CEBQ

APPPETITIVE TRAITS AND THEIR RELATIONSHIPS TO WEIGHT AND
WEIGHT MANAGEMENT
Claudia Madeleine Elizabeth Hunot
A thesis submitted for the degree of Doctor of Philosophy
UNIVERSITY COLLEGE LONDON

Declaration
Declaration
I, Claudia Madeleine Elizabeth Hunot, confirm that the work presented in this thesis is my
own. Where information has been derived from other sources, I confirm that this has been
indicated in the thesis.
Acknowledgements
Acknowledgements
Firstly, I would like to thank Professor Jane Wardle for her support, and for giving me the
opportunity to study at the Health Behaviour Research Centre (HBRC), and learn from such
a talented group of researchers. She sadly passed away in October 2015 and I was lucky to
be one of her last students. To my primary supervisor Dr Rebecca Beeken my greatest
thanks for your continual guidance, as well as to Dr Alison Fildes, Dr Helen Croker and Dr
Fiona Johnson, thank you all for your helpful comments and constant support throughout
this process. I would also like to thank Dr Clare Llewellyn for your statistical direction.
I would like to acknowledge the financial backing that I received from the ‘Programa para el
Desarrollo Profesional Docente’ (PRODEP) - Secretaría de Educación Pública (SEP) – México;
as well as from the Health Sciences University Centre, of the University of Guadalajara that
allowed me the time to complete this research. To my journey partner Priscila López
Torres, may this time spent away be helpful to us both and our students upon our return to
work.
I would like to thank all my colleagues at the HBRC, with whom there has been a great
exchange of ideas and encouragement to grow as a researcher, as well as for all your
valuable support. To Lindsay Kobayashi, thank you for all your help and friendship
throughout this process. To my three amigas, Elaine Douglas, Merle Owens and Lesley
McGregor, thank you for your laughter, inspiration and endless encouragement.
To Kate C., thank you for your unconditional backing and friendship over the last 30 years.
For opening your home to us, giving us Fleet and allowing us to share this family
experience.
To my parents Patrick and Jane who put up with me, my children, my health, and who
always believed in me and never gave up. To my sisters and my brother, and all my family,
all of my love and endless gratitude.
To my husband Gabo who taught me that the other PhD that had to be done and was
worth fighting for, was relationship and family; thank you for your love and support. To my
sons, Mateo and Alex, you are my inspiration and to you both I dedicate this work.
3
Abstract
Abstract
There is a need for novel approaches to weight management (WM) for adults to address
the increasing prevalence of obesity. Appetitive traits (ATs) are potentially modifiable
stable predispositions towards food, which could be targeted by tailored WM interventions.
Research has demonstrated associations between ATs and BMI in children, measured using
the parent report ‘Child Eating Behaviour Questionnaire’ (CEBQ). This thesis systematically
reviews the psychometric measures of ATs currently available for adults and children (Study
1). This review highlighted that the specific ATs captured by the CEBQ have not been
measured in adults and so their relationships to weight remains unexplored beyond
childhood. This review therefore demonstrated a need for a self-report version of the
CEBQ, the ‘Adult Eating Behaviour Questionnaire’ (AEBQ). Study 2 describes the
development of the AEBQ as a reliable measure of ATs in adults. Study 3 confirmed the
AEBQ factor structure in a different sample, and showed that ATs were associated with BMI
in adults. Study 4 describes the development and preliminary testing of a brief Appetitive
Trait Tailored Intervention (ATTI) based on participants’ AEBQ scores, to help with WM in
overweight and obese adults. Study 5 involved qualitative analysis of semi-structured
interviews with participants from Study 4 to provide in-depth understanding of their
experiences of the ATTI. Overall, findings suggest that ATs can be measured in adults using
the AEBQ, and they have similar associations with BMI to those seen in children. Using
AEBQ scores to provide tailored AT feedback for WM shows promise, however refinement
of the tips and delivery method is needed prior to further testing of this approach.
4
Table of Contents
Table of Contents
Declaration ............................................................................................................................... 2
Acknowledgements .................................................................................................................. 3
Abstract .................................................................................................................................... 4
Table of Contents ..................................................................................................................... 5
List of Tables ........................................................................................................................... 12
List of Figures.......................................................................................................................... 14
List of Appendices .................................................................................................................. 15
Abbreviations ......................................................................................................................... 17
Chapter 1. Weight management .......................................................................................... 19

1.1 The need for weight management interventions ....................................................... 19
1.2 Approaches to weight management ........................................................................... 20
1.2.1 Structural approaches ......................................................................................... 21
1.2.2 Pharmacological interventions ............................................................................ 22
1.2.3 Surgical interventions .......................................................................................... 23
1.2.4 Lifestyle interventions ......................................................................................... 24
1.3 Individual factors influencing the likelihood of successful weight management ....... 27
1.4 Tailoring weight management interventions to improve outcomes .......................... 29
1.5 Summary of the findings ............................................................................................. 31
Chapter 2. Appetitive traits and weight............................................................................... 32

2.1 Introduction................................................................................................................. 32
2.2 Appetite ....................................................................................................................... 32
2.2.1 The homeostatic control of appetite .................................................................. 33
2.2.2 The hedonic control of appetite .......................................................................... 34
2.2.3 Interplay of homeostatic and hedonic control of appetite ................................. 34
2.3 Appetitive traits ........................................................................................................... 35
2.4 Measures of appetite and appetitive traits................................................................. 36
5
Table of Contents
2.4.1 Experimental measures of appetite and appetitive traits .................................. 37
2.4.2 Psychometric measures of appetite and appetitive traits .................................. 41
2.5 Appetitive traits and weight ........................................................................................ 43
2.5.1 The genetics of obesity........................................................................................ 43
2.5.2 The “Behavioural Susceptibility Theory” of obesity ............................................ 45
2.5.3 Evidence for the “Behavioural Susceptibility Theory” ........................................ 45
2.6 Summary of the findings ............................................................................................. 48
Chapter 3. Aims of the thesis ............................................................................................... 51

3.1 My contributions to the research in this thesis .......................................................... 52
Chapter 4. Study 1: Systematic review on questionnaire measures of appetite and

appetitive traits ..................................................................................................................... 54
4.1 Introduction................................................................................................................. 54
4.2 Aim .............................................................................................................................. 55
4.3 Methods ...................................................................................................................... 55
4.3.1 Information sources and search strategy ............................................................ 55
4.3.2 Eligibility criteria .................................................................................................. 56
4.3.3 Study selection .................................................................................................... 57
4.3.4 Data extraction process – Classifying and coding studies ................................... 58
4.3.5 Assessing the robustness of the questionnaires ................................................. 58
4.3.6 Most commonly measured appetitive traits by age group ................................. 60
4.4 Results ......................................................................................................................... 60
4.4.1 Search results ...................................................................................................... 60
4.4.2 Evaluation of psychometric properties: Reliability and validity of the

questionnaires ..................................................................................................................... 78
4.4.3 Overall robustness of the questionnaires ........................................................... 78
4.4.4 Most commonly measured appetitive traits by age group ................................. 84
4.5 Discussion .................................................................................................................... 85
4.5.1 Measures that emcompass ‘emotional’ aspects of appetite .............................. 87
6
Table of Contents
4.5.2 Measures that emcompass ‘restraint’ aspects of appetite ................................. 89
4.5.3 Measures that emcompass ‘food and eating/externality’ aspects of appetite .. 89
4.5.4 Limitations ........................................................................................................... 90
4.5.5 Conclusions.......................................................................................................... 94
Chapter 5. Study 2: Development of the ‘Adult Eating Behaviour Questionnaire’ ............ 96

5.1 Background.................................................................................................................. 96
5.2 Aims ............................................................................................................................. 96
5.3 Methods ...................................................................................................................... 98
5.3.1 Translation of the CEBQ into the AEBQ ............................................................... 98
5.3.2 Review of items from other questionnaires on appetite from existing

literature 99
5.3.3 Piloting in a sample of adults ............................................................................ 101
5.3.4 Assessing the factor structure of the AEBQ ...................................................... 102
5.3.5 Summary statistics............................................................................................. 107
5.3.6 Internal reliability .............................................................................................. 107
5.4 Ethical approval ......................................................................................................... 108
5.5 Results ....................................................................................................................... 108
5.5.1 Assessment of the factor structure ................................................................... 108
5.5.2 Criteria used for eliminating items .................................................................... 109
5.5.3 Summary statistics............................................................................................. 114
5.5.4 Reliability ........................................................................................................... 118
5.6 Discussion .................................................................................................................. 120
5.6.1 Factor structure ................................................................................................. 120
5.6.2 Correlations between sub-scales ...................................................................... 121
5.6.3 Limitations ......................................................................................................... 122
5.6.4 Conclusions........................................................................................................ 123
7
Table of Contents
Chapter 6. Study 3: Confirmation of the factor structure of the ‘Adult Eating Behaviour
Questionnaire’ ..................................................................................................................... 124
6.1 Background................................................................................................................ 124
6.2 Aim and hypothesis ................................................................................................... 125
6.3 Methods .................................................................................................................... 125
6.3.1 Design and study population ............................................................................. 125
6.3.2 Measures ........................................................................................................... 126
6.3.3 Statistical analysis .............................................................................................. 127
6.5 Results ....................................................................................................................... 131
6.5.1 Sample ............................................................................................................... 131
6.5.2 Confirmatory Factor Analysis ............................................................................ 132
6.5.3 Internal and external reliability ......................................................................... 143
6.5.4 Descriptive statistics of the appetitive trait ...................................................... 143
6.5.5 Relationships between BMI and appetitive traits ............................................. 144
6.6 Discussion .................................................................................................................. 145
6.6.1 Relationships with BMI and appetitive traits .................................................... 146
6.6.2 Limitations ......................................................................................................... 148
6.6.3 Conclusions........................................................................................................ 149
Chapter 7. Study 4: Development of a brief appetitive trait tailored intervention in a

sample of overweight and obese adults ............................................................................ 150
7.1 Background................................................................................................................ 150
7.2 Aims and objectives................................................................................................... 151
7.3 Intervention development ........................................................................................ 152
7.3.1 Step 1. Define and understand the problem and its causes ............................. 152
7.3.2 Step 2. Clarify which causal or contextual factors are malleable and have
greatest scope for change ................................................................................................. 153
7.3.3 Step 3. Identify how to bring about change: the change mechanism .............. 155
8
Table of Contents
7.3.4 Step 4. Identify how to deliver the change mechanism .................................... 159
7.4 Methods - Step 5. Testing the intervention on a small scale .................................... 160
7.4.1 Participants ........................................................................................................ 160
7.4.2 Inclusion and exclusion criteria ......................................................................... 160
7.4.3 Measures ........................................................................................................... 160
7.4.4 Recruitment ....................................................................................................... 163
7.4.5 The intervention ................................................................................................ 163
7.5 Analyses ..................................................................................................................... 163
7.5.1 Descriptive statistics .......................................................................................... 163
7.7 Results ....................................................................................................................... 164
7.7.1 Number of tips followed ................................................................................... 164
7.7.2 Response rate and loss to follow-up ................................................................. 165
7.7.3 Effects on weight ............................................................................................... 168
7.7.4 Compliance, perceived usefulness and barriers to use of the tips ................... 168
7.7.5 Use of other weight loss programs followed alongside the tips ....................... 178
7.8 Discussion .................................................................................................................. 178
7.8.1 Development and testing of the ATTI ............................................................... 178
7.8.2 Effects on weight ............................................................................................... 180
7.8.3 Compliance, perceived usefulness and barriers to use of the tips ................... 181
7.8.4 Limitations ......................................................................................................... 183
7.8.5 Conclusions........................................................................................................ 186
Chapter 8. Study 5: Participant experiences of a brief appetite-based weight

management intervention (ATTI)........................................................................................ 187
8.1 Background................................................................................................................ 187
8.2 Aim ............................................................................................................................ 187
8.3 Methods .................................................................................................................... 188
8.3.1 Study design ...................................................................................................... 188
9
Table of Contents
8.3.2 Participants & recruitment ................................................................................ 188
8.3.3 Interview topic guide......................................................................................... 189
8.4 Analyses ..................................................................................................................... 189
8.4.1 Thematic analysis of interviews ........................................................................ 189
8.4.2 Coding the interviews........................................................................................ 190
8.4.3 Data saturation .................................................................................................. 190
8.4.4 Themes .............................................................................................................. 190
8.6 Results ....................................................................................................................... 191
8.6.1 Participants ........................................................................................................ 191
8.6.2 Themes .............................................................................................................. 191
8.7 Discussion .................................................................................................................. 201
8.7.1 Experience of the intervention.......................................................................... 202
8.7.2 Consequences of the intervention .................................................................... 205
8.7.3 Barriers and facilitators to adherence............................................................... 205
8.7.4 Limitations ......................................................................................................... 207
8.7.5 Conclusions........................................................................................................ 209
Chapter 9. General Discussion............................................................................................ 210

9.1 Introduction............................................................................................................... 210
9.2 Summary of findings and contribution to the literature ........................................... 210
9.2.1 What psychometric measures of appetitive traits currently exist and how do
they relate to weight? ....................................................................................................... 211
9.2.2 Can the parent report Ćhild Eating Behaviour Questionnaire´ (CEBQ) be
adapted into a valid and reliable measure of appetitive traits in adults? ........................ 212
9.2.3 How do appetitive traits relate to BMI in adults? ............................................. 212
9.2.4 Can a weight management intervention tailored to an individual´s appetitive

traits be developed that is acceptable and potentially useful? ........................................ 213
9.3 Limitations ................................................................................................................. 215
10
Table of Contents
9.3.1 Self-reported data ............................................................................................. 216
9.3.2 Cross-sectional data .......................................................................................... 216
9.3.3 Loss to follow-up ............................................................................................... 217
9.4 Implications for future research, practice and policy ............................................... 218
9.5 Concluding remarks ................................................................................................... 222
References ............................................................................................................................ 224

Appendices ........................................................................................................................... 253
11
List of Tables
List of Tables
Chapter 2
Table 2.1 Appetitive traits ........................................................................................................ 36
Chapter 4
Table 4.1 Inclusion and exclusion criteria for a systematic review of psychometric measures
of appetite................................................................................................................................. 57
Table 4.2 Characteristics of studies included in the systematic review .................................. 63
Table 4.3 Validity and reliability measures of questionnaires from the systematic review .... 80
Table 4.4 The three most commonly used psychometric measures of appetite and the traits
they measure by age group: The TFEQ (3020 citations), the DEBQ (1700 citations) and the
CEBQ (460 citations) ................................................................................................................. 84
Chapter 5
Table 5.1 Characteristics of the sample used to carry out PCA of the 47-item AEBQ (n=708)108
Table 5.2 Criteria used for the removal of items in the AEBQ factor analysis ....................... 110
Table 5.3 Factor loadings of a 35-item AEBQ......................................................................... 112
Table 5.4 Descriptive statistics of appetitive traits in the 35-item AEBQ (n=708)................. 115
Table 5.5 Correlations between appetitive traits (n=708) ..................................................... 117
Table 5.6 Internal reliability of appetitive trait scales for the AEBQ ..................................... 119
Chapter 6
Table 6.1 Descriptive statistics of adult samples used to carry out PCA (Sample 1), and CFA
and re-test sample (Sample 2) ................................................................................................ 132
Table 6.2 AEBQ components with indicator names and numbers (Model 1) ....................... 135
Table 6.3 AEBQ components with indicator names and numbers (Model 2) ....................... 140
Table 6.4 Model fit indices for Models 1 and 2 in CFA of the AEBQ ...................................... 142
Table 6.5 Internal and test-retest reliability measures for the AEBQ in an adult sample ..... 143
12
List of Tables
Table 6.6 Descriptive statistics of appetitive trait mean scores (n = 954) ............................. 143
Table 6.7 Correlations between appetitive traits (n=954) ..................................................... 144
Table 6.8 Correlations between BMI and appetitive traits in the total adult sample (n=940)144
Table 6.9 Multiple linear regression for BMI and appetitive traits (n=940) .......................... 145
Chapter 7
Table 7.1 Six steps in public health intervention development ............................................. 152
Table 7.2 Appetitive trait weight management tips for, ‘food responsiveness’, emotional
over-eating’, ‘satiety responsiveness’, and ‘fast eating’......................................................... 157
Table 7.3 Demographic characteristics and initial BMI of participants (n=53) ...................... 166
Table 7.4 Number of replies to weekly follow-up questionnaires by weight change category
in completers .......................................................................................................................... 168
Table 7.5 Number of participants that reported following each tip...................................... 170
Table 7.6 Perceived helpfulness of the tips ........................................................................... 172
Table 7.7 Reported goal setting by participants receiving tips .............................................. 172
Table 7.8 Barriers to following appetitive trait tips ............................................................... 174
13
List of Figures
List of Figures
Chapter 2
Figure 2.1 The psychobiological network of appetite regulation ............................................ 33
Figure 2.2 Change in the distribution of BMI between 1976–1980 and 1999–2004, for
adults aged 20-74 years in the United States of America (USA)............................................... 44
Figure 2.3 Relationships of appetitive traits to the genetically determined susceptibility of

the environment ....................................................................................................................... 45
Chapter 4
Figure 4.1 Flow chart of studies in review (based on PRISMA 2009 flow diagram). ............... 62
Chapter 6
Figure 6.1 CFA model for a 35 item, 7-factor AEBQ (Model 1) .............................................. 134
Figure 6.2 CFA model for a 35 item, 7-factor AEBQ with covariances between errors (Model
1) ............................................................................................................................................. 137
Figure 6.3 CFA for model a 35 item, 8-factor AEBQ, with ‘hunger’ and ‘food responsiveness’
separated (Model 2) ............................................................................................................... 139
Chapter 7
Figure 7.1 Flow chart of participants ..................................................................................... 167
14
List of Appendices
List of Appendices
Chapter 4
Appendix 4.1 Complete electronic search strategy for the systematic review in Chapter 2 253
Appendix 4.2 Standards for educational and psychological testing – does not include all of
the standards relevant to the development of high stake instrument, fairness in testing and
issues of cultural sensitivity .................................................................................................... 254
Appendix 4.3 Validity and reliability measures of questionnaires from the systematic
review in Chapter 2 – Extended version ................................................................................. 256
Chapter 5
Appendix 5.1 Accepted manuscript ........................................................................................ 264
Appendix 5.2 Conference presentations ................................................................................ 271
Appendix 5.3 The Child Eating Behaviour Questionnaire (CEBQ) ........................................... 272
Appendix 5.4 Weight Concern ‘Shape-Up’ manual ‘hunger’ or ‘craving’ questions.............. 275
Appendix 5.5 ‘Adult Eating Behaviour Questionnaire’ used for piloting (49-item) in Study 2 276
Appendix 5.6 Illustrative example of themes obtained from adult piloting (Study 2) .......... 284
Appendix 5.7 Adult Eating Behaviour Questionnaire (47-items) (Study 2, Sample 1)........... 285
Appendix 5.8 Notification of Ethical Approval ....................................................................... 294
Appendix 5.9 AEBQ items compared to the original CEBQ items .......................................... 296
Appendix 5.10 Adult Eating Behaviour Questionnaire with scoring system ......................... 298
Chapter 6
Appendix 6.1 Relevant parts of the Self-Regulation of Eating Behaviour Questionnaire

which contained the Adult Eating Behaviour Questionnaire items (Study 3, Sample 2) ....... 302
Appendix 6.2 Ethical approval ............................................................................................... 311
15
List of Appendices
Chapter 7
Appendix 7.1 Feasibility questions obtained from an on-line panel who completed the
AEBQ (n=711)a......................................................................................................................... 313
Appendix 7.2 Individualised appetitive trait tip feedback for high 'food responsiveness' and
high ‘emotional over-eating’, low 'satiety responsiveness' and fast eating (low 'slowness in
eating') .................................................................................................................................... 314
Appendix 7.3 Individualised appetitive trait tip feedback for high 'food responsiveness', low
'satiety responsiveness' and fast eating (low 'slowness in eating') ........................................ 321
Appendix 7.4. Initial Survey Monkey questionnaire sent to members of the ‘Big Panel’ in
Study 4 .................................................................................................................................... 327
Appendix 7.5 Survey Monkey questionnaire sent to potential ‘Big Panel’ members after a
previous first contact, to assess inclusion criteria for participation in Study 4. ..................... 334
Appendix 7.6 Weekly follow-up questionnaires (WFQ) (Study 4) ......................................... 339
Appendix 7.7 Consent form for participation in the Appetitive Trait Tailored Intervention
(ATTI) (Study 4) ....................................................................................................................... 345
Appendix 7.8 Ethical approval ............................................................................................... 346
Appendix 7.9 Number of tips given to each participant for high ‘food responsiveness’ and
‘emotional over-eating’ and low ‘satiety responsiveness’ and ‘slowness in eating’ scores
(n=53) ...................................................................................................................................... 348
Appendix 7.10 Initial BMI, final BMI, initial weight, final weight and change in weight for
each participant in the appetitive trait intervention (n=53). ................................................. 350
Chapter 8
Appendix 8.1 Semi-structured interview guide for partitipants of the ‘Appetitive Trait
Tailored Intervention’ (ATTI) (Study 5) ................................................................................... 352
Appendix 8.2 Consolidated criteria for reporting qualitative research (COREQ); 32 item
checklist .................................................................................................................................. 357
Appendix 8.3 Appetitive traits tips given, Initial BMI, final BMI, initial weight, final weight
and change in weight for each participant interviewed after the appetitive trait intervention
(n=21) ...................................................................................................................................... 359
16
Abbreviations
Abbreviations
ASBP The American Society of Bariatric Physicians

