much”.

Data analysis

In order to assess the fit of the theoretical supposed three-factor structure and
the factor invariance of the CISS-21 in a younger target group and in
adolescents and young adults with and without chronic digestive disorders,
confirmatory factor analysis was applied, following the procedure of Rahim
and Magner [10]. Confirmatory factor analysis is a powerful method of
investigating the construct validity of a scale [5,10,11]. In the present study
confirmatory analysis was performed with the LISREL 8 computer package
[12].
In order to test the three-factor structure, a series of confirmatory factor
analyses was performed: in 4 age groups (age 12-14, 15-17, 18-21 and 21-24
years); in patients (as 1 group); controls (1 group) and in 5 separate
diagnostic groups (IBD, chronic liver diseases, congenital digestive
disorders, coeliac disease and food allergy). In the measurement model, each
of the 21 items was allowed to load on only its associated factor (which was
identified a priori), and the factors (three coping strategies) were allowed to
correlate. The covariance matrix for the 21 items was used for performing
the analysis, and parameter estimates were made under the maximum-
likelihood method.
Following Rahim and Magner [10], the extent to which the theoretical three-
factor model fits the data in the different samples was assessed first. For this
purpose, LISREL provides several measures, one of which is the chi-square
statistic. Non-significant chi-squares suggest a satisfactory fit for the tested
model; significant chi-squares an unsatisfactory fit. However, the chi-square
is dependent on sample size, such that a large sample is likely to produce a

The social position of adolescents and young adults with chronic digestive disorders 89
significant result, even when there is a reasonable good fit to the data [13].
Except the chi-square, LISREL provides other statistics, such as the
goodness-of-fit index (GFI), which is based on the chi-square. In general,
this measure ranges between 0 and 1, with higher values indicating a better
fit. Next to the GFI, the normed-fit index (NFI) was computed. This measure
assesses the fit of the proposed model relative to that of the null model, and
is independent of sample size. It is suggested that .90 is a minimum value for
satisfactory fit when using both indices [13]. In addition, the Root Mean
Square Error of Approximation (RMSEA) was used (12,14,15]. The
RMSEA takes account of the error of approximation in the population and
the precision of the fit measure itself. A close fit of the model is indicated by
a value lower than of .05 [14] or .06 [16], whereas values up to .08 represent
reasonable errors in the population [14]. According to Jöreskog and Sörbom
[12], we used a test of RSMEA < .05 (90 percent confidence interval).
To provide additional reference points [10,13], indices were also computed
for a null model, i.e no relationship between the observed variables, and a
one-factor model, i.e. all observed variables refer to only one factor: an
undifferentiated coping strategy.
However, models with many variables and degrees of freedom will almost
always have significant chi-squares due to high levels of random error found
in typical items and the many parameters that must be estimated [10]. To
address this problem, several authors [17,18] proposed a method in which
subsets of items within factors are summed to create aggregate variables.
Using these parcels, it is appropriate to have two aggregate variables per
factor when the number of measured items per factor is five to seven. Like
Rahim and Magner [10], we followed this method, and thus formed two
parcels for each factor, creating six parcels in total. In the formation of these
parcels, the first four items and the second three items within a factor (see
Appendix) were summed (divided by respectively 4 and 3 to obtain the same
scale scores) to create two parcels per factor. So, instead of 21 variables
(seven items per factor, three factors), we created 6 aggregate variables (two
variables per factor, three factors). Next to analyses based on the 21
observed items, measures of fit were also computed on the basis of these
parcels.
After testing the three-factor model, two multi group analyses were
conducted in order to examine the invariance of the three-factor model
across different groups: across age groups (1) and across patients, controls