AEBQ Adult Eating Behaviour Questionnaire
AIC Akaike’s Information Criteria
ATTI Appetitive Trait Tailored Intervention
BEBQ Baby Eating Behaviour Questionnaire
BIC Bayesian Information Criterion
BMI Body Mass Index
BMI-SDS BMI standard deviation scores
BST Behavioural Susceptibility Theory of obesity
CCK Cholecystokinin
CEBQ Child Eating Behaviour Questionnaire
CFA Confirmatory Factor Analysis
CNS Central Nervous System
CoEQ Control of Eating Questionnaire
CFI Comparative Fit Index
CR Cognitive restraint
CVD Cardiovascular disease
DEBQ Dutch Eating Behaviour Questionnaire
D Disinhibition
DM2 Type 2 diabetes
ecSI ecSatter Inventory
EMAQ Emotional Appetite Questionnaire
EE Emotional Eating
EES Emotional Eating Scale
EAH Eating in the absence of hunger
EF Enjoyment of food
EMAQ Emotional Appetite Questionnaire
EOE Emotional over-eating
EUE Emotional under-eating
FF Food fussiness
FNS Food Neophobia Scale
FR Food responsiveness
17
Abbreviations
EPI-C Eating Pattern Inventory for Children

FTO Fat mass and obesity-associated gene
fMRI Functional Magnetic Resonance Imaging
FCQ-S State Food-Cravings Questionnaires
FCQ-T Trait Food-Cravings Questionnaires
GWAS Genome-wide association studies
H Hunger
HFR Hunger-Food responsiveness
HSS Hunger Sensitivity Scale
IES Intuitive Eating Scale
ICC Intra-class correlation coefficients
KMO Kaiser-Meyer-Olkin
MRI Magnetic Resonance Imaging
MES Mindful Eating Scale
MFES Motivation for Eating Scale
NWCR National Weight Control Registry
NFI Normed Fixed Index
OTS Over-eating Tension Scales
PCA Principal Component Analysis
PEMS Palatable Eating Motives Scale
PFS Power of Food Scale
PET Positron Emission Tomography
RMSEA Root Mean Square Error of Approximation
RS Restraint Scale
SE Slowness in eating
SNPs Single nucleotide polymorphisms
SR Satiety responsiveness
SSBs Sugar-sweetened beverages
TEDS Twins Early Development Study
TFEQ Three Factor Eating Questionnaire
WFQ Weekly follow-up questionnaire
WHO World Health Organisation
18
Chapter 1: Weight management
Chapter 1. Weight management
1.1 The need for weight management interventions
The worldwide problem of obesity is acknowledged as having far-reaching consequences

for health and wellbeing (Kleinert & Horton, 2015). Obesity can be defined as “abnormal or
excessive fat accumulation that may impair health” (WHO, 2015). Obesity is most often
classified using a proxy measure of weight adjusted for height known as Body Mass Index
(BMI). BMI is calculated by dividing a person’s body weight in kilograms by height in metres
squared (kg/m2). In adults, overweight is defined as a BMI between 25 and 29.9 and obesity as
a BMI equal to or above 30. There are different classes of obesity defined by the extent to which
an individual’s BMI is above 30 (Class I: BMI of 30 to 34.9; Class II: BMI of 35 to 39.9; and Class III:
BMI of 40 and above) (WHO, 2000, 2014). Normal weight is defined as a BMI between 18.5 and
24.9, and underweight, equal to or below 18.5. BMI is widely used in obesity research, as it is
simple to measure and has predictive validity for a range of health outcomes (Frühbeck et
al., 2013). BMI and weight will be used interchangeably throughout this thesis.
Excess body fat is a risk factor for the development of a range of chronic diseases, such as
Type 2 diabetes (DM2), cardiovascular disease (CVD) and cancer (Frühbeck et al., 2013).
Obesity increases the risk of metabolic syndrome, DM2 and hypertension, which together
substantially increase the risk of CVD and stroke (Brown & Kuk, 2015; Shamseddeen, Zelada
Getty, Hamdallah, & Ali, 2011). Overweight and obesity carry stereotypes of laziness and a
lack of self-discipline (Puhl & Heuer, 2009). This public prejudice makes obesity highly
stigmatized (Wee, Davis, Huskey, Jones, & Hamel, 2013) and weight-based discrimination is
common (Higgs & Thomas, 2016; Stok, Verkooijen, de Ridder, de Wit, & de Vet, 2014).
Another potential consequence of obesity is low self-esteem and poor self-image, as well as
disordered eating issues (Cruwys, Leverington, & Sheldon, 2015). Obesity also has great
economic costs, because the health consequences place a burden on health care systems
and result in losses to both productivity and disability-free life expectancy (Roberto et al.,
2015).
There is seemingly no relief in sight, given the continuously rising prevalence of obesity
around the globe. Between 1980 and 2013, the global proportion of overweight and obese
adults increased from 28.8% to 36.9% in men and from 29.8% to 38.0% in women; the
19
combined increase in prevalence of overweight and obesity was 27.5% (Ng et al., 2014). In
the United Kingdom (UK) in 2013, 41% of men were overweight and 26% were obese, while
33% of women were overweight and 26% were obese (Health and Social Care Information
Centre, 2014). Although overweight and obese individuals frequently report both a desire
to lose weight (Yaemsiri, Slining, & Agarwal, 2010) and actual attempts at losing weight
(Nicklas, Huskey, Davis, & Wee, 2012; Wardle & Johnson, 2002), only approximately 20%
manage to achieve significant weight loss and maintain it over the long-term (Wing &
Phelan, 2005).
In summary, given the physical and psychological health consequence of obesity, it’s
increasing prevalence and the difficulties of weight loss, weight management has become a
top priority for public health (NICE Clinical Guideline 189, 2014; The Obesity Society and
American College of Cardiology/American Heart Association Task Force on Practice
Guidelines. Based on a systematic review from The Obesity Expert Panel, 2014). Weight
management refers to both the prevention of weight gain and weight loss in order to
achieve and maintain a healthy weight. The ultimate aim of this thesis is to develop a novel
weight management intervention for the purposes of weight loss in individuals that are
already either overweight or obese.
1.2 Approaches to weight management
At a basic level, obesity results from a sustained positive energy balance (i.e. energy intake
exceeding energy expenditure). However the overall picture is far more complex, and
weight is known to be influenced by a range of factors, both at an environmental and
individual level (Vandenbroeck, Goossens, & Clemens, 2007). A variety of options therefore
exist to reduce obesity, which target some of these different causes. The majority of
approaches fall into four main categories: structural which include policy approaches (such
as taxation), or changes to the environment which influence active living; pharmacological
strategies; surgical interventions; and lifestyle interventions and counselling (which include
adherence to dietary changes and increasing physical activity through the use of
behavioural strategies) (Brownell & Roberto, 2015; NICE Clinical Guideline 189, 2014).
Structural approaches to obesity require a multi-level systems approach (Malik, Willett, &
Hu, 2012), and address the environmental and socio-cultural factors that contribute to
obesity. On the other hand, while lifestyle, pharmacological and surgical interventions may
ultimately impact at a population level, they are focused more on the individual (Malik &
20
Hu, 2007; Malik et al., 2012), and predominantly address the behavioural factors which
contribute to obesity development.
1.2.1 Structural approaches
We are currently living in an ‘obesogenic’ environment that promotes an overconsumption

of energy and reduces total energy expenditure (Swinburn, Sacks, & Ravussin, 2009).
Availability and access to convenient, inexpensive, palatable, energy-dense foods in large
portions is widespread and has increased energy consumption (French, 2003; French et al.,
2014; Kral, Roe, & Rolls, 2004; Piernas, Ng, & Popkin, 2013; Rolls, Roe, Kral, Meengs, &
Wall, 2004). A dependency on vehicles for transport, shifts in leisure time to include
greater amounts of screen time, and the move from manual labour to automation have all
reduced energy expenditure (Chaput, Klingenberg, Astrup, & Sjödin, 2011; Church et al.,
2011; Goodman, 2013). Various socio-economic and socio-cultural factors such as
education and time constraints affect energy intake and energy expenditure, with busy
families relying on fast-foods or restaurants for food consumption (Patrick & Nicklas, 2005;
Vandenbroeck, Goossens, & Clemens, 2007). Thus, a structural approach to obesity
management targets these environmental risk factors.
Applying structural approaches to manage the global obesity epidemic requires input from
elected leaders, government agencies and non-governmental organisations, industry,
health-care systems, schools, urban planners, agricultural and service sectors, and global
institutions such as the World Bank or the WHO, which can impact on the regulations of
sustained population-wide interventions and policy recommendations (Brownell & Roberto,
2015). One example of this is the implementation of the excise tax on sugar-sweetened
beverages (SSBs) in Mexico in January 2014. Here, purchases of taxed beverages decreased
by an average of 6% (-12mL/capita/day) and non-taxed beverages increased by 4%
(36mL/capita/day) one year after tax implementation (Colchero, Popkin, Rivera, & Ng,
2016). In the UK, proposals to tax SSBs have come under considerable criticism and it has
been argued that it will have minimal impact as consumption of sugary drinks only account
for approximately 20% of sugar intake in UK children (Neville & Pickard, 2016).
Other environmental interventions to tackle obesity include the development of cycle lanes
to promote increased physical activity and social capital (Torres, Sarmiento, Stauber, &
Zarama, 2013), the inclusion of family fitness zones in urban public parks (Cohen, Marsh,
Williamson, Golinelli, & McKenzie, 2012), and the structuring of urban planning codes that
21
impact on physical activity by increasing walking in residential environments (Christian et

al., 2013). However, assessing the actual use of structural or environmental approaches to
obesity management is subject to the limited evidence of the impact these changes have on
a community. Results are typically available only from cross-sectional studies which lack
control groups (Torres et al., 2013). Also, follow-up data tend to be collected at different
times of the year, which for example, limits the interpretation of park use from one follow-
up to the next (Cohen et al., 2012). Stronger evaluation of these interventions is needed to
encourage the use of such approaches for the management of obesity (Christian et al.,
2013; Cohen et al., 2012; Torres et al., 2013). Structural approaches to weight
management are challenging as their implementation requires the cooperation of many
parties, they necessitate high-level input, and have substantial financial costs (Brownell &
Roberto, 2015).
1.2.2 Pharmacological interventions
Pharmacotherapy is a treatment option for weight management targeted at the individual-

level. Current recommendations are that pharmacological treatment should only be
considered once dietary, physical activity and behavioural approaches have been
exhausted; or for those patients who cannot reach their target weight, or have reached a
plateau on dietary, physical activity and behavioural modifications (NICE Clinical Guideline
189, 2014). Preferably, pharmacological treatment of obesity should be used as an adjunct
to comprehensive lifestyles changes (The Obesity Society and American College of
Cardiology/American Heart Association Task Force on Practice Guidelines. Based on a
systematic review from The Obesity Expert Panel, 2014). Behavioural changes will be
discussed in Section 1.2.4.
Approved obesity medications work through effects on appetite, acting as anorexigenics on

satiety centres in the hypothalamic and limbic regions of the brain (e.g. sibutramine,
phentermine, fluoxetine, bupropion, topiramate); or as lipase inhibitors (e.g. orlistat),
reducing the absorption of dietary fat in the gastrointestinal tract (Domecq et al., 2015;
Kushner, 2014). With respect to the effectiveness of pharmacological interventions,
placebo-controlled trials have shown an average weight loss of approximately 2.7kg to
3.19kg with Orlistat (Kushner, 2014), the only obesity medication currently available in the
UK (NHS Choices. Your health, 2016). However, there are also a number of side effects
associated with the use of these drugs. In anorexigenics, these include restlessness,
insomnia, dry mouth, constipation and increased heart rate; for lipase inhibitors side
22
effects might include steatorrhea, bloating and abdominal distension, as well as anal
leakage (Kushner, 2014). The rationale for their use, as well as possible side effects should
therefore be discussed between the patient and the health professional team treating the
patient (Seger et al., 2013).
1.2.3 Surgical interventions
A more intensive approach to managing obesity at the individual-level is surgical treatment.

Long-term outcomes of bariatric surgery are better than for lifestyle changes or lifestyle
change and pharmacotherapy (Nguyen et al., 2012). Bariatric surgeries typically reduce
BMI by 12 to 17 points five years post-surgery (Chang et al., 2014). The most commonly
performed procedures are laparoscopic adjustable gastric banding (making up around a
third of all surgeries), laparoscopic sleeve gastrectomy, and Roux-en-Y gastric bypass
(comprising around half of all bariatric surgeries) (Kushner, 2014). Laparoscopic adjustable
gastric banding and sleeve gastrectomy aim to induce weight loss through restricting food
intake, whereas biliopancreatic diversion (gastric bypass) prevents food absorption, and
techniques such as a gastric bypass provide a combination of restriction and malabsorption
techniques (Seger et al., 2013). A recent meta-analysis of 164 studies, from 2003 to 2012,
revealed that gastric bypass was more effective than adjustable gastric banding but was
associated with more complications. Sleeve gastrectomy appears to be more effective than
gastric bypass in producing weight loss, and both were more effective than adjustable
gastric banding (Chang et al., 2014).
However not all patients with obesity are eligible for surgery. The American Society of
Bariatric Physicians (ASBP) published an algorithm for the decision to surgically treat
patients, which specified patients should only be considered if they have a BMI ≥ 30 with
one or more adverse health consequences or a BMI ≥ 40 with or without adverse health
consequences (Seger et al., 2013). Currently evidence is insufficient to recommend
bariatric surgery for individuals with a BMI < 35 and no co-morbidities (Published by the
Obesity Society and American College of Cardiology/American Heart Association Task Force
on Practice Guidelines. Based on a systematic review from The Obesity Expert Panel, 2014).
Furthermore, many patients do not qualify for surgery, due to the risk of complications
during and after surgery (Nguyen et al., 2012). Complications, occur in 10% to 17% of
patients, with repeat operations in approximately 7% (Chang et al., 2014; Nguyen et al.,
2012). Complications result from an altered anatomy, including malabsorption issues and
the effects of dietary changes due to reduced gastric size (Kushner, 2014). These risks
23
should be considered for each patient, and a team of health professionals should be
involved in assessing every individual case. However, for those patients that do qualify,
bariatric surgery leads to weight loss and improvements in comorbidities post-surgery,
attributed to changes in physiological responses to gut hormones and adipose tissue
metabolism (Kushner, 2014). As with pharmacological treatments, surgery is ineffective
without corresponding lifestyle changes, including dietary and physical activity
modifications. Although surgery can make it easier for patients to make these
modifications, patients may also require psychological interventions alongside their surgery
to manage the drivers behind their eating behaviours (The Obesity Society and American
College of Cardiology/American Heart Association Task Force on Practice Guidelines. Based
on a systematic review from The Obesity Expert Panel, 2014).
1.2.4 Lifestyle interventions
Even though global efforts exist to try and make changes to the environment, and
pharmacotherapy and surgery are becoming more common, the safest and less invasive
approach to manage obesity is to try and change people’s behaviour through lifestyle
interventions (Dansinger, Gleason, Griffith, Selker, & Schaefer, 2005; Franz et al., 2007;
Truby et al., 2006). Furthermore, both pharmacological and surgical approaches are only
effective in combination with lifestyle change, particularly over the long term.
Lifestyle interventions include dietary advice and recommendations for increased physical
activity and decreased sedentary behaviour, normally in combination with behavioural
counselling to facilitate weight reduction behaviours (Göhner, Schlatterer, Frey, Berg, &
Fuchs, 2012; Rapoport, Clark, & Wardle, 2000). Most guidelines recommend such
programmes aim for weight losses of 5% to 10% of body weight for adults, as such losses
have been associated with health improvements (Kirk, Tytus, Tsuyuki, & Sharma, 2012;
National Institute of Health, 1998; NICE Clinical Guideline 189, 2014; Willett, Dietz, &
Colditz, 1999). However, recent guidance suggests sustained weight losses of 3% to 5% can
produce clinically meaningful health benefits and should be encouraged (Published by The
Obesity Society and American College of Cardiology/American Heart Association Task Force
on Practice Guidelines. Based on a systematic review from the The Obesity Expert Panel,
2014).
An array of dietary recommendations have been used for weight loss including: low
carbohydrate diets (Naude et al., 2014); reduced intake of sugars (Te Morenga, Mallard, &
24
Mann, 2012); reduced intake of fat (Hooper et al., 2015); higher intake of ‘healthy’ fat as in
‘Mediterranean diets’ (Nordmann et al., 2011); diets high in protein such as the ‘Atkins
diet’; and low carbohydrate diets which recommend the intake of high quantities of protein
(Dansinger, Gleason, Griffith, Selker, & Schaefer, 2005; Gardner et al., 2007; Westerterp-
Plantenga, Lemmens, & Westerterp, 2012). Weight loss can also be achieved by the explicit
manipulation of the energy content of the diet, as opposed to dietary manipulation only
(Kirk et al., 2012). Very-low calorie diets (VLCD) (800 kilocalories per day or less), have been
found to induce greater short-term weight loss than low calorie diets (Tsai & Wadden,
2006). Commercial weight loss diets popularised by the growing slimming industry, such as
Nutrisystem (a diet delivery program which includes low calorie meal replacement
delivered to your door and promotes exercise and self-monitoring) have shown better short
term weight outcomes than educational control or behavioural counselling (Gudzune et al.,
2015). Meal replacement diets such as Slim-Fast have also had positive results in the short-
term (Truby et al., 2006).
Evidence points towards structured, individualised nutritional counselling and personal

support as being more important for success than the macronutrient content of the diet
(Johnston et al., 2014; Kirk, Penney, McHugh, & Sharma, 2012). Although a wide range of
dietary recommendations have been used for weight loss, current research has not shown
convincingly that one type of diet is more successful than another. For example, in a
randomised non-blinded controlled trial known as the BBC “diet trials”, the effectiveness of
the Atkins diet (a self-monitored low carbohydrate eating plan), Weight Watchers (an
energy controlled diet with weekly group meetings), Slim-Fast plan (a meal replacement
program) and Rosemary Conley (a low calorie diet with a weekly group exercise session)
programs were compared in a group of adults over a six-month period. An average weight
loss of 5.9 kg and an average fat loss of 4.4 kg was achieved based on an intention to treat
analysis, and no diet had greater success at achieving weight loss than the other (Truby et
al., 2006). A recent meta-analysis and meta-regression of 48 randomised trials of diet
classes and programs similarly found that no diet was better at achieving weight change at
6 or 12 months from baseline (Johnston et al., 2014). In this review, weight loss was
achieved with either low carbohydrate or low fat diets and individual differences between
weight loss diets was minimal; the authors suggested that individuals should choose the
diet they prefer and success is better predicted by how well individuals adhere to a diet
(Johnston et al., 2014).
25
There is some evidence to suggest long term outcomes can be improved if diet and physical
activity are combined (Johns, Hartmann-Boyce, Jebb, & Aveyard, 2014). A systematic
review of eight studies which combined either diet or physical activity interventions, found
these to be more effective at achieving weight loss compared to just exercise in both the
short term (-5.33 kg, 95%CI -7.61 to -3.04) and the long term (-6.29 kg, 95%CI -7.33 to -5.25)
(Johns et al., 2014). Although increasing physical activity in itself produces only modest
weight loss, it protects against the loss of lean tissue and has significant independent
benefits for cardiovascular health (The Obesity Society and American College of
Cardiology/American Heart Association Task Force on Practice Guidelines. Based on a
systematic review from The Obesity Expert Panel, 2014).
Behaviour change techniques are also important. For example, stimulus control1
techniques (Hartmann-Boyce, Aveyard, Koshiaris, & Jebb, 2016; Wardle & Johnson, 2015;
Wardle et al., 2013; Wardle, Liao, et al., 2001), self-monitoring of diet and physical activity
(Burke, Wang, & Sevick, 2011; Burke, Conroy, et al., 2011), relapse prevention strategies
(Strayhorn, 2002), and goal setting (Dalle Grave, Centis, Marzocchi, El Ghoch, & Marchesini,
2013; Hartmann-Boyce et al., 2016). A 2009 meta-regression of 101 studies reporting 122
evaluations of physical activity and healthy eating, found that those interventions which
include self-monitoring, with any other technique derived from the “Control theory”2
(Carver & Scheier, 1982), such as specific goal setting, review of goal settings, feedback of
performance or intention formation, are most likely to be effective for behaviour change
(Michie, Abraham, Whittington, McAteer, & Gupta, 2009). Interventions which target these
as well as dietary intake and physical activity levels (as a three-component intervention)
appear to be more successful compared to those interventions targeting diet and exercise
either in conjunction or alone (Kirk et al., 2012). Current guidance in the UK therefore
advocates this three pronged approach to lifestyle-based obesity management (NICE
Clinical Guideline 189, 2014). However, although the use of such techniques can enhance
1
Stimulus control refers to the reduction of exposure to eating cues, which helps limit the number of
occasions eating can occur (Wardle & Johnson, 2015; Wardle et al., 2013).
2
The “Control theory” refers to a model of self-regulation, which is presented as a feedback-loop,
where a person’s perception of their current state is compared against a goal state (Carver &
Scheier, 1982; Michie, West, et al., 2014).
26
weight loss in the short term, studies with long term outcomes suggest weight is still
gradually regained once treatment stops (Wardle & Johnson, 2015), and there is still a great
deal of individual variability in the success of these interventions (Stubbs et al., 2011).
1.3 Individual factors influencing the likelihood of successful weight

management
Given the variation in success with weight management, a number of studies have sought
to explore individual factors that might influence the likelihood of successful weight
management. Reviews have found some evidence for demographic predictors of weight
loss (Stubbs et al., 2011; Teixeira et al., 2010; Teixeira, Going, Sardinha, & Lohman, 2005),
with age and gender consistently related to success. Men consistently tend to lose more
weight than women (Stubbs et al., 2011), and women typically have higher attrition rates
than men (Fabricatore et al., 2009). With respect to age, although obesity is higher among
older than younger adults (Health and Social Care Information Centre, 2014), being of
younger age predicts greater attrition from weight loss interventions (Fabricatore et al.,
2009). However, attrition is also influenced by initial weight loss (i.e. higher initial weight
loss is associated with less attrition), which is itself correlated to attendance at a weight
loss program (Stubbs et al., 2011). Being older could also be beneficial, as it may bring the
benefit of greater awareness of dealing with relapses and of developing stable eating and
physical activity patterns. However, older age may also hinder potential weight loss due to
the physiological effects of numerous previous weight loss attempts (see effects below)
(Stubbs et al., 2011).
Prior weight loss attempts and participation in weight loss programs appear to predict
future weight loss failure (Stubbs et al., 2011). For example, Teixeira et al. (2004) found, in
158 overweight and obese middle-aged healthy women, that a history of weight loss
attempts was independently associated with non-completion in a behavioural weight
management program. However, these results cannot be generalised to other ages or to
men. Similarly, in the Australian Longitudinal Study of Women’s Health, 79.9% of women
reported using at least one weight loss strategy over the course of a year. The participants
described the strategies used and the number of times they had lost on purpose more than
5kg. These strategies were categorised into four clusters; ‘dieting – those who used a
variety of strategies to control their weight’ (39.7%), ‘healthy living – eat less move more’
(30.2%), ‘do nothing’ (20%), and ‘perpetual dieters – used all strategies, including unhealthy
27
behaviours’ (10.7%). Results showed that, despite most women trying to control their
weight, they gained an average of 700g per year over the nine-year period that weight
change data was assessed and the ‘perpetual dieters’ gained significant more weight (210g)
than the ‘do nothing’ group (p<0.01) (Madigan, Daley, Kabir, Aveyard, & Brown, 2015).
The negative impact of prior weight loss attempts on weight management success may be a
consequence of the physiological changes induced by weight loss. Weight loss is associated
with decreases in metabolic rate, reduced total energy expenditure greater than the weight
loss achieved, and changes in hormone profile which regulate appetite (Chapter 2)
(Rosenbaum, Hirsch, Gallagher, & Leibel, 2008; Sumithran & Proietto, 2013). Reductions in
circulating levels of leptin3, cholecystokinin4, insulin5 and other hormones involved in
appetite regulation accompany weight loss, and these changes do not appear to be
transient (Sumithran & Proietto, 2013). Therefore, previous weight loss attempts can
become negative weight loss predictors through diminishing total energy expenditure and
metabolic rate, and a lower resting metabolic rate is associated with less weight loss
(Stubbs et al., 2011). Additionally, a higher initial weight or BMI predicts greater weight loss
(Stubbs et al., 2011; Teixeira et al., 2004, 2010).
On the other hand, results from an exhaustive review suggest that baseline measurements
of psychosocial variables such as mood, depression and personality disorders are not
predictive of treatment outcomes (Teixeira et al., 2005). Evidence for other psychological
pre-treatment predictors of weight loss, such as self-esteem, body image and weight-
related quality of life, have also been mixed. Authors have suggested that this might be
due to measurement issues given the diversity of different assessment constructs6 or scales
found in different questionnaires used to measure these variables (Teixeira et al., 2005).
Also, the lack of associations found between these measures and weight, may be the
3
Leptin is an appetite suppressant, made in fat tissue.
4
Cholecystokinin is a hormone released in the gastrointestinal tract which stimulates fat and protein
digestion.
5
Insulin is a hormone produced by the pancreas which regulates blood glucose levels.
6
Throughout this thesis the terms ‘constructs’, ‘factors’, ‘scales’, ‘sub-scales’ or ‘dimensions’ are
used interchangeably to refer to the grouping of items which describe a certain type of trait,
measured by a questionnaire.
28
results of confounding, with other factors difficult to measure (e.g. personality, upbringing),
causing mixed results or none at all to be obtained (Stubbs et al., 2011).
Similar results have been observed for eating behaviour variables such as, binge eating,
‘disinhibition’ and ‘restrained eating’, which have shown very few or no associations with
weight changes during treatment (Teixeira et al., 2005). Interest in measuring ‘restraint’
started after the proposal of the “Externality” theory (Schachter, 1968), which suggests that
tendencies toward over-eating, trigger an individual’s need to restrict their food intake,
described as the “Restraint” theory of obesity (Herman & Polivy, 1975; Polivy, Herman,
Younger, & Erskine, 1979). However, ‘restraint’ has also been linked with inducing counter-
regulatory responses which result in binge-like or disinhibited eating patterns (Johnson et
al., 2012; Stunkard & Messick, 1985), and so these variables could confound one another
depending on the measures used to assess them. ‘Disinhibition’ is usually considered to be
highly variable between individuals and associated with factors such as ‘restraint’ and
weight gain. In general, binge eating and ‘disinhibition’ have been negatively associated
with weight control (Polivy & Herman, 1976a, 1976b). However, although initial studies
showed that ‘restraint’ was associated with dieting failure (Herman & Polivy, 1975; Polivy et
al., 1979), later evidence has suggested that ‘restraint’ could be associated with better
weight loss outcomes (Johnson et al., 2012). Weight loss maintenance studies have also
shown that higher dietary ‘restraint’ scores are associated with greater weight maintenance
over a 10 year period (Thomas et al., 2014). This has led to the suggestion that measures of
‘restraint’ may be capturing aspects of ‘self-regulation’ or ‘self-control’ (Johnson et al.,
2012), factors which may promote successful weight management. Another behavioural
trait which has been positively correlated to weight loss success is slow rate of eating
(Stubbs et al., 2011). However, more research is needed on this and the potential for other
eating behaviours to predict successful weight loss.
1.4 Tailoring weight management interventions to improve outcomes
To date there appear to be few consistent predictors of weight loss success, with the
exception of prior weight loss attempts. This may reflect the fact that different individuals
have varying success dependent on the programme they are following, and reviews have
highlighted a need to conduct further research into individualised approaches to weight
management (Stubbs et al., 2011; Teixeira et al., 2005). As previously mentioned in Section
1.2.4, results from behavioural weight loss treatments, have shown substantial individual
29
variability in weight loss achieved (Dansinger et al., 2005; Franz et al., 2007; Hartmann-
Boyce, Johns, Jebb, Summerbell, & Aveyard, 2014; Truby et al., 2006, 2008) and there has
therefore been increasing interest in the benefits of tailoring weight management
treatments to an individual’s biological and psychological characteristics in order to
promote better weight loss (Almirall, Nahum-Shani, Sherwood, & Murphy, 2014; Celis-
Morales, Lara, & Mathers, 2015). Improving understanding of the success of this type of
personalised medicine could facilitate recommendations for weight management (Finer,
2015; Gardner, 2012).
A recent review and meta-analysis of 39 trials that compared self-help interventions with
each other or with minimal control for weight loss in overweight and obese adults,
suggested that tailoring appeared to increase weight loss when compared to non-tailored
approaches (Hartmann-Boyce, Jebb, Fletcher, & Aveyard, 2015). Tailoring of diets was
based on the patients’ baseline information, such as personalised weight loss goals based
on initial weight, height and waist circumference (WC), or on progress reports generating
automated personalised feedback based on diary entries (Carter, Burley, Nykjaer, & Cade,
2013; Collins, Morgan, Hutchesson, & Callister, 2013). Tailored nutrition education based
on dietary intake, food purchases and anthropometric measures have also been found to
be useful at improving diets over the long term (Kirk et al., 2012; Pagoto & Appelhans,
2013). Tailoring has been used for giving personalized information on obesity risk in
vignette studies (Frosch, Mello, & Lerman, 2005). Those participants assigned to an
increased obesity risk vignette condition, indicated they had a greater intention of changing
their behaviours than those assigned to lower risk conditions. However, overall, there are
only a small number of studies to date in this area, and tailoring has been focused mainly
on baseline physiological conditions, such as the presence of DM2, or cardiovascular risk
factors (The Obesity Society and American College of Cardiology/American Heart
Association Task Force on Practice Guidelines. Based on a systematic review from The
Obesity Expert Panel, 2014). No scientific study has explored tailoring based on an
individual’s eating behaviours.
30
A potentially novel approach to improve weight management outcomes could be using

information about individuals’ eating behaviours or appetitive traits7 to inform weight loss
recommendations. Given that certain pharmacological and surgical approaches to obesity
treatment involve suppressing appetitive pathways, developing behavioural strategies to
help an individual manage their appetitive traits might be a less invasive way to help
overweight and obese individuals to lose weight. Appetitive traits and their influence on
weight will be addressed in the following Chapter 2.
1.5 Summary of the findings
Effective weight loss strategies are much needed in the prevailing global obesity landscape.
The physical and psychological health consequences of obesity, as well as its increasing
prevalence, mean that reducing obesity through weight management has become a
priority. Different approaches to weight management have been developed, ranging from
environmental and policy change to the more individual pharmacological, surgical and
lifestyle approaches. There is little evidence that individual factors affect how successful
any given person is at achieving weight loss, but there may be merit in matching individuals
to different treatments better suited to their biological and psychological profiles. There is
emerging evidence for the use of such factors to tailor weight management interventions,
and tailoring based on eating behaviour phenotypes or appetitive traits could be a helpful
approach. Appetitive traits and their influence on weight will therefore be discussed in
greater depth in the following chapter.
7
Appetitive traits are a set of stable predispositions towards food (Carnell et al., 2013). The
relationship that they have to weight and weight management will be explained in detail in Chapter
2.
31
Chapter 2. Appetitive traits and weight
2.1 Introduction
To understand how appetitive traits might be used to tailor weight management advice to
individuals who are overweight or obese, it is important to first understand what appetitive
traits are and how they are associated with weight. The following chapter discusses
definitions of appetite and appetitive traits, how they can be measured and how they are
thought to relate to weight.
2.2 Appetite
Appetite can be defined as the process of food selection and intake, and its consumption in
appropriate amounts in relation to the maintenance of body weight (Blundell et al., 2009).
It is related to hunger and satiety, the psychological experiences that determine meal-by-
meal eating behaviour8 (Blundell et al., 2009). In states of overweight, obesity, and eating
disorders, appetite becomes deregulated. The expression of appetite is a complex web of
biopsychological aspects, postulated to be controlled by the balance between ‘homeostatic’
and ‘hedonic’ pathways. These factors have also been posited to play a key role in an
individual’s vulnerability to gain weight (Blundell et al., 2005).
The ‘homeostatic’ mechanisms are mediated by a need to maintain energy balance, for
example increasing motivation to eat after depletion of energy stores. There are two types
of ‘homeostatic’ signals: ‘tonic’ signals and ‘episodic’ signals. ‘Tonic’ signals, such as leptin
secretion, provide a message of hunger and are involved in more stable long term energy
reserves. ‘Episodic’ signals are mediated by peripheral satiety signals from the gut (e.g.
8
In the literature, eating behaviour is a term sometimes used to refer to appetitive traits. For the
purpose of this thesis I will make a distinction between these terms. ‘Eating behaviour’ will refer to a
broader spectrum of behaviours related to eating, including all the processes around the ingestion of
food, such as thoughts, actions, and intents; whereas ‘appetitive traits’ will refer to stable
predispositions towards food, which are thought to be susceptible to environmental interaction,
potentially predisposing an individual to weight gain.
32
cholecystokinin secretion) and are involved in short-term energy maintenance in response

to recent consumption. In contrast to these ‘homeostatic’ mechanisms, ‘hedonic’ control
of appetite is related to reward-based pathways and maintains a drive to eat. There is a
strong interplay between the ‘homeostatic’ and ‘hedonic’ pathways (termed the ‘satiety
cascade’) (Blundell et al., 2009; Harrold, Dovey, Blundell, & Halford, 2012) (Figure 2.1).
Neurotransmitter
and metabolic
interaction in the
brain
Physiological and
metabolic events
Psychological
and
behavioural
events –
‘Satiety
Cascade’
Source: (Harrold et al., 2012)

5-HT, serotonin; AA, amino acids; AgRP, agouti-related peptide; CART, cocaine and amphetamine-
regulated transcript; CCK, cholecystokinin; CRF, corticotrophin releasing factor; FFA, free fatty
acids; GI, gastrointestinal; GLP-1, glucagon-like peptide-1; GRP, gastric releasing peptide; MC,
melanocortin; NPY, neuropeptide Y; NST, nucleus tractus solitarius; PYY, peptide YY; T:LNAA,
tryptophan large neutral amino acid ratio.
Figure 2.1 The psychobiological network of appetite regulation
The psychobiological network of appetite regulation shown in Figure 2.1 represents the
different aspects of ‘homeostatic’ and ‘hedonic’ control of appetite, which will be briefly
explained in the following sections.
2.2.1 The homeostatic control of appetite
‘Homeostatic’ control of appetite originates pre-prandially (prior to meal ingestion).

Hunger signals are activated by the sight and smell of food, and they signal the brain via
33
cranial nerves to promote food intake. The proposed interactions between the ‘satiety
cascade’ (the physiological and metabolic events occurring during the satiation and satiety
processes) and the neurotransmitter and metabolic interactions which occur at the Central
Nervous System (CNS) level and control both ‘episodic’ and ‘tonic’ processes of appetite
control are seen in Figure 2.1. The short-term control of food intake is an ‘episodic’ activity
(i.e. a pattern of food episodes) that is primarily modulated by the gastrointestinal tract.
These temporary oscillations in energy influx are mainly caused by periodic meal intake.
The ‘tonic’ control of appetite responds to the depletion and repletion of energy stores,
representing the longer-term control of food intake. It is controlled by glucose metabolism
and fat storage in the adipose tissue, liver and pancreas via leptin, insulin and glucagon9.
Signals are released from storage tissues when energy is depleted and these signals then
stimulate energy intake. ‘Tonic’ signals characterise a more accurate representation of
energy needs than those driven by periodic ‘episodic’ signals (Harrold et al., 2012).
2.2.2 The hedonic control of appetite
The ‘hedonic’ control of appetite is mediated by reward, maintained by the drive to eat
highly palatable foods, which in turn stimulates over-consumption and maintains a system
based on pleasure (mainly sensory), compared to a biological need (as in ‘homeostatic’
control). It relates to the feeling of pleasure arising from or associated with eating.
However, it also involves other more complicated processes around the incentive value of
foods (‘liking’) and the reward value of foods (‘wanting’) – and is referred to as the major
driving force for food ingestion (Finlayson, King, & Blundell, 2007). Brain neurotransmitters
are involved with ‘hedonic’ processes, including glutamate, opioids, endocannabinoid and
dopamine, and they have been postulated to be involved in the ‘wanting’ and ‘liking’
control processes (Harrold et al., 2012).
2.2.3 Interplay of homeostatic and hedonic control of appetite
In the current obesogenic environment where highly palatable foods are freely available,
regulating pleasure and reward from the ‘hedonic’ pathways can lead to a diminished
9
Glucagon is a pancreatic hormone which promotes glycogen breakdown for glucose formation in
the liver.
34
control of the ‘homeostatic’ mechanisms, which in turn leads to hyperphagia (excessive

eating) and obesity. People may become over-responsive to the pleasure of eating and
homeostasis (physiological consequences of ingestion) is overridden by the ‘hedonic’
mechanisms. However, these systems do not operate independently; a careful balance is
required to maintain energy balance and thus control of body weight (Blundell & Finlayson,
2004).
The susceptibility to overeat is extremely variable among individuals. Because of the

complex nature of appetite, and the different aspects involved, a number of distinct
appetitive traits have been defined to facilitate its measurement.
2.3 Appetitive traits
Appetitive traits can be defined as stable predispositions towards food (Carnell, Benson,
Pryor, & Driggin, 2013). Appetitive traits fall into two broad groups: ‘food approach’ and
‘food avoidance’ traits (Viana, Sinde, & Saxton, 2008; Wardle & Gibson, 2001). ‘Food
approach’ (or eating-onset) traits such as ‘food responsiveness’, ‘external eating’,
‘disinhibition’, ‘emotional over-eating’, ‘enjoyment of food’, and ‘hunger’ are associated
with larger appetites or greater interest in food. ‘Food avoidance’ (or eating-offset) traits,
such as: ‘restraint’, ‘satiety responsiveness’, ‘emotional under-eating’, ‘food fussiness’ and
‘slowness in eating’, are associated with better appetitive control and/or a lower interest in
food (Table 2.1).
35
Table 2.1 Appetitive traits