90 The social position of adolescents and young adults with chronic digestive disorders
 and different diagnostic groups (2). For each multi group analysis a
covariance matrix was computed for each group. Then the following four
models were estimated and compared sequentially on the basis of fit [10,11]:
model 1, in which the pattern of factor loadings was held invariant across
groups; model 2, the pattern of factor loadings and the factor loadings were
held invariant across groups; model 3, the pattern of factor loadings, the
factor loadings and the errors were held invariant across groups; and model
4, the most restrictive model, in which the pattern of factor loadings, the
factor loadings, the errors as well as the variances and covariances were held
invariant across groups. For each model, the covariance matrices for all
groups were analyzed simultaneously. In performing these tests, the
following statistics were computed: chi-square, GFI, NFI and RMSEA. As a
test for equal factor loadings across groups, the chi-square of model 2 was
compared with the chi-square of model 1. A non-significant difference
means that the hypothesis of equal factor loadings cannot be rejected on a
statistical basis. In the same way, model 3 was compared with model 2 and
model 4 with model 3.
Finally, the internal consistency (Cronbach’s alpha) of the constructed scales
was computed in the several samples.

Results
Factor structure, separate items
Table 2 (columns left) shows indices that were used to assess the extent to
which the proposed three-factor model fits the data in the four age samples.
For comparative purposes, fit indices are also presented for a null model and
a one-factor model.
Chi-square tests for the three-factor model were significant, suggesting an
unsatisfactory fit. However, the chi-square is dependent on sample size: a
large sample is likely to produce a significant result even when there is a
reasonably good fit to the data (see Methods section). Applying the .90
criterion for GFI and NFI, the three-factor model only has a moderate fit: the
GFI, ranges from .74 to .87, showing the best fit in the oldest sample and the
worst fit in the youngest sample. The NFI also indicates a moderate fit,
ranging from .59 in the youngest sample to .77 in the oldest sample.
Applying the .05-.08 criterium for the RMSEA, results show a moderate fit

The social position of adolescents and young adults with chronic digestive disorders 91
of the three-factor model in two age-groups. However, significant p-tests
indicate high errors of approximation.
After this, the three-factor model was tested in patients (as 1 group) and the
control group and in the different diagnostic groups. The results are
summarized in Table 3 (columns left). Again, significant chi-squares were
found in combination with measures of fit ranging from .53 to .87, and
significant RMSEA’s. These results only indicate a moderate fit.
As expected, these results suggest that the theoretical three-factor model has
only a moderate fit to the data. As pointed out before, parcels were
introduced in the analyses instead of separate items.

Factor structure, parcels

Table 2 (columns right) presents goodness-of-fit indices for the three-factor
model in the four age samples based on analyses with parcels. Only in the
two oldest age groups significant chi-squares were found. The GFI was high,
ranging from .97 to .99. Furthermore, the NFI also exceeds the .90 criterion
in all age samples. The RMSEA ranged from zero to .10, only significant in
the oldest age group. These results indicate that the three-factor model has a
satisfactory fit in the four age samples.
The same analyses were conducted in patients (as one group) and in controls,
and in the five different diagnostic groups (Table 3, columns right). GFI and
NFI both exceeded the .90 criterion in all samples. In two samples the
RMSEA was found significant. These results indicate a reasonable to
satisfactory fit in the different samples.

92 The social position of adolescents and young adults with chronic digestive disorders
Table 2 Measures of fit for separate age samples, based on analysis of 21 separate variables (columns left)
and 6 parcels (columns right) of the CISS-21
Sample and n Chi2 df GFI NFI RMSEA Chi2 df GFI NFI RMSEA
model
Age 21-24 years 381
null model 2,675* 210 809* 15
1-factor model 1,813* 189 .60 .32 .15* 539* 9 .73 .33 .39*
3-factor model 606* 186 .87 .77 .08* 30* 6 .97 .96 .10*

Age 18-20 years 167

null model 1,434* 210 380* 15
1-factor model 913* 189 .58 .36 .15* 194* 9 .77 .49 .35*
3-factor model 445* 186 .80 .69 .09* 15* 6 .97 .96 .10 (ns)

Age 15-17 years 131

null model 1,007* 210 302* 15
1-factor model 653* 189 .60 .35 .14* 156* 9 .76 .48 .35*
3-factor model 303* 186 .84 .70 .07* 3 6 .99 .99 .00 (ns)

Age 12-14 years 103

null model 923* 210 245* 15
1-factor model 554* 189 .62 .40 .14* 95* 9 .78 .61 .31*
3-factor model 376* 186 .74 .59 .10* 10 6 .97 .96 .08 (ns)
Note: GFI=goodness-of-fit index; NFI=normed fit index; RMSEA=Root Mean Square Error of Approximation
 p ≤ .05; ns = non significant
Table 3 Measures of fit for controls, patients (total group) and separate diagnostic groups, based
on analysis of 21 separate variables (columns left) and 6 parcels (columns right)

2 2
Sample and model n Chi df GFI NFI RMSEA Chi df GFI NFI RMSEA

Controls 269
null model 1,941* 210 601* 15
1-factor model 1,282* 189 .60 .34 .15* 371* 9 .73 .38 .39*
3-factor model 477* 186 .85 .75 .08* 22* 6 .97 .96 .10*

Patients (total group)

null model 513 3,736* 210 1,197* 15
1-factor model 2,279* 189 .59 .39 .15* 728* 9 .72 .39 .39*
3-factor model 760* 186 .87 .80 .08* 39* 6 .97 .97 .10 (ns)

IBD 190
null model 1,512* 210 428* 15
1-factor model 1,037* 189 .57 .31 .15* 285* 9 .72 .33 .40*
3-factor model 464* 186 .81 .69 .09* 27* 6 .96 .94 .14*

Chronic liver disease 51

null model 618* 210 137* 15
1-factor model 466* 189 .50 .25 .17* 79* 9 .71 .42 .40*
3-factor model 275* 186 .69 .55 .10* 7 6 .95 .95 .04 (ns)
Table 3 continued

2 2
Sample and model n Chi df GFI NFI RMSEA Chi df GFI NFI RMSEA

Congenital disorders 115

null model 1,071* 210 273* 15
1-factor model 633* 189 .61 .41 .14* 106* 9 .79 .61 .31*
3-factor model 393* 186 .76 .63 .10* 11 6 .97 .96 .09 (ns)

Coeliac disease 60
null model 711* 210 162* 15
1-factor model 488* 189 .53 .31 .16* 99* 9 .71 .39 .41*
3-factor model 336* 186 .67 .53 .12* 7 6 .96 .95 .07 (ns)

Food allergy 97
null model 1,258* 210 254* 15
1-factor model 923* 189 .50 .27 .20* 162* 9 .69 .36 .42*
3-factor model 591* 186 .69 .53 .15* 10 6 .97 .96 .09 (ns)

Note: GFI=goodness-of-fit index; NFI=normed fit index; RMSEA=Root Mean Square Error of Approximation
Note: measures of fit in sample chronic liver diseases are based on preliminary solutions (model did not converge after X iterations)
* p.≤05; ns=not significant
Factor invariance
Two multi-group analyses were conducted to examine the invariance of the
three-factor model of the CISS-21 across different samples: in age groups
and in patient groups. In these analyses we continued using parcels. In each
multi-group analysis 4 models were tested (see methods section). The first
multi-group analysis was performed in age samples to test the invariance of
the structure across age groups. Results of this analysis are presented in
Table 4. Although the chi-square of each model was significant, the other
indices (GFI, NFI and RMSEA) provided evidence that the first three
models have a good fit: the goodness-of-fit index for each of these models
was found .95 or higher, each normed fit index .94 or higher and each
RMSEA lower than .05. Furthermore, as a test of the hypothesis of equal
factor loadings across the groups, the chi-square of model 2 was compared
with the chi-square of model 1, whereas the difference was found to be non-
significant, meaning that the hypothesis of equal factor loadings cannot be
rejected on a statistical basis. In the same manner, model 3 was compared
with model 2 and model 4 with model 3, to test -respectively- the hypothesis
of equal errors and equal variances. Table 4 shows that the hypotheses of
equal pattern, loadings and errors cannot be rejected; the hypothesis of equal
variances / covariances should be rejected. In other words, these findings not
only indicate a satisfactory fit of the hypothesized three-factor structure, also
the factor loadings and error terms appear to be practically identical in the
different samples. Only the variance was found to be different in the various
samples.