Appetitive Traits Traits Description
Measures interest in food
‘Food responsiveness’
and drive to eat.
Increased consumption of
food due to the response
‘External eating’
to the sight and smell of
food.
The tendency to over-eat
in response to external
‘Disinhibition’ eating and/or eating in
response to negative
‘Food approach’ – food-onset moods.
traits Assesses tendencies to
‘Emotional over-eating’ over-eat in negative
emotional states.
Measures the level of
‘Enjoyment of food’ subjective pleasure
experienced from eating.
An individual’s perception
of their level of
‘Hunger’ motivation to eat and the
extent to which this elicits
food intake.
The tendency of some
persons to restrict their
‘Restraint’
food intake in order to
control their body weight.
Measures an individual’s
‘Satiety responsiveness’
fullness threshold.
Assesses tendencies to
‘Food avoidance’ – food-offset ‘Emotional under-eating’ under-eat in negative
traits emotional states.
Assesses pickiness with
regard to the type of food
‘Food fussiness’
an individual is willing to
eat.
Evaluates the pace at
‘Slowness in eating’ which an individual
consumes their food.
2.4 Measures of appetite and appetitive traits
Appetite and appetitive traits have been measured experimentally and psychometrically.
Experimental measures include both standard laboratory and neurological/neuroimaging
36
measures, whilst psychometric measures are typically questionnaires. Neurological studies

have the potential to link the ‘homeostatic’ and ‘hedonic’ neurological pathways of
appetite with obesity (Carnell, Gibson, Benson, Ochner, & Geliebter, 2012). In contrast, the
phenotypic expression of appetitive traits is more frequently investigated under laboratory
conditions or through psychometric measures (Blundell et al., 2005).
This section reviews the main methods of measuring appetite and appetitive traits as
identified in a number of key reviews and book chapters (Allison & Baskin, 2009; Blundell et
al., 2009; Faith, Carnell, & Kral, 2013; French, Epstein, Jeffery, Blundell, & Wardle, 2012;
Llewellyn, Carnell, & Wardle, 2011).
2.4.1 Experimental measures of appetite and appetitive traits
Experimental measures of appetite and appetitive traits can be subdivided into neurological
and laboratory measures.
2.4.1.1 Neurological measures
Neurological markers of appetite capture the neural appetite response pathways by using
brain activation imaging. Typically, neuroimaging studies have used positron emission
tomography (PET), functional magnetic resonance imaging (fMRI) and magnetic resonance
imaging (MRI) to assess appetite. In these studies, the appetite of a study participant is
triggered by a food cue and the measured appetite response could represent both normal
neurological responses (‘homeostatic’) and reward anticipation responses (‘hedonic’), as
well as cognitive attempts to inhibit those responses (Carnell et al., 2013).
2.4.1.2 Laboratory measures
Laboratory measures of appetite can be subdivided into those designed to measure

prandial (within-meal effects or effects that occur during the eating process) such as eating
speed or post-prandial (effects that occur following eating) such as satiation. These usually
use a pre-load test (as the independent variable) that is strictly fixed in terms of factors
such as weight, volume, energy density, macronutrient content, with only the variable
under investigation allowed to vary (e.g. measuring response to food cues, where a snack is
presented 15 minutes after the consumption of a standardised meal). Satiety can be
measured using time until the next eating episode and by the characteristics of food
consumed under strict fixed conditions (weight, energy density, etc.) (Blundell et al., 2009).
Examples of some of these measures for certain appetitive traits are given below.
37
‘Food responsiveness’, ‘external eating’ and ‘enjoyment of food’
Experimental studies of the sensory activation of eating (Jansen et al., 2003) expose
participants to sensory food cues (intense smell of tasty food) versus no food cues (control
task, such as an activity) and measure the amount consumed of a particular food after
exposure (Jansen et al., 2003). These experiments test participants’ response to external
food cues (i.e. ‘food responsiveness’, ‘external eating’, and ‘enjoyment of food’), and
measure participants’ capacity to down-regulate their appetite after food consumption (i.e.
‘satiety responsiveness’) (Carnell & Wardle, 2008b).
In children, ‘external eating’ behaviour, or eating in the absence of hunger (EAH) (Birch,
Fisher, & Davison, 2003; Fisher & Birch, 1999), has also been studied, by measuring a
tendency to over-eat palatable food (‘food responsiveness’ and ‘enjoyment of food’). The
laboratory setting uses a ‘free-access procedure’, where access to low nutrient and high
energy density foods (such as pretzels, chocolate, or popcorn), presented as snacks, 15
minutes after the consumption of a standardised meal and the child feels full and no longer
hungry (Fisher & Birch, 2002).
The value of food has also been studied in the laboratory through exploring in what
circumstances children choose palatable food over food with lower palatability or over
other enjoyable activities (Temple, Legierski, Giacomelli, Salvy, & Epstein, 2008). This
model of appetite measurement assesses how much a participant is willing to ‘work’ to
access food of higher versus lower palatability, or for a food reward versus a non-food
reward (such as a video game), thus measuring the motivational drive to eat for pleasure or
reward. Progressive schedules of reinforcement are set up to measure the amount of work
a person is willing to do to obtain a food reward (Lappalainen & Epstein, 1990). The
‘reinforcing value of food’ is dictated by the time it takes for a person to choose a non-food
reward versus a palatable food, which acts as a measure of responsiveness to external food
cues (i.e. ‘food responsiveness’ and ‘external eating’).
Other laboratory methods for exploring ‘food responsiveness’ and ‘enjoyment of food’
include food preference studies (i.e. the extent to which an individual likes certain foods)
38
(Birch, 1998). These studies often use taste tests10 whereby participants sample a range of
foods and rank their preferences among the tested foods (Blundell & Cooling, 1999; Halford
et al., 2008). Preferences for highly palatable foods suggest someone is motivated by
‘hedonic’ pathways which are transmitted by reward circuits that over-ride ‘homeostatic’
pathways. This represents an interaction between ‘liking’ of food linked to affect and
incentive, versus a more motivational ‘wanting’ component (Blundell et al., 2009; Finlayson
et al., 2007).
‘Satiety responsiveness’ and ‘slowness in eating’
Laboratory measures of caloric compensation assess the ability of a participant to adjust

food intake according to the energy level of a pre-load. The compensation can be
quantified in both adults and children by comparing the amount of an ad libitum meal
eaten a short time after a pre-load, which can be higher or lower in energy (Carnell &
Wardle, 2007b; Johnson & Birch, 1994; Mattes, Pierce, & Friedman, 1988). These measures
test the idea that an individual who is responsive to internal satiety cues (i.e. has high
‘satiety responsiveness’) is able to compensate their food intake according to the energy
content of the pre-load given before a meal. Those individuals who are not sensitive to
their internal satiety cues will not compensate (Carnell & Wardle, 2007). A short-term
energy-compensation procedure (COMPX score) is given to indicate how much a meal is
compensated for after pre-load ingestion (Johnson & Birch, 1994).
Microstructural analysis of ingestive patterns is a method used to measure eating rate and
the trajectory or stages of eating, breaking it up into smaller structures, such as quantity of
food per unit of time per meal (Guss & Kissileff, 2000). A slowing down of eating rate,
demonstrated by a decelerating cumulative intake curve is associated with a ‘normal’
pattern of satiety in adults, while non-deceleration is associated with low ‘satiety
responsiveness’ (Meyer & Pudel, 1972). Eating rates are thought to signify one’s level of
hunger and motivation to eat. ‘Slowness in eating’ is operationalised as the total amount of
energy consumed (calories or mouthfuls) within a given time and is measured as kcal/min
10
Taste-tests are more specific measures of ‘liking’ (i.e. the incentive values of food) (Finlayson et al.,
2007).
39
or bites/min. A faster eating rate has been associated with a greater intake of food (Kaplan,
1980).
‘Food fussiness’
‘Food fussiness’ or picky eating is a tendency to be extremely selective about foods (Taylor,
Wernimont, Northstone, & Emmett, 2015). It refers to both the rejection of unfamiliar
foods (neophobia), as well as known or familiar foods. Picky eating behaviour in both
adults and children has been measured in laboratory settings using food selection
situations (Pliner & Hobden, 1992). Taste tests can also be used as a proxy measure of
fussiness, with participants who report liking fewer foods or a narrow range of foods rated
as more ‘picky’ (Blundell & Cooling, 1999; Halford et al., 2008).
‘Restraint’ and ‘disinhibition’
‘Restraint’ is measured under laboratory conditions by observing whether individuals

consciously attempt to control their energy intake, by restricting food intake in response to
a high calorie pre-load (Herman & Polivy, 1975; Polivy et al., 1979). Increased food intake
described as counter-regulation11 or ‘disinhibition’ (Johnson et al., 2012; Stunkard &
Messick, 1985) (Section 1.3, Chapter 2), has been observed in response to dysphoric mood
(Herman & Polivy, 1975) and alcohol ingestion (Polivy & Herman, 1976b). Restrained eaters
who are induced to break their ‘restraint’ with a high calorie pre-load and then asked to eat
unlimited palatable food have shown counter-regulation (Wardle & Beales, 1988), but
these behaviours cannot be extrapolated to real world situations (Johnson et al., 2012).
2.4.1.3 Limitations of experimental measures
Experimental measures of appetite have the advantage of being objective measures of

eating behaviour under different conditions and are used to accurately measure particular
aspects of food response such as taste or preference. However, they have limitations, and
11
From the “Restraint” theory, counter-regulation refers to control over eating being undermined,
which results from trying to control eating cognitively (Herman & Mack, 1975; Herman & Polivy,
1975).
40
concerns have been expressed about their validity12 (i.e. whether they are actually
capturing the aspect of appetite which they intend to measure), reliability and whether the
findings are reproducible (Carnell & Wardle, 2007). For example, it is difficult to control the
entire diet for a group of individuals over a prolonged period of time and therefore
laboratory measures of appetite primarily focus on one mealtime without assessing food
intake throughout the day (Carnell & Wardle, 2007). In the case of neurological measures,
it is difficult to expose participants to different food cues (e.g. visual and olfactory)
simultaneously in order to track the neurological response of one particular system (Carnell
et al., 2013), and these studies can be difficult and inconvenient to run. Experimental
studies, which assess a particular meal condition are more viable, but they have significant
limitations in their application to human obesity, as behaviour is only captured on one
occasion in an artificial context; therefore they cannot claim to be true measures of traits
(Carnell & Wardle, 2008b). Furthermore, given that conditions have to be very strict for
reproducibility, experimental studies do not resemble everyday life (Blundell et al., 2009).
Experimental studies of appetite can be expensive as they require special laboratory
settings, and are typically only possible in small samples, providing potential challenges for
statistical power and external generalisability (Carnell et al., 2013; Carnell & Wardle, 2007).
However, experimental measures are used to validate specific aspects of appetite captured
by psychometric assessment, and together both types of measures strengthen each other.
2.4.2 Psychometric measures of appetite and appetitive traits
The use of validated and reliable questionnaires to measure appetitive traits removes the
costly obstacles of laboratory and neurological measurements. Psychometric measures are
standardised quantitative questionnaires concerned with the study of psychological
dimensions. They are convenient to administer to large numbers of participants and are
beneficial for statistical power and may better reflect ‘real-world’ conditions (Streiner &
Norman, 2015). They can also be used to incorporate behaviours over many different
12
Different types of validity can be measured, mainly content validity which examines the content of
the items; criterion validity that measures how well the scores on a test agree with the performance
on a task it was meant to predict; and construct validity which refers to the nomologies embedded in
the scale (i.e. the meaning of the construct/trait being measured) and it can be either convergent or
discriminant (Streiner & Norman, 2015).
41
situations (e.g. ‘Do you eat more when you: smell food/ see others eating’). These in turn
may be used to reveal untapped behavioural ‘traits’ which are more stable or ‘tonic’ in
nature (Carnell & Wardle, 2007), as opposed to ‘states’13 related to periodic measurements
of appetite that tend to fluctuate (Blundell et al., 2009; Harrold et al., 2012).
An array of psychometric questionnaires have been used to demonstrate that different

aspects of appetite are associated with: parental feeding practices, composition of dietary
intake, food preferences, and dietary patterns (Birch & Fisher, 1998; Birch et al., 2001;
Deglaire et al., 2012; Emmett, Jones, & Northstone, 2015). Psychometrically measured
aspects of appetite have also been linked with socio-environmental factors such as
frequency of family meals, healthy and unhealthy food availability, and parental or peer
group support (Cutler, Flood, Hannan, & Neumark-Sztainer, 2011); attempts to control
weight (Schembre, Greene, & Melanson, 2009; Tapper & Pothos, 2010); or pathological
aspects of appetite which include eating disorders in adults measured with the ‘Eating
Attitude Test’ (Garner & Garfinkel, 1979), disordered eating patterns in adolescents and
young adults (Larson, Neumark-Sztainer, & Story, 2009), and body image disturbances
(Kroon Van Diest & Tylka, 2010).
Some of the most commonly used tools for measuring appetite, include the ‘Three Factor
Eating Questionnaire’ (TFEQ) (Stunkard & Messick, 1985) the ‘Dutch Eating Behaviour
Questionnaire’ (DEBQ) (van Strein, Frijters, Bergers, & Defares, 1986), and in children the
‘Child Eating Behaviour Questionnaire’. The original TFEQ measures ‘restraint’,
‘disinhibition’ and ‘hunger’ in a 51-item questionnaire. The TFEQ has been revised into an
18 item TFEQ-R18, which measures ‘uncontrolled eating’ which includes ‘disinhibition’ and
‘hunger’ items from the original TFEQ, ‘cognitive restraint’ and ‘emotional eating’ (Karlsson,
Persson, Sjöström, & Sullivan, 2000). The DEBQ is a 33-item questionnaire that measures
13
Behavioural acts of eating and food selection are also accompanied by subjective states, so a
person experiencing strong hunger sensations may eat faster, quicker and more food than a person
who is not as hungry. Psychological aspects of eating motivation, however, allow the theoretical
distinction between different ‘states’ and ‘traits’ in order to study and measure appetite. ‘States’ are
related to periodic sensation of hunger, fullness and ‘wanting’ (the drive to eat), which occur
episodically, tend to fluctuate and are part of our eating patterns. ‘Traits’, on the other hand, are
more stable across time and situations and can be identified using psychometric questionnaires
(Blundell et al., 2009; Harrold et al., 2012).
42
‘external eating’, ‘restraint’ and ‘emotional eating’ in adults, as well as through parent
report (DEBQ-P) (Braet & van Strein, 1997) and through self-report (DEBQ-C) in children
(van Strein & Oosterveld, 2008). The CEBQ is a 35-item questionnaire, measuring eight
appetitive traits; ‘food responsiveness’, ‘emotional over-eating’, ‘enjoyment of food’,
‘desire to drink’, ‘satiety responsiveness’, ‘emotional under-eating’, ‘food fussiness’, and
‘slowness in eating’ (Wardle, Guthrie, et al., 2001). The CEBQ is a parent-report measure
for 3-13 year old children that has also been adapted to measure similar traits in infants
using the ‘Baby Eating Behaviour Questionnaire’ (BEBQ) (Llewellyn, van Jaarsveld, Johnson,
Carnell, & Wardle, 2011).
However, there is a need to systematically review all available measures of appetite and
appetitive traits. This is therefore the aim of Study 1 in Chapter 4 of this thesis.
2.4.2.1 Limitations of psychometric studies
Psychometric measures are not objective measures of appetite, but they have the potential
to reflect behaviour over a wide range of situations. Questionnaires lack detail about the
complexity of their subject, as the individual is not able to express fully how he/she feels
about his/her appetite (Oppenheim, 2003). Because self-report questionnaires about
behaviours are subjective, their reliability and validity should be tested and the potential
for self-report errors should also be taken into account (Streiner & Norman, 2015). One of
the most common problems with psychometric questionnaires is related to response set
issues, where a participant’s responses to questions are based on reasons other than those
intended by the researcher. For example; excessive positive or negative checking of
statements (acquiescence); a tendency to answer the extremes of the response format on
the questionnaire (extreme response set); and social desirability (as a tendency to choose
items in terms of the perceived desirability to others, rather than those reflecting the
person’s actual feelings or behaviour); are all problems of response set common to
psychometric measures (Allison & Baskin, 2009).
2.5 Appetitive traits and weight
2.5.1 The genetics of obesity
Obese individuals show the greatest vulnerability to weight gain caused by the obesogenic
environment (Ogden, Yanovski, Carroll, & Flegal, 2007). This concentration of weight gain
among the top end of the BMI spectrum, has been proposed to demonstrate a gene-
43
environment interaction in the development of obesity (i.e. that environmental

susceptibility may be genetically determined) (Carnell, Haworth, Plomin, & Wardle, 2008;
Llewellyn & Wardle, 2015; Wardle et al., 2008).
Source: (Ogden, Yanovski, Carroll, & Flegal, 2007)
Figure 2.2 Change in the distribution of BMI between 1976–1980 and 1999–2004, for
adults aged 20-74 years in the United States of America (USA)
There is considerable evidence for a genetic influence on obesity (O’Rahilly & Farooqi,
2008). The tendency for obesity to run in families has been demonstrated through family
studies, where obese parents were found to have a 40% chance of having an obese child;
two obese parents have double that possibility (Stunkard, Harris, Pedersen, & McClearn,
1990). Twin and adoption studies have provided the most useful evidence for the
heritability of weight so far, distinguishing between genetic and shared environmental
effects on body weight (Llewellyn & Wardle, 2015). A review of adoption studies showed
that children’s weight status was associated with that of their biological parents, but that
there was no association between the weight statuses of adopted children and their
adoptive parents (Grilo & Pogue-Geile, 1991), indicating a genetic basis for weight.
Additionally, twin studies have shown that monozygotic (i.e. identical) twins, who share
100% of their genes, have more similar BMI and WC measurements than dizygotic (i.e. non-
identical) twins, who on average share 50% of their genes (Clark, 1956). Thus, variation in
BMI has been attributed to genetic differences (Stunkard et al., 1990) and estimates of BMI
heritability are around 70% in adults (Maes, Neale, & Eaves, 1997; Schousboe et al., 2003;
Silventoinen, Rokholm, Kaprio, & Sørensen, 2010).
44
2.5.2 The “Behavioural Susceptibility Theory” of obesity
The “Behavioural Susceptibility Theory” (BST) of obesity posits that environmental and
genetic factors interact to promote weight gain (Carnell & Wardle, 2008a; Llewellyn &
Wardle, 2015). This model postulates that the genetic risk of obesity is expressed in terms
of appetitive traits which are genetically determined, and which are associated to different
eating behaviour phenotypes, across the weight spectrum (Croker, Cooke, & Wardle, 2011).
At an individual level, and under the appropriate environmental circumstances, the
presence of adverse appetitive traits could lead to a positive energy balance and possible
weight gain. The BST of obesity is depicted below Figure 2.3.
Individual
Weight
Environment
Appetitive Positive energy gain
traits balance
Genes
Figure 2.3 Relationships of appetitive traits to the genetically determined susceptibility

of the environment
2.5.3 Evidence for the “Behavioural Susceptibility Theory”
The appetitive traits posited by the BST to play a role in an individual’s susceptibility to
obesity, are those measured by the CEBQ. As such, evidence for the BST to date has been
mostly provided through studies in children (Carnell & Wardle, 2008a; Croker et al., 2011;
Sleddens, Kremers, & Thijs, 2008; Spence, Carson, Casey, & Boule, 2011; Viana et al., 2008),
and infants (Llewellyn, van Jaarsveld, et al., 2011; Llewellyn & Wardle, 2015; van Jaarsveld,
Llewellyn, Johnson, & Wardle, 2011) using the CEBQ and BEBQ respectively.
These traits have been shown to be heritable both in children (Carnell et al., 2008), and in
infants (Llewellyn, van Jaarsveld, Plomin, Fisher, & Wardle, 2012). In a study of 2402 infant
45
twin pairs, the heritability of ‘slowness in eating’ and ‘satiety responsiveness’ traits was
large, at 84% and 72%, respectively, and heritability was moderate for ‘food
responsiveness’ and ‘enjoyment of food’, at 59% and 53%, respectively (Llewellyn, van
Jaarsveld, Johnson, Carnell, & Wardle, 2010). Similarly, in a sample of twin pairs aged 8 to
11 years, ‘food responsiveness’ and ‘satiety responsiveness’ were estimated to have
heritable components, at 75% and 63%, respectively (Carnell et al., 2008). These findings
suggest that genes play an important role in the regulation of appetite in an environment
which is rich with food (Piernas, Ng, & Popkin, 2013) from an early age, and may continue
to regulate these traits over the life course (Llewellyn et al., 2010).
Furthermore, a number of studies have demonstrated an association between these traits