96 The social position of adolescents and young adults with chronic digestive disorders
Table 4 Invariance analysis across four age samples
2 2 2
Model and sample Chi - Chi Chi df GFI RMSE
contribution df (differenc NFI A
(total e) (total
model) model)

Null model 1,735*

60

Equal factor
58* .97 .97 .04(ns)
pattern(1) 24
*12-14 10(18%)
*15-17 3(5%)
*18-20 15(26%)
*21-24 30(51%)

Equal factor 88* 30(ns) 18 .96 .95 .04(ns

pattern and 42
laodings(2)
*12-14
*15-17 15(17%)
14(16%)
*18-20 21(24%)
*21-24 38(44%)

Equal 98* 10(ns) 18 .95 .94 .03(ns)

factorpattern, 60
laodings and
errors(3)
*12-14
*15-17 16(16%)
*18-20 24(24%)
*21-24 41(42%)

Equal factor 1,648* 1,550* 9 .60 .05 .17*

pattern, loadings, 69
errors, and
variances /
covariances(4)
*12-14
*15-17 290(18%)
*18-20 355(22%)
*21-24 790(48%)

Note: GFI=goodness-of-fit index; NFI=normed fit index; RMSEA=Root Mean Square Error
of Approximation
p ≤.05; ns=not significant

The social position of adolescents and young adults with chronic digestive disorders 97
Table 5 presents results of the second multi group analysis of the three-factor
model: across patients and controls (two samples). Again, the chi-square of
each model was significant, but the other indices provide evidence that the
first three models have a good fit: the goodness-of-fit index for each of these
models was found .97 or higher, each normed fit index .96 or higher and the
RMSEA between .05 and .07. In the first model the RMSEA was found
significant, however, with a score of .07 representing reasonable errors.
Comparisons between the models provide evidence for the hypotheses of an
equal pattern of factor loadings, equal factor loadings and invariant errors.
The significant chi-square difference between model 4 and model 3
indicates, again, that the hypothesis of equal variances and covariances
should be rejected.
The LISREL-model did not converge in testing the factor invariance across
the five different diagnostic groups. Therefore, these analyses were only
carried out in patients as one group, compared with the control group.

In general, these results provide strong support for the invariance of the
three-factor model across age groups, as well as controls and patients, with
respect to factor pattern, factor loadings and error terms.

98 The social position of adolescents and young adults with chronic digestive disorders
Table 5 Invariance analysis across control sample and patient sample
2 2 2
Model and sample Chi - Chi Chi df GFI RMSE
contribution df (differenc NFI A
(total e) (total
model) model)

Null model 1,797*

30

Equal factor 61* .97 .97 .07*

pattern(1) 12
*patients
39(64%)
*controls 22(36%)

Equal factor 63* 2(ns) 6 .98 .96 .06(ns)

pattern and 18
laodings(2)
*patients
*controls 40(63%)
23(37%)

Equal 73* 10(ns) 6 .97 .96 .05(ns)

factorpattern, 24
laodings and
errors(3)
*patients
*controls 30(41%)

Equal factor 1,663* 1,590* 3 .66 .08 .28*

pattern, loadings, 27
errors, and
variances /
covariances(4)
*patients
*controls 567(34%)

Note: GFI=goodness-of-fit index; NFI=normed fit index; RMSEA=Root Mean Square Error
of Approximation
p ≤.05; ns=not significant

Scale construction
Although the previous results of confirmatory factor analysis provide
evidence for a satisfactory fit of the total model, the factor loading of the
first item 'Take some time off and get away from the situation' on the factor

The social position of adolescents and young adults with chronic digestive disorders 99
 avoidance (see Appendix) was not significant in most of the samples (except
for the age sample 18-20 years and the food allergy sample, p≤.01).
Therefore, we suggest to exclude this item from the scale 'avoidance' in
samples of adolescents and young adults with chronic digestive disorders
and in these age groups without chronic digestive disorders. In the so
constructed scales, Cronbach's alpha’s varied in the different samples from
.79 to .86 in task-oriented coping (7 items), .79 to .86 in emotion-oriented
coping (7 items) and from .78 to .85 in the avoidance coping scale (6 items).

Discussion

The aim of the present study was to determine whether the theoretically
assumed three-factor structure of the CISS-21, a valid and reliable coping
measurement instrument in adults, could be reproduced in adolescents and
young adults aged 12 to 25 years, with and without various chronic digestive
disorders. For this purpose, the fit of the three-factor model was investigated
by means of confirmatory factor analysis. Results indicate that the three-
factor structure of the CISS-21 has a satisfactory fit, in the younger age
samples as well as in the different diagnostic groups and in the control group
with healthy peers. In adult samples, the CISS-21 has proven to have good
psychometric characteristics; apparently, in younger age samples, generic
coping strategies can also be categorized as task-focused, emotion-focused
and avoidance coping.
Except a satisfactory fit of the hypothesized three-factor structure of the
CISS-21, general support was obtained for factor invariance across age
samples and across patient and control samples. For this purpose, 4 models
were successively compared in multi group analyses, running up from an
equal factor pattern in model 1 to the most restrictive model 4, in which,
apart from the factor pattern, the factor loadings and the error terms, the
variance also was held invariant. Through the first three models, satisfactory
results were obtained. This means that in age groups as well as in patients
and controls, the factor pattern, the factor loadings and the error terms
appeared to be practically identical in the various samples. These results
strongly support the theoretical three-factor structure of the CISS-21 in a
younger target group and in adolescents and young adults with chronic
digestive disorders. Only the variance - added in the latter model to be held

100 The social position of adolescents and young adults with chronic digestive disorders
 invariant across the samples - did not lead to satisfactory results.
Discrepancies in variance between groups possibly indicate that differences
exist in the use of coping strategies between age groups and between patient
and controls.
Some comments have to be made upon these results. First, in testing the
factor structure, the measures of fit in the sample chronic liver diseases were
based on preliminary solutions (the model did not converge). This is
probably due to small numbers. However, exploratory factor analyses (data
not shown) also demonstrated the theoretical model structure in this sample,
supporting the preliminary solutions of the confirmatory factor analyses. In
addition, with regard to the factor invariance tests, patients as a whole group
were compared with controls first, resulting in evidence for a high extent of
invariance across these samples. After that, multi-group analyses were
carried out in the five different diagnostic groups. However, the LISREL-
model did not converge in these analyses, probably as a consequence of the
application of a complex model in relatively small subgroups. Again, results
of exploratory factor analyses (data not shown) supported our hypothesis
that the theoretical three-factor structure does exist across the separate
diagnostic groups as well. Taking these comments into account, it can be
concluded that the factor structure of the CISS-21 is being maintained in
younger adolescents with and without various chronic digestive disorders.

References
1 De Ridder D. What is wrong with coping assessment? A review of conceptual and
methodological issues. Psychology and Health 1997;12:417-431.
2 Boekaerts M, Röder I. Stress, coping and adjustment in children with a chronic disease: a
review of the literature. Disability and Rehabilitation 1999;7: 311-337.
3 Calsbeek H, Rijken PM, Bekkers JTM, Kerssens JJ, Dekker J, Van Berge Henegouwen
GP, participating centres. Social position of adolescents with chronic digestive disorders.
European Journal of Gastroenterology & Hepatology 2002;14:543-549.
4 Ayers TS, Sandler IN, Twohey JL. Conceptualization and measurement
of coping in children and adolescents. In: Ollendick TH, Prinz RJ (Eds.). Advances in
Clinical Child Psychology (pp. 243-301). New York: Plenum Press; 1998.
5 Compas BE, Connor-Smith JK, Saltzman H, Thomsen AH, Wadsworth ME. Coping with
stress during childhood and adolescence: problems, progress an potential in theory and
research. Psychological Bulletin 2001;1:87-127.