and weight. In observational studies, ‘food responsiveness’ and ‘enjoyment of food’ have
consistently been found to positively correlate with BMI-SDS14 in children (Carnell &
Wardle, 2008a; Fuemmeler, Lovelady, Zucker, & Ostbye, 2013; Mackenbach et al., 2012;
Rodenburg, Kremers, Oenema, & van de Mheen, 2012; Santos et al., 2011; Sleddens et al.,
2008; Soussignan, Schaal, Boulanger, Gaillet, & Jiang, 2012; Svensson et al., 2011; Viana et
al., 2008; Webber, Hill, Saxton, Van Jaarsveld, & Wardle, 2009). In a comparison of a
community sample and a clinical sample of children referred to a hospital obesity
programme, ‘food responsiveness’ was highest in the clinical sample, although this was not
observed for ‘enjoyment of food’ (Croker et al, 2011). The observed differences could have
been due to a lack of power given the small sample size of the clinical vs. the community
sample (n=66 vs. n=406). These differences were not observed in a study in Portugal where
240 children aged three to 13 years of age were drawn from both community and clinical
settings (Viana et al., 2008), although the authors failed to state the proportion of
participants taken from each setting.
Studies using the CEBQ have also consistently found negative associations between BMI-
SDS and ‘satiety responsiveness’ (Fuemmeler et al., 2013; Mackenbach et al., 2012;
Rodenburg et al., 2012; Santos et al., 2011; Sleddens et al., 2008; Soussignan et al., 2012;
14
BMI standard deviation scores (BMI-SDS) are measures of relative weight in children and
adolescents that are gender and age independent. They are calculated from BMI values by adjusting
for age and gender using British 1990 reference data (Freeman et al., 1995).
46
Svensson et al., 2011; Viana et al., 2008). In Croker et al.’s study described above, ‘satiety
responsiveness’ was lower in a clinical sample of obese children attending a weight
management programme than obese children from a community sample, suggesting that
those children with greater obesity are less able to feel internal satiety cues (Croker et al.,
2011). ‘Slowness in eating’, also measured using the CEBQ, has also been shown to be
negatively associated with weight in a number of studies (Croker et al., 2011; Mallan et al.,
2013; Parkinson, Drewett, Le Couteur, & Adamson, 2010; Santos et al., 2011; Sleddens et
al., 2008; Soussignan et al., 2012; Sparks & Radnitz, 2012; Spence et al., 2011; Viana et al.,
2008; Webber et al., 2009).
Another trait measured using the CEBQ, ‘emotional over-eating’, has consistently been
reported to positively associate with BMI-SDS in children (Mallan et al., 2013; Rodenburg et
al., 2012; Soussignan et al., 2012; Sparks & Radnitz, 2012; Svensson et al., 2011). However,
associations between ‘emotional under-eating’ and weight have been somewhat
inconsistent. Most studies have found negative correlations in children (Mallan, Nambiar,
Magarey, & Daniels, 2014; Rodenburg et al., 2012; Svensson et al., 2011). Clinical groups
also scored lower for ‘emotional under-eating’ than community groups (Croker et al.,
2011). However, other studies have reported no relationship between ‘emotional under-
eating’ and weight (Hill, Saxton, Webber, Blundell, & Wardle, 2009; Loh, Moy, Zaharan, &
Mohamed, 2013; Mackenbach et al., 2012; Parkinson et al., 2010; Sparks & Radnitz, 2012;
Spence et al., 2011). The studies reporting no association included one using a self-report
version of the CEBQ developed for 13 year old adolescents (Loh et al., 2013); a study using a
longitudinal birth cohort of maternal responses to the CEBQ at six weeks, 12 months and
five to six year old infants and children (Parkinson et al., 2010); and studies from diverse
socio-economic groups where confirmatory factor analysis revealed a different structure
for the CEBQ compared to the original (Loh et al., 2013; Sparks & Radnitz, 2012).
Fussy eating or pickiness has been associated with failure to thrive (Wright & Birks, 2000),
although these findings have been somewhat inconsistent (Carruth & Skinner, 2000), and it
has also been suggested to confer protection against weight gain (Llewellyn, Carnell, et al.,
2011; Wardle, Guthrie, Sanderson, & Rapoport, 2001). ‘Food fussiness’ measured using the
CEBQ has similarly shown inconsistent relationships with weight. Some studies have found
no relationship between ‘food fussiness’ and weight (Santos et al., 2011; Svensson et al.,
2011), whereas others have reported negative associations with weight (Hill et al., 2009;
Loh et al., 2013; Mallan et al., 2013; Rodenburg et al., 2012; Spence et al., 2011; Viana et
47
al., 2008). In the previously mentioned study comparing a community sample and a clinical
sample referred to a hospital obesity programme, the clinical group scored higher for ‘food
fussiness’ than the community group (Croker et al, 2011). The authors suggested that
clinical studies tend to select samples of children with greater feeding difficulties, which
could have been over-represented this sample.
Lastly, evidence for these appetitive traits mediating the relationship between genes and
BMI has recently been provided through a study from the Twins Early Development Study
(TEDS) in 2258 unrelated, ten-year-old children. Polygenic obesity scores comprising 28
known obesity-related variants were associated with ‘satiety responsiveness’ assessed
using the CEBQ. This study showed that whilst BMI-SDS and WC increased with an increase
in the genetic risk of obesity, ‘satiety responsiveness’ decreased (Llewellyn, Trzaskowski,
van Jaarsveld, Plomin, & Wardle, 2014).
Some evidence relating appetitive traits to weight in adulthood has come from other
measures, mainly the TFEQ and the DEBQ. ‘External eating’, measured using the DEBQ, has
been positively associated with weight in a number of studies in adults (Koenders & van
Strien, 2011; van Strien et al., 1986). Conversely in a study of adolescents, Wardle et al. in
1992 showed ‘external eating’ was highest among the lowest BMI groups. In this study,
‘external eating’ was lower in those who perceived themselves as being fatter and also for
those who were more ‘restrained’. ‘Emotional eating’ assessed by the DEBQ is generally
associated with increased weight in adults in clinical settings (van Strein et al., 1986;
Wardle, 1987a). Similarly, when ‘emotional eating’ was measured in children using the
parent version of the DEBQ (DEBQ-P), it was higher in obese vs. non-obese children drawn
from clinical samples (Braet & van Strein, 1997). Although others did not find these
associations in adolescents (Wardle et al., 1992). Using the DEBQ to measure ‘restraint’,
obese adult participants had significantly higher scores on the ‘restraint’ scale than normal
weight subjects (van Strein et al., 1986). In children, BMI-SDS also correlated positively
with DEBQ-C ‘restraint’ but only in the normal weight groups (van Strein & Oosterveld,
2008). In pre-adolescents, using the parental report version of the DEBQ, obese and
overweight subjects had higher values of ‘restraint’ than the normal-weight pre-
adolescents (Caccialanza et al., 2004). However, in general, the DEBQ has primarily been
used in the context of disordered eating behaviours (Johnson & Wardle, 2005; van Strein et
al., 1986; Wardle, 1987a), compared to the evidence provided for the BST of obesity which
48
shows a relationships between appetitive traits and weight across the weight spectrum
(Croker et al., 2011; Llewellyn & Wardle, 2015).
The TFEQ also measures a form of ‘emotional eating’ through its ‘disinhibition’ sub-scale.
The ‘disinhibition’ sub-scale is comprised of two aspects involving weight fluctuation, as
well as ‘emotional eating’ and ‘external eating’ (Arnow et al., 1995; Stunkard & Messick,
1985). Higher positive ‘disinhibition’ scores have been associated with higher energy intake
and higher BMI (Stunkard & Messick, 1985). Further analyses of the TFEQ have failed to
replicate the factor structure of the original 51-items, and have led to the loading of items
onto an ‘emotional eating’ sub-scale to produce two revised (shortened) versions of the
TFEQ, the TFEQ-R18 and the TFEQ-R21 (Cappelleri, Bushmakin, Gerber, Leidy, Sexton, Lowe,
et al., 2009; Karlsson, Persson, Sjöström, & Sullivan, 2000). Positive associations between
‘cognitive restraint’ and weight have been found in normal weight, but not overweight
subjects when using the TFEQ-R18 (de Lauzon-Guillain et al., 2006). Similar associations
were also found using the TFEQ-R18, where higher BMI was associated with higher levels of
‘cognitive restraint’ and ‘emotional eating’, but not with ‘uncontrolled eating’ (Anglé et al.,
2009). TFEQ-R21 ‘cognitive restraint’ and BMI correlations have been found to be
significant in clinical samples (Cappelleri, Bushmakin, Gerber, Leidy, Sexton, Lowe, et al.,
2009). However, negative associations have also been reported using the TFEQ, where
lower ‘cognitive restraint’ and higher ‘disinhibition’ scores were associated with higher BMI
(Williamson et al., 1995). Differences in the associations between weight and ‘restraint’
appear to vary according to the weight status of the samples being studied; positive
associations have been observed in individuals of normal weight (Williamson et al., 1995),
whereas in obese populations, negative associations have been reported (Cappelleri,
Bushmakin, Gerber, Leidy, Sexton, Lowe, et al., 2009). It has therefore been suggested that
‘restraint’ might help to diminish the adverse effects of appetitive traits on weight gain in
obese populations (Johnson et al., 2012).
The TFEQ also measures ‘hunger’, as a measure which relates to an individual’s perception
of their level of motivation to eat and the extent to which this elicits food intake (Stunkard
& Messick, 1985). ‘Hunger’ (and ‘disinhibition’) when combined with an increase in
‘restraint’ were associated with weight loss parameters in 58 overweight and obese
participants who completed a 12-weeks exercise supervised program (Bryant, Caudwell,
Hopkins, King, & Blundell, 2012). However this could be due to confounding effects of
‘disinhibition’, and it has been suggested that people who usually feel hungry could also
49
present other eating behaviours such as cravings and disordered eating (Elfhag & Rössner,
2005). In a study of Finnish adolescent women aged 17 to 20 years, ‘cognitive restraint’
and ‘emotional eating’ measured using the TFEQ-R18 showed positive associations with
BMI, but no associations were seen for ‘uncontrolled eating’ (which includes ‘hunger’
items) (Anglé et al., 2009). Again, it is unclear whether the relationship between ‘hunger’
and weight is being confounded by ‘disinhibition’. Also problematic is the measurement of
‘hunger’ itself as it is subject to great variability dependent on how it is measured and also
the timing of the measurement (Wardle, 1987b).
2.6 Summary of the findings
Appetitive traits are stable predispositions towards food, which make individuals more or
less susceptible to certain environmental exposures that can contribute to the
development of obesity. Given the complex processes involved in ‘homeostatic’ and
‘hedonic’ regulation of appetite, measurement of appetite and appetitive traits has
included both experimental and psychometric assessments. Experimental measures, which
include both neurological and laboratory-based measures, objectively assess different
aspects of appetite under very specific conditions, limiting their generalisability to natural
eating conditions. Psychometric measures of appetite, which measure different dimensions
of appetite, have endeavoured to address some of the limitations of experimental
measures and are useful for obtaining data from large populations in real-world settings.
The “Behavioural Susceptibility Theory” (BST) of obesity proposes that individual

differences in weight are due to variation in appetitive traits. However, thus far the
evidence for the BST has primarily come from studies in children and there is a lack of
empirical research regarding many of the traits captured by the CEBQ in adults. Such
studies would provide the evidence needed to demonstrate if the relationship between
appetitive traits and weight still holds into adulthood, and if this could inform the
development of tailored interventions to help individuals manage these traits and in turn
their weight. The lack of such studies may be because there is no adult measure of the
appetitive traits captured by the CEBQ. However, it is possible that other existing measures
may capture similar traits and so a systematic review of existing measures for different age
groups and the traits they capture is warranted.
50
Chapter 3. Aims of the thesis
Chapters 1 and 2 describe the need for weight management interventions and how a
person’s appetitive traits interact with the current obesogenic environment to determine
their individual susceptibility to overweight or obesity. Identifying an individual’s specific
pattern of appetitive traits could potentially enable personalised and targeted feedback for
weight management interventions.
Evidence supporting the association between appetitive traits and weight comes primarily
from paediatric studies (Llewellyn & Wardle, 2015). The majority of this research has
measured appetite in children using the CEBQ or the infant version, the BEBQ (Llewellyn,
van Jaarsveld, et al., 2011; Wardle, Guthrie, et al., 2001). However, it is unclear whether
these relationships hold into adulthood, and whether existing psychometric measures of
appetite can adequately assess these traits in adult populations. A systematic review of
existing measures is necessary to explore the justification for the development of a new
measure of appetitive traits for use in adults. A standardised psychometric measure of
appetitive traits during adulthood, measuring traits comparable to those measured by the
CEBQ, would allow large-scale studies to establish relationships with BMI at different stages
of the life course. Furthermore, if associations between specific dimensions of adult
appetite and BMI were established, this information could be used to tailor individualised
weight management advice to overweight and obese adults as part of a behaviour change
intervention.
This thesis aims to address the following research questions:
1. What psychometric measures of appetitive traits currently exist?

2. Can the parent report Ćhild Eating Behaviour Questionnaire´ (CEBQ) be adapted into
a valid and reliable measure of appetitive traits in adults?
3. How do appetitive traits relate to BMI in adults?
4. Can a weight management intervention tailored to an individual´s appetitive traits be
developed that is acceptable and potentially useful?
The studies in Chapters 4 to 8 attempt to address the above questions. Study 1 will explore
psychometric measures of appetite previously used in adults and children through a
systematic review. Study 2 describes the adaptation of the CEBQ - a parent report
51
questionnaire - into a self-report ‘Adult Eating Behaviour Questionnaire’ (AEBQ), and an

exploration of its factor structure in a sample of adults aged 18+ years old (Sample 1).
Study 3 aims to validate the AEBQ through confirmatory factor analysis (CFA) in a different
sample of adults (Sample 2), and to establish the internal and test-retest reliability of the
scales. Associations between appetitive traits and BMI will also be established. Study 4 will
develop and test a brief intervention, tailoring weight management tips to overweight and
obese individuals, based on their individual appetitive trait profile from AEBQ scores.
Finally, Study 5 aims to assess participants’ experiences of participating in this intervention
through qualitative interviews.
3.1 My contributions to the research in this thesis
I played a key role in developing the aims of this thesis and the design of the studies in
conjunction with my supervisors Professor Jane Wardle (who sadly passed away in October
2015, shortly after I had completed my final study), Dr Rebecca Beeken (my primary
supervisor) (RJB), Dr Alison Fildes (AF), Dr Helen Croker (HC), and Dr Fiona Johnson (FJ). I
performed all of the statistical analysis and interpreted the results with the help of my
supervisors.
I carried out the systematic review in Study 1 with help from UCL Librarians to obtain
adequate search terms. I selected the appetite measures according to the eligibility
criteria, discussed previously with Dr Beeken and Dr Croker. Final study selection was also
reviewed with the help of Dr Fildes and the rest of my supervisors.
For Study 2, I applied for and obtained ethical approval and was involved in all stages of the
development of the AEBQ, including the translation of the items, piloting of the preliminary
questionnaires and coordinated the data collection with the research sampling company. I
also processed and cleaned all the data (Sample 1) and conducted the Principal Component
Analysis (PCA), with additional statistical support provided by Dr Clare Llewellyn. I carried
out the iteration processes, and final AEBQ item selection was agreed in collaboration with
Professor Wardle, Dr Beeken and Dr Croker.
For Study 3 (Sample 2), data collection was conducted by myself and a fellow PhD student
Nathalie Kliemann, coordinating with the same research sampling company used in Study 2
(Sample 1). Ms. Kliemann and I jointly achieved ethical approval for this research. I carried
out the Confirmatory Factor Analysis (CFA) (Study 3) with the help of statistician, Tao Ding.
52
However, I ran the analysis myself using SPSS AMOS and interpreted the results with
minimal statistical assistance. I independently conducted the remaining statistical analyses
for Study 3.
In collaboration with my supervisors I designed and achieved ethical approval jointly with
Ms. Andrea Smith a fellow PhD student, and carried out the intervention for Study 4 and 5.
I conducted all 21 qualitative interviews in Study 5 and personally transcribed three
interviews, with the remaining transcriptions conducted by an independent company. I
carried out all the coding using NVivo, generated the themes and these were finalised with
the help of Dr Beeken and Dr Fildes.
53
Chapter 4. Systematic review on questionnaire measures of appetite and appetitive traits
Chapter 4. Study 1: Systematic review on questionnaire

measures of appetite and appetitive traits
4.1 Introduction
Appetite is a process involved in food selection and intake that demonstrates both trait and
state-type elements (Chapter 2). It is stimulated by ‘homeostatic’ and ‘hedonic’ processes
and, within an individual, is ‘episodic’ in nature, characterised by sensations of hunger or
fullness (Blundell et al., 2009; Harrold et al., 2012). Appetitive traits are stable
predispositions and have a more ‘tonic’ form of expression. They encompass a range of
eating behaviour dimensions such as responding to internal and external food cues, or
eating at a faster or slower rate, and are posited to play a key role in an individual’s
vulnerability to gain weight (Blundell et al., 2005, 2009; French et al., 2012; Harrold et al.,
2012; Wardle & Carnell, 2009). Appetitive traits are thought to drive different expressions
of an individual’s appetite. Given the wide variability that has been shown in body weight
and weight gain (Ogden, Yanovski, Carroll, & Flegal, 2007; Wardle & Boniface, 2008)
(Chapter 2, Section 2.5.1), individual characteristics that interact with the environment,
such as appetitive traits, have the potential to increase or decrease environmental risk.
Appetite can be measured experimentally or psychometrically, and there are advantages

and disadvantages to both of these methods, as discussed in Chapter 2. While
psychometric measures do not have the objectivity of experimental measures, they may be
used to collect information from large numbers of people in a practical and inexpensive
way. In 2012, French et al., published a selective review of the psychometric and
experimental measures used to assess the relationship between appetitive traits and
weight, and to capture key dimensions of eating behaviour. The specified inclusion criteria
included reported associations with energy intake, food choices, body weight, or weight
gain. From these measures, the authors identified seven eating dimensions that are
thought to influence energy intake when expressed in a permissive food environment:
‘food responsiveness’, ‘enjoyment of eating’, ‘satiety responsiveness’, ‘eating in the
absence of hunger’, ‘reinforcing value of food’, ‘eating disinhibition’ and ‘impulsivity/self-
control’ (French et al., 2012). This review provides a helpful overview of appetitive traits
that are related to energy intake, however, it was not conducted in a systematic manner,
54
and therefore some published questionnaires are likely to have been omitted. The
psychometric strengths of existing appetitive trait measures were also not assessed.
An up-to-date systematic review of psychometric measures of appetite and appetitive traits

that may influence body weight is needed to provide a better understanding of the
research landscape in this area. It is important to recognise how different aspects of
appetite and eating behaviour are conceptualised that could help explain individual
differences in weight, and to identify measures that are psychometrically sound and age
appropriate to assess these traits across the life course.
4.2 Aim
This study aimed to: (1) systematically review the relevant literature to identify existing
psychometric measures of appetite; (2) assess the psychometric properties of the measure
through their reliability and validity; and (3) identify the most commonly measured
appetitive traits in different age groups.
4.3 Methods
4.3.1 Information sources and search strategy
This review followed the Preferred Reporting Items for Systematic Reviews and Meta-
Analyses (PRISMA) statement (Moher, Liberati, Tetzlaff, & Altman, 2009). The search
strategy was developed with the assistance of Dr Croker (HAC), Dr Beeken (RJB), and Dr
Fildes (AF), following the instructions of University College London (UCL) librarians. A
systematic literature search was carried out using the Embase, MEDLINE®, PsycINFO and
PsychEXTRA databases to find relevant articles conducted in any country or language that
were published in English until the 26th of January 2016. The search strategy included
terms relating to appetite and eating behaviour; questionnaires, scales, measures and
instruments; food and eating; validation, reliability, development and adaptation; and was
limited to humans. See Appendix 4.1 for the complete electronic search strategy. The
reference lists of all included articles were manually checked for other relevant articles.
Once the questionnaires were selected from the search and included in the qualitative
synthesis, further searches using Google Scholar were carried out to make sure inclusion of
validation studies, including validation of measures through experimental studies (as stated
55
in the inclusion criteria), and test-retest reliability. Any additional articles were then
included to assess the psychometric properties of the questionnaires, but were not
included as part of the original search, just referenced.
4.3.2 Eligibility criteria
Initially, all psychometric measures of appetite for adults or children published to date were
included. Peer-reviewed articles in English were considered for inclusion, with the
exclusion of reviews, conference abstracts, or dissertation abstracts. Measures were
included if they sought to measure appetite or appetitive traits, and if the traits being
measured were proposed by the authors to be related to body weight. I was interested in
measures used for both general population-based samples and overweight, obese and/or
clinical samples, as appetitive traits have been shown to be linearly related to weight across
the whole weight spectrum (Croker et al., 2011). Measures were excluded if they: were not
the original questionnaire; had not been assessed for validity or reliability and this was not
included in any additional study; were experimental or laboratory-based, although these
methods could be used to validate a psychometric measure; assessed attempts to control
or modify appetite (including measures of self-regulation); monitored food intake, single
food items, or nutrients, or food frequency; included measurement using nutritional
software; measured weight control and dieting; were designed for use as a clinical
diagnostic tool; or, were designed for exclusive use in a clinical population, including
measures of symptoms associated with pathologically disordered eating (e.g. Anorexia
Nervosa). If the measure was later used in non-pathological participants and in relation to
obesity, it was still included. The inclusion and exclusion criteria for the review are shown
in Table 4.1.
56
Table 4.1 Inclusion and exclusion criteria for a systematic review of psychometric
measures of appetite
Inclusion criteria Exclusion criteria
• In humans. • Reviews.
• Adults and • Conference abstracts.
children.
• Dissertation abstracts.
• General
• Not published in the English language.
population- based
samples as well as • Did not contain original questionnaire.
overweight, obese
• Did not assess the validity, reliability.
and/or clinical
samples. • Laboratory measures of appetite or validation of
laboratory measures or recording sessions
• Cross-sectional and
(observational methodology).
longitudinal
studies. • Measures that monitor eating, single food items, or
nutrients, food frequency, use of visual analogue scales
• Published in
or nutritional software.
English.
• Questionnaires which measure change or attempts to
• Peer reviewed
control or modify appetite.
articles
• Questionnaires relating to body image, anthropometry,
• Psychometric
malnutrition
measures of
appetite which use • Questionnaires relating to diseases or medical and
the definition surgical treatments (e.g. cancer, Prader-Willi
specified in Syndrome, bariatric surgery) or disabilities.
sections 2.2 and
• Measures of parental/caregiver feeding
2.3 related to
practices/strategies, home/school enviroment, social
weight or
or cultural enviroments, external influences.
proneness to
obesity. • Measures of eating disorder symptoms or eating
pathology (addictions, anorexia nervosa, bulimia
nervosa, binge eating, etc.).
• Measures related to weight control and dieting.
• Diagnostic tools.
4.3.3 Study selection
Returned article titles and abstracts were initially screened using EndNote X7® referencing
software, to see if any should be excluded. Those that appeared to meet the inclusion
criteria were downloaded in full-text. The full-text articles were then read to confirm that
they met the inclusion and exclusion criteria. When multiple articles relating to a single
questionnaire were identified, either the article that was published first or the one that
presented the development of the questionnaire was selected. Where appropriate, other
articles were used as evidence for validity and reliability to assess the robustness of the
57
measures, but were not included in the final search and simply referenced. Child, adult,
and parental versions of the same questionnaires were retained for inclusion. One
reviewer (CH) performed the initial search, screening, and data extraction. The second
reviewer (AF) checked all included articles and a sub-set of the excluded articles against the
eligibility criteria.
4.3.4 Data extraction process – Classifying and coding studies
Data from the first paper describing the questionnaire or the development paper were
extracted in agreement with HAC, RJB, and AF. The extracted data items, which were
defined a priori, were: (1) Study reference, including the year of publication and country of
origin, (2) aim of the measure, (3) sample size and participants involved (children,
adolescents or adults), (4) age (mean±sd), (5) gender (Male/Female), (6) BMI (kg/m2;
mean±sd), (7) measure response options, (8) the statistical test(s) used, as well as the
number of factors and items obtained.
4.3.5 Assessing the robustness of the questionnaires
The robustness15 of the questionnaires was evaluated by assessing the psychometric