The social position of adolescents and young adults with chronic digestive disorders 101
6 Connor-Smith JK, Compas BE, Wadsworth ME, Thomsen AH, Saltzman H. Responses
to stress in adolescence: Measurement of coping and involuntary stress responses. Journal
of Consulting and Clinical Psychology 2001;6:976-992.
7 Endler NS, Parker DA. Coping Inventory for Stressful Situations (CISS): Manual (2nd
ed.). Toronto: Multi Health Systems;1999.
8 De Ridder DTD & Maes S. Dutch translation of the CISS-21. Lisse: Swets & Zeitlinger
(Swets Test Publishers); in press.
9 Endler NS, Parker JDA. The multidimensional assessment of coping: a critical evaluation.
Journal of Personality and Social Psychology 1990;58:844-855.
10 Rahim MA, Magner NR. Confirmatory factor analysis of the styles of handling
interpersonal conflict: first-order factor model and its invariance across groups. Journal of
Applied Psychology 1995;1:122-132.
11 Schmitt N, Stults DN. Methodology review: Analysis of multi trait-multi method
matrices. Applied Psychological Measurement 1986;10:1-22.
12 Jöreskog KG, Sörbom D. LISREL 8. Chicago: SPSS;1993.
13 Bentler PM, Bonnett DG. Significance tests and goodness-of-fit in the analysis of
covariance structures. Psychological Bulletin 1980;88:588-606.
14 Browne MW, Cudeck R. Alternative ways of assessing model fit. In: Bollen KA, Long JS
(Eds.). Testing structural equation models (p. 136-162). Newbury Park: Sage;1993.
15 Ferrando PJ, Lorenza-Seva U. Unrestricted versus restricted factor analysis of
multidimensional test items: some aspects of the problem and some suggestions.
Psicológica 2000;21:301-323.
16 Hu L, Bentler PM. Cutoff criteria for fit indices in covariance structure analysis:
conventional criteria versus new alternatives. Structural Equation Modeling 1999;1:1-55.
17 Bagozzi RP, Heatherton TF. A general approach to representing multifaceted personality
constructs: application to state self-esteem. Structural Equation Modeling 1994;1:35-67.
18 Floyd FJ, Widaman KF. Factor analysis in the development and refinement of clinical
assessment instruments. Psychological assessment 1995;7:286-299.
19 Jöriskog KG. Simultaneous factor analysis in several populations.
Psychometrika 1971;36:409-426.

102 The social position of adolescents and young adults with chronic digestive disorders
Appendix
Shortened version (21-items) Coping Inventory for Stressful Situations
(CISS-21)

1. Take some time off and get away from the situation (A) *
2. Focus on the problem and see how I can solve it (T)
3. Blame myself for having gotten into this situation (E)
4. Treat myself to a favorite food or snack (A)
5. Feel anxious about not being able to cope (E)
6. Think about how I solved similar problems (T)
7. Visit a friend (A)
8. Determine a course of action and follow it (T)
9. Buy myself something (A)
10. Blame myself for being too emotional about the situation (E)
11. Work to understand the situation (T)
12. Become very upset (E)
13. Take corrective action immediately (T)
14. Blame myself for not knowing what to do (E)
15. Spend time with a special person (A)
16. Think about the event and learn from my mistakes (T)
17. Wish that I could change what had happened or how I felt (E)
18. Go out for a snack or meal (A)
19. Analyze my problem before reacting (T)
20. Focus on my general inadequacies (E)
21. Phone a friend (A)

T = Task-oriented coping
E = Emotion-oriented coping
A = Avoidance coping

* On the basis of our results it is suggested to exclude this item from the scale 'Avoidance
coping' in samples of adolescents and young adults.

The social position of adolescents and young adults with chronic digestive disorders 103