properties of the measures: (1) internal reliability; (2) test-retest reliability; (3) convergent
validity or when this information was not available, content or criterion validity; and, (4)
discriminant validity. An overall measure of robustness based on a point system was
developed. One point was awarded for each criteria met, for a total achievable score of
four. Measures scoring four points were defined as being ‘robust’.
1. Evaluation of psychometric properties
The internal reliability, test-retest (external) reliability, convergent validity, and discriminant
validity of measures was assessed using a pre-determined scoring system from the
standards jointly published by the American Education Research Association, the American
Psychological Association and the National Council on Measurement in Education
(American Educational Research Association, American Psychological Association, &
15
The robustness of a measure is a term used to provide an indication of how ‘good’ the
questionnaire is, i.e. it serves to assess the ‘quality’ of the questionnaire.
58
National Council on Measurement in Education, 1999; Streiner & Norman, 2015). The full
joint standards for assessment of psychometric measures are shown in Appendix 4.2. One
point was given for each of the following:
(1) Internal reliability16: Cronbach’s alpha ≥ 0.7 (Field, 2013);
(2) Test-retest reliability (or external)17: Intra-class correlation coefficients ≥ 0.7 or

significant Pearson’s correlation coefficient or Student t-test between two time points
(Field, 2013);
(3) Convergent validity18: Significant positive Pearson’s correlations coefficient against

another questionnaire measure of appetite; r’s in the mid-range of 0.4 to 0.8 to ensure
similar attributes are being measured (Streiner & Norman, 2015). Some authors used
Cohen’s criteria to indicate an effect size (i.e., r >0.50 a large/strong effect size; r around
0.30, a medium/moderate effect size; and r around 0.10, a slight/negligible effect size)
(Cohen, 1988). In some cases, convergent validity was measured using other questionnaire
measures not related to appetite but were correlated (e.g. self-esteem or social
desirability), and these were also included. When convergent validity against other
measures was not calculated, content, or criterion validity against the scales within the
measure were included, although these last forms of validation were not considered the
preferred scoring method;
16
Internal reliability shows the degree of inter-correlations which exists between the items in a
scale. It measures the consistency of the scale. Assessed by Cronbach’s alpha (Allison & Baskin, 2009;
Streiner & Norman, 2015).
17
Test-retest (external) reliability is measured through test-retests as a measure of external
consistency over time, from the first time the test was taken, to the next (Allison & Baskin, 2009;
Streiner & Norman, 2015).
18
Convergent validity refers to the relationship between the measure and another questionnaire
which measures similar constructs (e.g. Correlations between the PFS and the TFEQ-R21 ‘emotional
eating’ scale) (Allison & Baskin, 2009).
59
(4) Discriminant validity19: Lack of correlation between different scales or sub-scales was
used to assess discriminant validity; ranges between 0.0 to ±0.3 are considered to be
negligible (Streiner & Norman, 2015).
4.3.6 Most commonly measured appetitive traits by age group
The most robust questionnaires were screened within Google Scholar to identify those that
were the most commonly used based on the number of citations they had. The top three
most cited were grouped by the target age group for whom the questionnaire was
designed, to show which scales have been most commonly used in children and in adults.
4.4 Results
4.4.1 Search results
A total of 38 studies conducted in nine different countries were included in this systematic
review (Figure 4.1). The predominant country of development of a measure was the United
States (n=18) (Measures # 1-7, 11-13, 17, 22-23, 25,27-30; Table 4.2); five in the UK
(Measures # 10, 14, 18, 35-36; Table 4.2); five in Canada (Measures # 6, 16, 19-20, 34; Table
4.2); four in the Netherlands (Measures # 9, 31-33; Table 4.2); three in Germany (Measures
# 8, 24, 38; Table 4.2); one in Malaysia (Measures # 37; Table 4.2); one in Sweden
(Measures # 26; Table 4.2); one in China (Measures # 15; Table 4.2); and one in Italy
(Measures # 21; Table 4.2). All the questionnaires were developed and tested using cross-
sectional studies and convenience sampling. Study populations varied from students
(Avalos & Tylka, 2006; Tylka & Kroon Van Diest, 2013; Tylka, 2006) to obese individuals
(Braet & van Strein, 1997; Schembre & Geller, 2011; Stunkard & Messick, 1985; Tanofsky-
Kraff et al., 2008).
A total of 14 questionnaires were developed for use in children and adolescents (Measures
# 2, 5, 15-16, 20-21, 23-24, 28, 32-33, 35-37; Table 4.2), ages two to thirteen years,
19
Discriminant validity refers to the lack of correlations which should exist between dissimilar
unrelated variables (e.g. No associations were found between the EES sub-scales and TFEQ ‘cognitive
restraint’) (Allison & Baskin, 2009; Streiner & Norman, 2015).
60
including one in infants up to 24 months (Llewellyn, van Jaarsveld, et al., 2011) and four in
adolescents aged 12 to 18 years (Boggiano, Wenger, Mrug, et al., 2015; Loh et al., 2013;
Rollins et al., 2014; Tanofsky-Kraff et al., 2007, 2008). Twenty-four questionnaires were
developed for use in adult populations (Measures # 1, 3-4, 6-14, 17-19, 22, 25-27, 29-31,
34, 38; Table 4.2). The majority of the questionnaires used Likert-style scale response
options from 1 to 5, with 14 measures using the ‘never’ to ‘always’ format (Measures # 5, 7,
9, 12, 14-15, 17, 25-28, 36-38; Table 4.2) and 10 measures using ‘strongly agree’ to ‘strongly
disagree’ format (Measures # 3, 6-7, 13, 18-19, 29-31, 34; Table 4.2). Two questionnaires
had a ‘true’ or ‘false’ response options (Measures # 25, 38; Table 4.2) and one
questionnaire had a dichotomous ‘yes’ or ‘no’ response option (Measures # 23; Table 4.2).
Other response options were seen in 13 questionnaires (Measures # 1-2, 4, 10-11, 16, 21-
22, 24-27, 32; Table 4.2), and two questionnaires did not report their response options
(Measures # 8, 20; Table 4.2).
61
Identification
Records identified through database Additional records identified from

searching (n=3891) other citations (n=12)
Screening
Records remaining following removal of duplicates (n=2330)
Records screened (n=2330) Records excluded (n=2253)

⋅ Did not fit inclusion criteria
Eligibility Full-text articles excluded with reasons

(n=39)
⋅ 11 parental feeding practices measures
Full-text articles assessed for eligibility ⋅ 8 did not contain original questionnaire
(n=77) ⋅ 5 did not add to the robustness of the
questionnaires
⋅ 4 food frequency questionnaire measures
⋅ 2 weight control and dieting measures
⋅ 1 no questionnaire/scale produced
⋅ 1 eating disorder syndrome measure
⋅ 1 measure of perceived norms, barriers,
value of health, intentions, assessment of
eating and sedentary behaviour
⋅ 1 measure for children with chronic illnesses
-Davies
⋅ 1 measure of parental affect towards child
feeding
⋅ 1 measure of the extent to which individuals
might try to control or change urges or
cravings
⋅ 1 measure of behaviour disorders in children
⋅ 1 measure of perceived body image
Included ⋅ 1 article in Japanese
Studies included in qualitative synthesis (n=38)
Figure 4.1 Flow chart of studies in review (based on PRISMA 2009 flow diagram).
62
Table 4.2 Characteristics of studies included in the systematic review

Reference Sample Age Gender BMI kg/m2 Statistical
# Aim Response option
(Country) composition (mean±sd) M/F (mean±sd) analysis/Factors/Items
Emotional Eating Scale (EES)
(Arnow, Kenardy, & S1: to develop the item pool and investigate S1: 47 obese S1: 23 to 64 S1: 0/47 S1: 26.1 to 51.7 5-point scale S1: PCA: 3 factors (25 items)
Agras, 1995) psychometric properties of a questionnaire to females (44.9±10.4) S2: 0/51 37.9±6.0 ”no desire to eat”, “a Anger/Frustration (11 items)
permit analysis of the relationship between S2: 51 obese S2: 21 to 65 S2: 26.6 to 55.8 small desire to eat”, “a Anxiety (9 items)
negative mood and disordered eating females 45.1±10.6. 38.9±7.2 moderate desire to Depression (5 items)
USA eat’’, ”a strong desire Cronbach α’s internal and test-
S2: to assess the construct, discriminant, and to eat”, “an retest reliability.
1
criterion validity against the TFEQ, BES and BDI, overwhelming urge to S2: PCA, association between
SCL-90-R, RSE measures. eat,” EES and TFEQ (and other
measures) for convergent and
S3: to assess the discriminant validity by discriminant validity
administering it to a group of subjects diagnosed
with an anxiety disorder (not reported)
Emotional Eating Scale for use in Children and Adolescents (EES-C)
(Tanofsky-Kraff et al., To adapt the EES (Arnow, 1995) for use in children 59 overweight 8 to 18 yo S1: 56.6%/44.4% S1: BMI-DS 5-point scale: PCA: 3-factores (23 items)
2007) and adolescents (EES-C) assessed in two samples: 100 non- (14.3±2.4) LOC (1.6±0.9) ‘‘I have no desire to Anxiety, anger, and frustration
S1: LOC overweight S1: (13.1±2.7) S2: 46%/54% no LOC eat’’, through ‘‘I have a (EES-C- AAF) (12 items);
USA S2: No LOC 64 test-retest LOC LOC S2: : BMI-SDS very Depressive symptoms (EES-C-
To assess convergent, discriminant and test-retest S1: 18 LOC S2: (14.4±2.3) no (1.0±1.1) No LOC strong desire to eat.’’ DEP) (7 items); Feeling unsettled
reliability. S2: 137 no LOC LOC (EES-C-UNS) (4 items)
5-point scale: ‘‘On Cronbach α’s internal reliability,
2 average, how many test-retest 3.4±2.6 month
days a week do you eat interval.
because you feel this Convergent validity: ANCOVA
way?’’ between LOC versus No LOC.
Discriminant validity: partial
correlations between EES-C
subscales and measures of
general psychopathology.
Eating Identity Type Inventory (EITI)
(Blake, Bell, To assess how different eating identity types are 968 adults 57.2 ± 14.5 21.6/79.4 N/A 5-point scale: CFA: revealed 11/12-items:
Freedman, related to dietary intake. 903 CFA ‘‘strongly agree’’ , to RMSEA (.070), CFI (.937), NNFI
Colabianchi, & Liese, To assess the construct validity of the EITI using 94 retest ‘‘strongly dis-agree’’ (.925), and SRMR (.058)
2013) CFA. reliability 4 – factors (11 items)
To establish the convergent validity (against Healthy; Meat; Picky; Emotional
USA dietary intake measures) and internal and test- (number of items not reported).
3
retest reliability of EITI Cronbach α’s internal and test-
retest reliability.
63
Convergent validity: by assessing

the hypothesized degree to
which each eating identity type
(healthy, emotional, picky, and

meat) corresponded with each
dietary intake measures.
Pearson’s correlation
coefficients to assess test–retest
reliability.
Palatable Eating Motives Scale (PEMS)

(Burgess, Turan, To identify individual motivations for eating tasty 150 College 17 to 60 44/106 16.4 to 51.0 5-responses to choice PCA: 4-factor (19 items)
Lokken, Morse, & foods and to determine if certain motives would students (Mean 24.4) (Mean 26.3) frequency items. Social (5 items)
Boggiano, 2014) be associated with BMI. Coping (4 items)
Enhancement (5 items)
4 USA Conformity (5 items)
Cronbach α’s internal reliability.
Convergent validity: Partial
correlation coefficients with BIS,
BAS, YFAS and BES.
Palatable Eating Motives Scale for kids (K-PEMS)
(Boggiano, Wenger, To provide a preliminary validation of the K-PEMS, 73 African 12 and 17 48%/52% BMI-SDS 5-point scale: PCA: 4 Factors (19 items)
Mrug, Burgess, & a self-report survey to identify individual motives American (14.7±0.9) (0.84±1.1) “Never/almost never” Social (5 items), Conformity (5
Morgan, 2015) for eating tasty foods in adolescents, for early adolescents to “Almost items), Reward Enhancement (5
identification of obesity and binge-eating risk. always/always” items), Coping motives (4 items).
USA To determine if any specific motive(s) can account Cronbach α’s internal reliability.
5 for variance in BMI and binge-eating disorder (C- Linear regressions between the
BED) (Risky Eating) traits which can exacerbate K-PEMS motives and BMI-SDS
obesity and Risky Eating.
Binary logistic regressions tested
associations between K-PEMS
motives and C-BEDS.
Power of Food Scale (PFS)
(Cappelleri, To examine the factor structure of the PFS from: S1: 1741 obese S1: 46.3±11.0 S1: 314/1427 S1: 38.6±6.7 5-point scale: S1: EFA: 21-item; CFA: 15-item;
Bushmakin, Gerber, S1: baseline pre-treatment data of phase 3 clinical adults S2: 52.5±12.8 S2: 39% women S2: 33.1±7.6 “do not agree at all” to CFI (0.95), PNFI (0.78), ECVI
Leidy, Sexton, trial candidates for weight loss (including non- S2: 1275 adults “strongly agree” (0.48)
Karlsson, et al., 2009) obese, overweight and obese subjects) Web-based survey: CFI (0.94)
S2: Web-based survey: US arm of the 2006 Cronbach α’s internal and test-
USA - Canada National Health and Wellness Survey (NHWS) retest reliability.
S2: – 15 items 3-F:
6 Food readily available in the
environment but not physically
present, Food present but not
tasted, and Food when first
64
tasted but not consumed

(number of items not reported).
Cronbach α’s internal and test-
retest reliability.

State and Trait Food-Cravings Questionnaires (FCQ-S and FCQ-T)
(Cepeda-Benito, To develop, validate, and cross-validate 2 S1: S1: S1: 34%/66% N/A 5-point scale. Different S1: CFA: FCQ-T: χ2[593] = 204, p
Gleaves, Williams, & inventories for food cravings: The Food Cravings 217 psychology 18 to 44 S2: 30/74 options: < 0.001; GFI=0.80; NFI=0.99;
Erath, 2000) Questionnaire-State (FCQ-S) and the Food Cravings students –last (21±2.98) S3: 169/121 “Never”, “rarely”, TLI=1.0; CFI = 1.0; RMSEA = 0.37
Questionnaire- Trait (FCQ-T). 100 test-retest S2: “sometimes”, “often”, FCQ-T: 9-Factors (37-Items)
USA S1: Confirmation of factor structure, test-retest S2: 19 to 27 “usually”, “always” Intention to consume food (3
and internal reliability 104 students in (21±1.25) items), Anticipation of positive
S2: Convergent and discriminant validity by (i) an elective S3: “strongly disagree”, reinforcement (5 items), Relief
comparing food deprivation versus food satiation psychology 17 to 33 “disagree”, “neutral”, from negative states (3 items),
(not reported); (ii) FCQ-T/TFEQ course (19.3±1.85) “agree”, and “strongly Lack of control over eating (6
S3: Cross-validation and CFA S3: 290 agree”. items), Preoccupation with food
psychology (6 items), Hunger (4 items),
students Emotions (4 items), Cues that
trigger cravings (4 items), Guilt
(3 items)
S2: CFA: FCQ-S: χ2[80] = 206,
p<0.001; GFI=0.89; NFI=0.98;
7
TLI=0.99; CFI = 0.99;
RMSEA=0.72
FCQ-S: 5-Factors (15-Items)
Intense desire to eat (3 items),
Anticipation of positive
reinforcement (3 items), Relief
from negative state (3 items),
Lack of control over eating (3
items), Hunger (3 items)
Cronbach α’s internal test-retest
reliability.
Convergent and discriminant
validity: Correlations between
FCQ-T and TFEQ
S3: CFA: (39-item0 (2 additional
emotions items) – confirmed
factor structure in S1
Brief version of the Food Craving Questionnaire-Trait (FCQ-T-r) (FCQ-T-r)
(Meule, Hermann, & To develop and validate a short form of the FCQ-T S1: 323 S1: 24.4±5.6 S1: F=271 S1: 22.0±3.4 Not reported (taken S1: PCA – one-factor (15-items)
Kübler, 2014) the FCQ-T-r: S2: 70 S1: 22.0±3.3 S2: F=70 S1: 21.5±2.8 from FCQ-T, German Cronbach α’s internal reliability.
8 S1: Factor structure – online questionnaire version [Meule, 2012]) S2: Pearson correlation
Germany S2: Working memory task of highly palatable foods coefficients with BMI and RS.
(not reported) and RS.
65
General index of food craving (G-FCQ-T and G-FCQ-S)

(Nijs, Franken, & To assess the factor structure, validity and S1: (i) 227 (G- S1: (i) 17 to 28 S1: (i) 39/188; (ii) - 5-point scale: “never” S1: PCA:
9 Muris, 2007) reliability of the modified questionnaires FCQ-T) and (ii) (19.98±2.2); (ii) 30/89 to “very often” G-FCQ-T: 4-factors (21 items)
(G-FCQ-T and G-FCQ-S), 119 (G-FCQ-S) 17 to 28 S2: 35/170 Preoccupation with food, Loss of

Netherlands Cross-sectional: psychology (19.98±2.3) control (once eating), Positive
S1: To test construction and EFA students S2: 17 to 41 outcome expectancy (from
S2: CFA, test-retest reliability and construct S2: 205 (19.86±3.2) eating), Emotional craving
validity (DEBQ). psychology (number of items per subscale
S3: Construct validity against experimental students - Test- not reported).
measures of satiety meal conditions (not retest (50
reported). students) G-FCQ-S: 5-factor (15 Items)
Desire to eat, Anticipation to
positive reinforcement,
Anticipation to negative
reinforcement, Obsessive
preoccupation, raving as a
physiological state (number of
items per subscale not
reported).
Cronbach α’s internal reliability,
ICC test-retest reliability,
Pearson correlations coefficients
to assess validity against DEBQ.
S2: CFA: G-FCQ-T: χ2/df=2.44,
TLI=0.86, CFI=0.88; RMSEA=0.08.
G-FCQ-T: χ2/df=2.44, TLI=0.86,
CFI=0.88; RMSEA=0.08.
Control of Eating Questionnaire (CoEQ)

(Dalton et al., 2014) To assess the severity and type of food cravings an S1: 80 S1: 18 to 54 S1: 26/54 S1: 18.5 to 37.7 Participants responded Originally CoEQ contained 21
individual experiences over the previous 7 days. S2: 50 (26.5±8.1) S2: 0/50 (24.2±4.3) about their Items – six sections
UK To examine the psychometric properties and S3: 30 S2: 18 to 41 S3: 0/30 S2:18.6 to 39.8 experiences over the PCA: 4 factors (17 items)
underlying component structure in 4 samples S1, S4: 55 (24.3±5.9) S4: 18/37 (27.1±5.4) last 7 days: Craving Control (5 items)
S2, S3 and S4. Total sample: 215 S3: 20 to 54 Total sample: S3: 18.8 to 29.1 19 Items (VAS) Positive Mood (4 items)
10 To examine construct validity by exploring (27.8±10.5) 20%/80% (23.2±2.9) Items 20 and 21 Craving for Savoury (4 items)
associations with body composition and TFEQ (S1, S4: 20 to 55 S4: 26.1 to 39.7 allowed for own Craving for Sweet (4 items)
S2, S3), and BES (S1, S2, S3 and S4). (41.0±8.7) (24.3±5.9) response. Cronbach α’s internal reliability.
Total sample: Total sample: Construct validity through
29.7±10.3 26.4±5.2 Pearson correlation coefficients
with TFEQ and BES scales.
Emotional Appetite Questionnaire (EMAQ)
(Geliebter & Aversa, To examine a wide array of both negative and 90/364 Underweight Underweight Underweight 9-point scale: ANOVA was used to analyse
2003) positive emotions and situations in relation to not questionnaires 29.2±9.6 15/15 18.9±1.4 responses for emotions and
only overweight and normal-weight but also to stratified by sex, ‘‘much less’’ and situations, with weight category
USA underweight individuals. and for each Normal Normal Normal ‘‘much more’’ as and gender as group factors,
11
66
Authors predicted that overweight individuals gender, the 15 28.9±6.6 15/15 22.2±0.80 anchors and 5 indicates followed by LSD post-hoc tests.
would tend to overeat, whereas underweight most overweight, ‘‘the same’’ EMAQ (22 items):
individuals would tend to under-eat, in response the 15 most Overweight Overweight Overweight Tendency to eat in response to
to both positive and negative emotions and underweight, and 33.5±11.2 15/15 29.8±2.7 For each item, there is positive and negative emotions

situations. the 15 closest to also the option to (14 items)
desirable body indicate ‘‘not To positive and negative
weight for height applicable’’ or ‘‘don’t situations (8 items).
know’’ The positive emotion (EMAQ-PE)
and positive situation (EMAQ-PS)
scores can be averaged to obtain
a positive EMAQ score (EMAQ-
P).
The negative emotion (EMAQ-
NE) and negative situation
(EMAQ-NS) scores can also be
averaged to obtain a negative
EMAQ score (EMAQ-N).
Cronbach α’s internal
consistency and test–retest
reliability with Pearson’s
correlations for subscales of the
questionnaire.
Motivation for Eating Scale (MFES)
(Hawks, Merrill, Gast, To develop and validate items for the Motivation 298 (156 college CM were older 20/224 245 normal 5-point scale: PCA: 4 factors: Environmental
& Hawks, 2004) for Eating Scale (MFES) as possible contributors to students [CS], than CS (35.5 vs. weight “almost”, “never”, eating (23 items)
obesity. 142 community 24.8 years, p<.01) 53 overweight “sometimes”, almost”, Emotional eating (12 items),
USA members [CM]). “always”. Physical eating (9 items), Social
To assess internal and test-retest reliability, as well eating (5 items)
12
as concurrent and convergent validity (TFEQ and Test retest Cronbach α’s internal and test-
the EES and BMI) reliability and retest reliability. Pearson
convergent correlation coefficient between
validity (n=103) scores on MFES and TFEQ, and
BMI.
Intuitive Eating Scale (IES-H)
(Hawks, Merrill, & To develop items and validate an instrument 391 college 20.6±3.4 227/162 - 5-point scale: “strongly PCA: 4-factors (27 items)
Madanat, 2004) designed to measure the concept of intuitive students M (21.1±2.7) agree” to “strongly Intrinsic eating (4 Items),
eating. Test-retest F (19.9± 4.2). disagree” Extrinsic eating (6 Items), Anti-
USA To test internal and test-retest reliability, as well (n=285) dieting (13 Items), Self-care (4
13 as concurrent and convergent validity against the Items).
CBDS Cronbach α’s internal and test-
retest reliability. Logistic
regression to assess convergent
validity with CBDS.
Mindful Eating Scale (MES)
(Hulbert-Williams, et Development of a self-report scale to measure 127 university 25.65 ± 8.89 23.8%/77.2% 23.59±3.54 4-point scale: EFA: 6-factors (28 items)
67
al.,, 2013) mindfulness with respect to eating behaviours. students “never”, “rarely”, Acceptance (6 items), Awareness
14 “sometimes”, “usually” (5 items), Non-reactivity (5
UK To explore the MES against other measures of items), Act with Awareness (4
mindfulness and body acceptance. items),

Routine (4 items) and
Unstructured eating (4 items)
Cronbach α’s internal reliability
calculated by mean inter-item
correlations. Pearson’s
correlation coefficient with
mindfulness and body
acceptance quest.
Chinese Pre-schoolers’ Eating Behaviour Questionnaire (CPEBQ)
(Jiang et al., 2014) To develop a questionnaire which can be used to S1: 313 children S1: 3 to 6 S1: S2: 5-point scale: S1: EFA
evaluate Chinese pre-schoolers’ problematic S2: 603 children (4.3 ± 1.4) 161/152 OW (12.6%) “never”, “seldom”, S2: CFA
China eating behaviours. S2: S2: OB “sometimes”, “often”, (NFI=0.88, NNFI=0.91, CFI=0.92,
S1: To assess the factor structure 3yo (21.7%) 322/281 (10.9%) “always”. RMSR=0.04, and SB-x2/df=1.79).
S2: To confirm the factor structure through CFA, 4yo (23.4%) 7-factors (38 items)
and assess reliability, convergent and discriminant 5yo (27.7%) Food fussiness (7 items), Food
validity 6yo (27.2%) responsiveness 6 items), Eating
habit (5 items), Satiety
15 responsiveness (5 items),
Exogenous eating (5 items),
Emotional eating (5 items),
Initiative eating (5 items)
Cronbach α’s internal and split-
half test-retest reliability.
Pearson’s correlations analysis
was used to evaluate content
validity and construct validity.
Food Situations Questionnaire (FSQ)
(Loewen & Pliner, To develop and validate of a self-report measure S1: 125 children S1: 5 to 12 Not reported - 4-point scale: S1: EFA: 2-factors (10 items)
2000) of food neophobia for children. S2: 335 children S2: 7 to 12 “very happy” , “ok”, S2: Addition of 12 filler items.
To validate the FSQ against measures of “so-so”, “very sad” EFA: 2-factors (10 items):
Canada willingness to try new foods under laboratory Willingness to try novel foods in
conditions, and parent-report measures of their highly stimulating circumstances
child’s neophobia. (HI-STIM) (5 items), Willingness
to try novel foods in non-
16 stimulating circumstances (LO-
STIM) (5 items).
consistency and test-retest
reliability. Correlation
coefficients of FSQ and
behavioural tasks.
68
ecSatter Inventory (ecSI)

(Lohse et. al., 2007) To assess validity of the ecSatter Inventory (ecSI) 370 – on-line 18 to 71 172/644 F (n=631) 5-point scale: EFA and CFA: 4-factors (16
17
to measure eating competence (EC). survey (36.2±13.4) 25.7±6 “always”, “often”, items):

USA 462 - paper M (n=172) “sometimes”, “rarely”, Eating attitudes (5 items), Food
To assess construct validity against other version. 27.0±4.9 “never” acceptance (3 items), Internal
measures (TFEQ,EDI, FPS) 832/863 usable regulation (3 items), Contextual
surveys skills (5 items)
Meaning of Food Questionnaire (MOF)
(Ogden, Liakopoulou, To assess beliefs about food and the role that S1: 451 S1: 37.8±11.1 S1: 6/444 S1: 5-point scale: S1: EFA – Oblimin rotation - 5-
Antilliou, & Gough, these play and to evaluate the effectiveness of S2: 170 S2: 20.1±4.1 S2: 84/86 Normal weight “totally disagree” to Factors
2012) interventions designed to change some of these 17.1% (n=77) “totally agree” S2: EFA – oblimin rotation - 6-
dimensions or in a clinical setting to help health Overweight following the Factors (8 subscales) 25 items:
UK professionals explore clients’ relationships to food. 35.8% (n=161) statement: Food and sex (3 items); Control
S1: Dieters Obese 47.1% “to what extent do you over life (6 items); Control over
18 S2: University students (n=212) agree with the food (6 items); Food and family
following…. “ (3 items); Food as a treat (2
items); Food and emotional
regulation (2 items); Food and
guilt (3 items); and Food and
social interaction (1 items).
Food Neophobia Scale (FNS)
(Pliner & Hobden, To examine this neophobia-neophilia continuum in S1: 21, 55, 2 S1: 18 to 74 18 to 49 (M=22.6) - 7-point bipolar rating S1: Inter-rater correlations
1992) humans. S2: 135, 75 (M=20.7) scale: “disagree revealed a 18 relating to food
S3: 41, 35, 80 The majority strongly” to “agree neophobia; 12 items measured a
Canada To develop a paper and pencil measure of food were between strongly” more general neophobia
neophobia and to examine some of the correlates the ages of 19 (General Neophobia Scale [GNS])
of neophobia as assessed by this measure. and 25. S2: Uncorrected item-whole
S1: Construction of the FNS scale correlations for each sample: 10
S2: Psychometric analysis of FNS and GNS item:
S3: Behavioural validation (food tastings). 5 positive (neophilic) and 5
19 Convergent and discriminant validity against negative (neophobic) statements
Fear/Anxiety, Foreign food familiarity, Finickiness, about food or situations related
and sensation seeking measures (not reported). to food consumption. Cronbach
α’s internal consistency and
test–retest reliability.
S3: Subject ratings of familiarity
of foods, averaged across foods
(not reported).
Correlations between FNS and
GNS.
Food Neophobia Scale for children (FNS-C)
(Pliner, 1994) To adapt behavioural (not reported) and paper 117 5, 8 and 11 year Age 5, M=7; Age - Not reported. Parent-report FNS, validated
and pencil trait measures to study food neophobia old children 5, F=18; Age 8, against their child’s behavioural
69
Canada in children; paired with corresponding parent’s M=20; measurements (10 items)
20
prediction of their child’s willingness to try familiar Age 8, F=13; Age 5 positive (neophilic) and 5
and unfamiliar foods and overall neophobic 11, M=19; negative (neophobic) statements
behaviour. Age 11, F=22 about food or situations related

to food consumption
Italian Food Neophobia Scale for children (ICFNS)
(Laureati, The aim of the present study was to develop and 491/594 6 to 9 7.9 ± 1.0 303-291 - 5-point facial (8 items) (4 neophobic and 4
Bergamaschi, & validate a self-report measure of food neophobia years expression scale: ‘‘Very neophilic items).
Pagliarini, 2015) designed for Italian primary school children by false for me,’’ ‘‘False Cronbach α’s internal
adapting the ICFNS. for me,’’ ‘‘So-so,’’ consistency and test-retest
21
Italy Validity of the questionnaire was assessed through ‘‘True for me,’’ ‘‘Very reliability - Mean values for each
behavioural measurements true for me’’ item in the test–retest
evaluation were compared
through paired t-tests (p < 0.05).
Overeating Tension Scales (OTS)
(Popkess-Vawter, To develop an Overeating Tension Scale (OTS), S1: 373 S1: 26 ± 9 S1: 201/172 S1: normal 10-point continuum: S1: 8 subscales (Semantic
Gerkovich, & Wendel, derived from Apter's Reversal Theory, to measure S2: 208 S2: 27 ± 10 S2: 197/111 weight 43%; “how they were feeling differential scales) from 48 to
2000) overall reported tension and motivation-specific S3: 330 S3: 37 ± 13 S3: 82/248 overweight; just before over- 32-Items. Internal consistency
tension. S4: 130 S4: 35 ± 7 S4: 0/130 obese 44%; eating” (X), “how they S2: 32 Items revised. Internal
USA S1: Item reduction testing Underweight 13% wanted to feel” (O). consistency. Pearson’s
S2: Content validity and internal consistency S2: normal A discrepancy score (D) correlations
S3: Testing contrast validity using contrasted weight 40%; equivalent to tension S3: Cronbach α’s internal
groups (social gatherings, college enrolment and overweight; (O – X = D) consistency. Pearson’s
examination). obese 48%; correlations to assess validity.
22
S4: Testing OTS in normal weight and overweight underweight 12% S4: EFA: 7 factors: serious;
women (included BULIT and MCSDS) playful; compliant; defiant; self-
S4: 62 normal centered mastery; self-centered
weight and 68 sympathy; other-centered
overweight. sympathy
consistency. Pearson’s
correlations to assess
convergent validity (BULIT)
Eating in Emotional Situations Questionnaire (EESQ)
(Rollins et al., 2014) To describe the frequency of eating in emotional 159/184 low- 11 to 17 (age: 45%/57.9% - Response option: “no”, CFA – 2-factors (11 Items):
situations (EES) among a sample of low-income income Latino M=9.4, SD=.6). “yes”. (χ2=45.05, p=.39; CFI=.999;
USA Latino elementary-school children fourth graders. RMSEA=.017).
A limited sample F1 – (6-items): Eating in

completed the response to psychological
external eating distress (e.g. anxiety).
23 (n=70) and junk F2 – (5 items): Triggered by
food (n=89) contextual cues (e.g. receiving a
subscales. bad grade).
70
consistency. Criterion validity of

the EESQ was evaluated by
correlating the EESQ scales with
the food frequency and eating

behaviour measures, stratified
by gender.
Eating Pattern Inventory for Children (EPI-C)
(Schacht, Richter- To evaluate and present the factor structure, 373 children 8 (2 children) 168/205 12.8 to 29.5 (17.9 4-point scale: 6 factors (39 items) were taken
appelt, & Schulte- psychometric properties, and initial validation data 9 (53.1%) ±2.8) “not at all”, “little” from the Problems Inventory for
markwort, 2006) of a new self-report questionnaire on 10 (42.1%) “mostly”, “totally” Children (EWI-C) (Diehl, 1999);
psychological dimensions of eating behaviour in 11 (14 children) Underweight 32 for 11 to 14 year olds:
Germany children. (8.6%) EFA: 4-factors (20 items)
Based on the Eating Behaviour and Weight Dietary restraint (8 items)
Overweight 57 External eating (5 items)
24 (15.3%) – 17 Parental pressure to eat (3
(4.6%) of these items)
were obese. Emotional eating (4 items)
consistency. Pearson’s
correlation coefficients between
scales and BMI-SDS
Three Factor Eating Questionnaire (TFEQ)

(Stunkard & Messick, To construct a measure that describes three S1: Restrained S1: 17 to 77 S1:18/60 S1: Restrained Different response S1: EFA on 67 items revealed 3-
1985) dimensions of human eating behaviour: (78), (44±12.8) Restrained, 22/40 eater 50% normal options: factors (57 items). Cronbach α’s
Cognitive restraint (cognitive and behavioural unrestrained (62) (combined) unrestrained, weight 15- response scale inter-scale reliability and inter-
USA aspects of controlling food intake), Disinhibition or intermediate 57/23 50% obese 6 true/false items. correlations. Correlations with
(susceptibility to emotional and social cues), eaters (80) intermediate weight.
Hunger (eating when hungry) Two ancillary S2:7/46 and “True/False” S2: EFA on 93 items revealed 3-
samples 52 and 5/13/27didn’t factors (58-items). Cronbach α’s
28. record gender “Rarely”, “sometimes”, inter-scale reliability and inter-
25
S2: 53 evangelical S3: Not reported “usually”, “always” correlations.
weight program, S3: EFA on 58 items revealed 3-
45 free eater “Not at all”, “slightly”, factors (51-items): Cognitive
S3: combined “moderately”, “very restraint (21 Items), Disinhibition
sample of 98 much”. (16 Items), Hunger (14 Items).
cases (dieters Cronbach α’s inter-scale
[n=53], free Etc. reliability and inter-correlations
eaters [n=45])
Three Factor Eating Questionnaire revised version-TFEQ-R18
(Karlsson et al., 2000) To evaluate the construct validity of the TFEQ in 4377 obese 37 to 57 1774/2603 M: 38.3±4.6 Different response Multi-trait/multi-item analysis
large samples of obese men and women. participants od (46.5±5.9) F: 41.2±6.0 options: (using EFA): 3-factors (18 Items):
Sweden To test if more efficient scales could be Swedish Obese “Definitely true”, Cognitive restraint (6 items)
constructed by item reduction. Subjects Study “mostly true”, “mostly Disinhibition and Hunger were
26 (SOS). false”, “definitely false” grouped into Uncontrolled
71
Eating (9 Items)
Two samples “almost never”, Emotional Eating (3 Items)
(2193, 2184) “seldom”, “usually”, Cronbach α’s internal reliability.
“almost always” Pearson’s correlations with BMI

and between scales to assess
“Unlikely/slightly”, convergent and discriminant
“likely/moderately”, validity
“likely/very likely”
Three Factor Eating Questionnaire revised version TFEQ-R21 _TFEQ-R18-V2
(Cappelleri, To evaluate the factor structure and reliability of S1: 1741 obese S1: S1: 18%/82% S1: 4-point scale for items S1: CFA in clinical sample (no
Bushmakin, Gerber, the TFEQ-R21, and its association with BMI, in a non-diabetic 46.3±11.0 S2: 61%/39% 38.6±6.7 1–20 (different constraint model). CFI: 18-Item
Leidy, Sexton, Lowe, large obese clinical sample from the United States S2: 1275 web- S2: 52.5±12.8 S2: response options) model (0.91) best fit (TFEQ-
et al., 2009) and Canada. survey 33.1±7.6 R18V2 Cronbach α’s internal
To modify the structure of the TFEQ-R21, if 8-point numerical reliability and Pearson’s
USA warranted, using the clinical sample and then test rating scale for item correlation coefficient between
Canada the refined model in a web-based sample of obese 21. BMI and TFEQ- R18V2.
and non-obese healthy individuals from the United S2: Same analysis- Web-based
27 States sample (constrained model). CFI:
(0.96)
Cognitive Restraint (6 items),
Uncontrolled Eating (9 items),
Emotional Eating (6 items).
Cronbach α’s internal reliability
and Pearson’s correlation
coefficient between BMI and
TFEQ- R18V2
Eating in the Absence of Hunger (EAH-C)
(Tanofsky-Kraff et al., To develop an Eating in the Absence of Hunger 153 obese 6 to 19 yo Non-obese 23% obese 5-point scale: PCA – 3-factors (14 items):
2008) Questionnaire to be administered to children and 73 non-obese (14.4±2.5) 48%/52% 68% non-obese ‘‘Never’’ through to Negative Affect (6 items),
adolescents (EAH-C) and to examine its Non-obese Obese ‘‘Always’’. External Eating (4 items),
USA psychometric properties. 14.2±2.5 53.8%/56.2% Fatigue/Boredom (4 items).
28 Cross-sectional Obese 14.9±2.4 Cronbach α’s internal and test-
retest reliability. Convergent and
discriminant validity (against
measures of depression and
anxiety).
Intuitive Eating Scale (IES)
(Tylka, 2006) To develop and psychometrically evaluate of a 1260 college S1: 17 to 61 S1: 0/391 S3: 17.5 to 34.9 5-point scale: S1: EFA – 3-factors (25 items),
measure of Intuitive Eating (IES) [(a) unconditional students (20.85±6.21) S2:0/476 (23.50±3.90) “strongly disagree”, Factor 1 (11 items), Factor 2 (8
USA permission to eat when hun-gry and what food is S1: 391 women- S2: 17 to 50 S3: 0/199 Self-reported “disagree”, “neutral”, items), Factor 3 (6 items).
desired, (b) eating for physical rather than EFA (19.70±4.5) S4: 0/194 weight and “agree”, “strongly Cronbach α’s internal reliability
emotional reasons, and (c) reliance on internal S2: 476 women S3: 17 to 55 height agree”. and construct validity with (EAT-
hunger and satiety cues to determine when and college students- (18.9±3.3) 26).
29
how much to eat] CFA S4: 17 to 55 S2: CFA – 3-factors (21 Items)
S3: 199 women (22.1±7.38) CFI=0.91, TLI=0.90, RMSEA=0.80,
72
(who knew about SRMR=0.07

study) Unconditional permission to eat
S4: 194 women (9 items); Eating for physical
rather than emotional reasons (6

items); Reliance on internal
hunger/satiety cues (6 items)
Internal reliability and construct
validity with other measures.
S3: Correlations of IES to BMI
S4: Test-retest reliability
Intuitive Eating Scale–2:
(Tylka & Kroon Van Item refinement and psychometric evaluation of S1: 878 S1: 18 to 56 S1: 391/487 Test- S2: F: 15.98 to 5-point scale: S1: EFA and CFA separate in M
Diest, 2013) IES with college women and men. Test-retest 219 (20.4±5.2) Test- retest 79/140 56.25 (24.0±5.7) “strongly disagree”, and F: 23-item IES-2 contained
S2: 1200 retest 18 to 47 S2: 520/680 M: 16.5 to 59.1 “disagree”, “neutral”, 11 original items and 12 added
USA S3: 522 (20.3±4.6) S3:284/238 (25.4±5.5) “agree”, “strongly items. 4 factors: 3 Original IES
S2: 18 to 53 agree” factor + Body–Food Choice
(20.5±5.1) Congruence.
S3: 18 to 56 Cronbach α’s internal reliability,
(20.3±4.8) contruct validity with IES, test-
retest reliability.
S2: CFA: Factor structure from S1
in M and F: CFI=0.96,
SRMR=0.06, RMSEA=0.05, 90%
CI [0.050, 0.057], χ2(206,
30 n=1200) = 908.31, p<0.001.
Unconditional permission to eat
(8 items); Eating for physical
rather than emotional reasons (6
items); Reliance on internal
hunger/satiety cues (6 items);
Body–Food Choice Congruence
(3 items)
Cronbach α’s internal reliability,
construct validity with eating
and body-related variables and
psychological wellbeing indices.
S3: Discriminant validity with
social desirability scales
Dutch Eating Behaviour Questionnaire (DEBQ)
(van Strein et al., To develop a questionnaire containing three S1: 120 subjects S1: M=(30.8±5.2) S1: 40/80 S1: M=(26.2± 5.4) 5-point scale: S1: Item pool development from
1986) scales: restrained eating, emotional eating and S2: (i) 264 F=(31.1±8.4) S2: (i) F=(25.2 ± 4.8) “never”, “seldom”, 100 items: PCA: 3-factors (51
external eating. (ii) 93 S2: (i) M=(23.6± M=103 (26.8±4.5) S2: (i) “sometimes”, “often”, Items).
Netherlands S1: Item pool development from 100 items. S3: 91 obese; 566 2.8) F=(22.9 ± F=161 M=(23.6±2.8) “very often” S2: PCA on 51 items
S2: To devise distinct scales for EE and ExtE and non-obese 4.1) (ii) M=(31.0± (29.9±4.7) F=(22.9 ± 4.1) (ii) administered to two samples: EE
31
administer to two samples. 8.3) F=(31.1 ± (ii) M=(31.1± 2.9) S3: full sample comprised 2 factors (clearly
73
S3: To develop a final item pool and also to assess S3: 8.6) F=(32.8 ± 6.2) reported in labelled emotions and diffuse
the dimensional stability of this item pool in sub- M=517 S3: full sample S3: full sample (Baecke et al., emotions).
samples of obese and non-obese subjects, and F=653 reported in reported in 1983) S3: From S2, items were revised
men and women, and then to replicate the factor (Baecke, Burema, (Baecke et al., and new items developed (48

structures obtained in the preliminary studies. Frijters, Hautvast, 1983) items). PCA: EE comprised 2
& van der Wiel- factors (same as above).
Wetzels, 1983) Construction of final scale (33
items):
Restrained (10 items), emotional
eating (13 items), and external
eating (10 items). Cronbach α’s
internal reliability. Descriptive
statistics and subscale Pearson’s
correlations coefficients.
Children’s Dutch Eating Behaviour Questionnaire (DEBQ-C)

(van Strein & To construct an age adapted version of the DEBQ S1: 769 children S1: 9.6±1.4 S1: 382/387 S1: 81.4% normal 3-point scale: S1: PCA on 37 items, revealed a
Oosterveld, 2008) for measurement of restrained, emotional and S2: 515 children – (7 to 12) S2: 252/263 weight “No”, “sometimes”, 3-factors (20 items).
external eating in 7- to 12-year-old children: the additional S2: 18.6% overweight “yes”. Cronbach α’s internal reliability.
Netherlands DEBQ-C validation B 9.30±1.44 S2: CFA: RMSEA=0.031, p=1.0,
S1: Item pool development. G 9.30±1.47 χ2(187) = 286, p<0.001,
S2: To determine the reliability, inter-correlations, χ2/df=1.71.
32 and correlations with other measures (e.g. other Restrained (7 items), emotional
risk factors for overweight such as snacking, (7 items), and external eating (6
skipping breakfast, physical inactivity, and time items)
spend with screen media, parental feeding styles Multi-group model for testing
and body dissatisfaction) construct invariance for BMI-
status.
Dutch Eating Behaviour Questionnaire parent version (DEBQ-P)

(Braet & van Strein, To assess eating patterns in children using the 292 children + 9 to 12 (10.5±0.9) Overweight Overweight 5-point scale: EFA: 3-factors (33 items)
1997) parent version of the DEBQ, the DEBQ-P, and focus parents 52/93 49.3%±19.7 “never”, “seldom”, Cronbach α’s internal reliability.
on obese and non-obese youngsters. 145 Overweight Normal weight Normal weight “sometimes”, “often”, ANOVA for each DEBQ scale: sex
Netherlands To explore specifically the relationships among EE, 147 Normal 58/89 ±9%IBW “very often” (male, female) by group (obese,
33
ExtE, and caloric intake. weight Items adapted to non-obese).
To test the relationship between EE and ExtE- parental report
induced eating against psychological measures of version.
emotionality and externality in children.
Hunger Sensitivity Scale (HSS)
(Walker, To develop and validate the hunger sensitivity S1: 556 university S1: 24.6±0.41 S1: 121/435 6-point scale: S1: Item analysis of the 29-item
Hadjistavropoulos, scale (a cognitive eating style associated with students S2: 22.4±5.76 S2: 15/86 (test- (0) “strongly disagree” scale. Parallel analysis, then EFA
Gagnon, & MacNab, heightened distress in response to hunger S2: 101 university retest 7/40) to (6) “strongly agree” (50% sample) and CFA (50%
2015) sensations) students on a sample): 1 factor (13 items)
S1: Conceptual grounds for item generation and diet or had been CMIN/df=2.01, RMSEA=0.6,
34
Canada factor analysis. on a diet (47/85 ECVI=0.66.
74
S2: Test-retest and discriminant validity (TFEQ plus test-retest) Cronbach α’s internal reliability.
other measures of general anxiety, depression and S2: Cronbach α’s internal and
anxiety sensitivity.) test-retest reliability.
Convergent and discriminant

validity (against TFEQ and
measures of depression and
anxiety).
Child Eating Behaviour Questionnaire (CEBQ)
(Wardle, Guthrie, et To develop and validate a questionnaire to assess S2: (i) 15 parents S2: (ii) 2 to 7 S3: 100/78 (4 no - 5-point scale: S2: (ii) PCA on 57 items revealed
al., 2001) eating style in young children through parental (2 to 6 yo); (ii) (4.2±1.3) gender indicated) ‘‘never’’, ‘‘seldom’’, 7-factors (35 items)
report. 131 parents S4: 5.6±1.5 S4: 111/97 ‘‘sometimes’’, ‘‘often’’, Cronbach α’s internal reliability.
UK S1: Evaluation of existing literature S3: 187 parents ‘‘always’’ S3: PCA: 7-factors (35 items):
S2: (i) Interviews with parents about their S4: 208 parents (scored 0–4). Food responsiveness (5 items);
children’s eating; (ii) Pilot study Enjoyment of food (4 items);
S3: Internal consistency in 2 samples of parents Test-retest: 160 Emotional over-eating (4 items);
parents Desire to drink (3 items); Satiety
35 S4: PCA to third sample of parents. Test-retest 2 responsiveness (5
weeks later. Gender and age differences were items)/Slowness in eating (4
analysed. items); Emotional under-eating
(4 items); Fussiness (6 items).
Cronbach α’s internal and test-
retest reliability. Correlations
between scales and age and
gender differences using
ANOVA.
Baby Eating Behaviour Questionnaire (BEBQ)
(Llewellyn, van To describe the development and factor structure S1: 33 mothers of S1: 2 to 24 S2: 1194/1208 Weight at birth: 5-point scale: S2: PCA – 4-factors (18 items):
Jaarsveld, et al., of the BEBQ an infant version of the CEBQ that twins months 2.5±0.55 “never”, “rarely”, Enjoyment of food (4 items);
2011) measures four appetitive traits in infants who are S2: 2402 infants S2: 4 to 20 “sometimes”, “often”, Food responsiveness (6 items);
still exclusively fed milk, related to weight. months (M: 8 “always” Slowness in eating (4 items);
UK S1: Development of the questionnaire. Pilot study months) Satiety responsiveness (4 items);
with 33 mothers of twins. plus (1 appetite item: ‘My baby
S2: Gemini Study – Assessment of the factor had a big appetite’)
structure Cronbach α’s internal reliability.
T-tests and ANOVAs were used
36 to assess group differences
across all of the scale. Pearson’s
correlation coefficient was used
to explore associations between
birth weight SDS and normally
distributed BEBQ scales,
Spearman’s r was used for
‘enjoyment of food’ and birth
weight SDS.
75
Self-report measure of the CEBQ for 13 year old adolescents (CEBQ-self-report)

(Loh et al., 2013) To adapt of the CEBQ as a self-report among S1: 362, test- 13 S1: IOTF cut-off 5-point scale: S1: CFA: 9-factors (35-item). FF
37 adolescents in a Malaysian population. retest n=133 M (59.7%) points “never”, “rarely”, was split into two.
Malaysia Cross-sectional – two phase study: S2: 646 S2: M=182(26.8), “sometimes”, “often”, Cronbach α’s internal and test-

S1: Construct validation F=464(73.2) S2: “always” retest reliability.
S2: Associations w/BMI Underweight=52( S2: CFA: 8-factors (30-item),
8.0%) Normal yielded an improved model fit.
weight=422(65.3 (χ2/df = 3.686, CFI =0.850, TLI =
%) 0.815, GFI = 0.736, AGFI = 0.773,
Overweight=104( RMSEA =0.065).
16.1%) Food responsiveness (5 items),
Obese=68(10.5%) Enjoyment of food (4 items),
Emotional over-eating (4 items),
Desire to drink (3 items),
Slowness in eating (4 items),
Emotional under-eating (4 items)
and Food fussiness 1 (FF1) (4
items); Food fussiness 2 (FF2) (4
items) (as two different
concepts; dislike towards food
(FF1) and trying new food (FF2).
Associations between eating
behaviour and BMI z-scores
were examined with complex
samples general linear model
(GLM) analyses, adjusted for
gender, ethnicity and maternal
educational level.
Flexible and Rigid Control Dimensions of Dietary Restraint

(Westenhoefer, 1991) To examine if restrained eaters in fact restrain 54,525 M 8393/ M Different response Cognitive restraint subscale of
food intake. participants in a 45.6±12.2 46132 28.2±3.2 options: the EI (Stunkard & Messick,
Germany S1: Participants were subdivided into 17 groups computer-aided F F “true”, “false” 1985) is divided into flexible and
according to their level of disinhibition. training program 43.6±12.7 27.2±3.8 rigid control strategies of dietary
S2: Examined whether there are distinctive types for weight “usually”, “always”, restraint.
of restrained eating behaviour, one associated reduction. “moderately”, “very S1: Mean scores for item on the
with high disinhibition, the other with low much”, “often”, RS were computed by subgroups
disinhibition of control. “always” of disinhibition and tested for
linear relation to the
38 disinhibition scores, and for
deviations from linearity.
S2: Test of linearity and
deviation between high and low
scores of disinhibition (ANOVAs)
From the results of discriminant
76
analysis, two ad hoc scales were

built from the restraint items
having the most discriminating
power:

Flexible control (12 items)
(FC12).
Rigid control (16 items) (RC16).
S1, S2, S3, etc. = Study 1, Study 2, Study 3
AGFI: Adjusted Goodness of Fit Index; CFA = Confirmatory Factor Analysis; CFI: Comparative Fit Index; ECVI: Expected Cross Validation Index;
EFA: Exploratory Factor Analysis; EPCA: Exploratory principal component analysis; GFI: Goodness-of-Fit Index; PNFI: Parsimonious Normed Fit Index;
RMSEA: Root Mean Square Error of Approximation; TLI: Tucker-Lewis Index; SEM = Structural Equation Modelling.
Mo = Mothers, Fa = Fathers; M=male, F=Female, B=boys, G=girls
EE: Emotional Eating; ExtE: External Eating; RE: Restrained Eating.
LOC: Loss of Control; No LOC: No Loss of Control.
AAQ: Acceptance and Action Questionnaire II; BAQ: Body Attitude Questionnaire; BAS: Behavioural Activation Scale; BDI: Psychological adjustment; BES: Binge Eating
Scale; BIS: Behavioural Inhibition Scale; BULIT: CBDS: Cognitive Behavioural Dieting Scale; EDE-Q: Eating Disorder Examination Questionnaire; EDI: Eating Disorder
Inventory;
FMPS: Frost Multidimensional Perfectionism Scale; FPS: Food Preference Survey; LOC: Loss of Control; MAAS: Mindfulness Attention Awareness Scale;
MCSDS: Marlowe-Crowne Social Desirability Scale; RSE: Self-esteem; SCL-90-R: Psychological adjustment; VAS=Visual Analogue Scale; WPI: Weight Problems Inventory;
YFAS: Yale Food Addiction Scale.
BEBQ: Baby Eating Behaviour Questionnaire; CEBQ: Child Eating Behaviour Questionnaire; CEBQ-self-report: Self-report measure of the CEBQ; CPEBQ:
Chinese Pre-schoolers’ Eating Behaviour Questionnaire; CoEQ: Control of Eating Questionnaire; DEBQ: Dutch Eating Behaviour Questionnaire; DEBQ-C:
Children’s Dutch Eating Behaviour Questionnaire; DEBQ-P: Dutch Eating Behaviour Questionnaire parent version; EAH-C: Eating in the Absence of Hunger;
ecSI: ecSatter Inventory; EES: Emotional Eating Scale; EESQ: Eating in Emotional Situations Questionnaire; EES-C: Emotional Eating Scale; EITI: Eating
Identity Type Inventory; EMAQ: Emotional Appetite Questionnaire; EPI-C: Eating Pattern Inventory for Children; FCQ-S and FCQ-T: State and Trait Food-
Cravings Questionnaires; FCQ-T-r and FCQ-T-r: Brief version of the Food Craving Questionnaire-Trait; FSQ: food Situations Questionnaire; FNS: Food
Neophobia Scale; FNS-C: Food Neophobia Scale for children; G-FCQ-T and G-FCQ-S: General index of food craving; HSS: Hunger Sensitivity Scale; ICFNS:
Italian Food Neophobia Scale for children; IES: Intuitive Eating Scale; IES-2: Intuitive Eating Scale-2; IES-H: Intuitive Eating Scale-H; K-PEMS: Palatable Eating
Motives Scale for kids; MES: Mindful Eating Scale; MOF: Meaning of Food Questionnaire; OTS: Overeating Tension Scales; PEMS: Palatable Eating Motives
Scale; PFS: Power of Food Scale; MFES: Motivation for Eating Scale; TFEQ: Three Factor Eating Questionnaire; TFEQ-R18: Three Factor Eating
Questionnaire revised version; Three Factor Eating Questionnaire revised version TFEQ-R21 _TFEQ-R18-V2.
77
4.4.2 Evaluation of psychometric properties: Reliability and validity of the

questionnaires
The majority of studies reported reliability and validity of their scales. Table 4.3 shows the
results of the psychometric evaluation of the questionnaires (see Appendix 4.3 for an
extended version of the psychometric evaluation of the questionnaires). All questionnaires
provided measures of internal reliability, and the majority carried out test-retest reliability,
except for 10 measures (Boggiano et al., 2015; Braet & van Strein, 1997; Cappelleri,
Bushmakin, Gerber, Leidy, Sexton, Lowe, et al., 2009; Dalton et al., 2014; Hulbert-Williams
et al., 2013; Karlsson et al., 2000; Llewellyn et al., 2011; Ogden et al., 2012; Pliner, 1994;
Popkess-Vawter et al., 2000; Schacht et al., 2006). The majority of questionnaires were
validated using convergent validity, except for five measures which used content validity
(Jiang et al., 2014; Llewellyn, van Jaarsveld, et al., 2011; van Strein et al., 1986; van Strein &
Oosterveld, 2008; Wardle, Guthrie, et al., 2001). One measure used criterion validity
(Rollins et al., 2014), and seven measures did not provide any validity results (Boggiano et
al., 2015; Cappelleri, Bushmakin, Gerber, Leidy, Sexton, Lowe, et al., 2009; Hawks et al.,
2004; Loh et al., 2013; Ogden et al., 2012; Schacht et al., 2006). Five questionnaires used
behavioural measures to validate the questionnaires: The Food Situations Questionnaire
(FSQ),the Food Neophobia Scale (FNS), the FNS-C in children, the Italian version of the FNS
(IFNC), and the CEBQ (Laureati et al., 2015; Loewen & Pliner, 2000; Pliner & Hobden, 1992;
Pliner, 1994; Wardle, Guthrie, et al., 2001). A total of 20 questionnaires did not measure
discriminant validity (Measures # 3, 5, 8, 10-16, 18, 20-24, 27, 32-33, 36; Table 4.3).
4.4.3 Overall robustness of the questionnaires
A total of 17 questionnaires obtained a 4-point score: the ‘Emotional Eating Scale’ (EES)
(Arnow et al., 1995), the ‘Emotional Eating Scale’ for children (EES-C) (Tanofsky-Kraff et al.,
2007), the ‘Palatable Eating Motives Scale’ (PEMS) (Burgess et al., 2014), the ‘Power of
Food Scale’ (PFS) (Cappelleri, Bushmakin, Gerber, Leidy, Sexton, Karlsson, et al., 2009), the
‘State and Trait Food-Cravings Questionnaires’ (FCQ-S and FCQ-T) (Cepeda-Benito et al.,
2000), the ‘General index of food craving’ (G-FCQ-T and G-FCQ-S) (Nijs et al., 2007), the
‘Emotional Appetite Questionnaire’ (EMAQ) (Geliebter & Aversa, 2003), the Intuitive Eating
Scale (IES-H) (Hawks et al., 2004), the ‘Food Neophobia Scale’ (FNS) (Pliner & Hobden,
1992), the TFEQ (Stunkard & Messick, 1985), the ‘Eating in the Absence of Hunger’
questionnaire (EAH-C) (Tanofsky-Kraff et al., 2008), a second ‘Intuitive Eating Scale’ scale
(IES) (Tylka, 2006), the ‘Intuitive Eating Scale–2’ (IES-2) (Tylka & Kroon Van Diest, 2013), the
78
DEBQ (van Strein et al., 1986), the ‘Hunger Sensitivity Scale’ (HSS) (Walker et al., 2015), the
CEBQ (Wardle, Guthrie, Sanderson, & Rapoport, 2001) and the ‘Flexible Control’ and ‘Rigid
Control’ dimensions of ‘dietary restraint’ (Westenhoefer, 1991). The majority of the
questionnaires obtained a score of 2 or 3 points (n=16). Five questionnaires obtained a 1-
point score (Boggiano, Wenger, Mrug, et al., 2015; Cappelleri, Bushmakin, Gerber, Leidy,
Sexton, Lowe, et al., 2009; Ogden et al., 2012; Pliner, 1994; Schacht et al., 2006).
The majority of the 17 robust questionnaires identified were adult measures, with only
three specifically for use in children: the ‘Emotional Eating Scale’ for children (EES-C)
(Tanofsky-Kraff et al., 2007), the ‘Eating in the Absence of Hunger’ questionnaire (EAH-C)
(Tanofsky-Kraff et al., 2008), and the CEBQ (Wardle, Guthrie, et al., 2001).
79
Table 4.3 Validity and reliability measures of questionnaires from the systematic review

Test-retest Convergent/content Psychometric
Reference Internal reliability Discriminant validity
reliability /criterion validity evaluation
1 (Arnow et al., 1995)
4
Emotional Eating Scale - Adapted for use in Children and Adolescents (EES-C)
2 (Tanofsky-Kraff et al., 2007
4
)
3 (Blake et al., 2013)
- 3
4 (Burgess et al., 2014) (Boggiano et al.,
2015) 4

5 (Boggiano, Wenger, Mrug,
- - - 1
Burgess, & Morgan, 2015)
6 (Cappelleri, Bushmakin, Gerber,
(Lowe et al., 2009)
Leidy, Sexton, Karlsson, et al., 4
2009)
7 (Cepeda-Benito et al., 2000)
4
Brief version of the Food Craving Questionnaire-Trait (FCQ-T) (FCQ-T-r)
8 (Meule et al., 2014) - - 2
9 (Nijs et al., 2007)
4
10 (Dalton et al., 2014)
- 4

11 (Geliebter & Aversa, 2003) (Nolan, Halperin, &
80
(Nolan et al., 2010)

Geliebter, 2010) 4
12 (Hawks, Merrill, & Madanat,
- 3

2004)
13 (Hawks et al., 2004) 4
14 (Hulbert-Williams et al., 2014) - - 2
15 (Jiang et al., 2014) - 3
16 (Loewen & Pliner, 2000) - 2
17 (Lohse et al., 2007) (Stotts & Lohse,
2007) - 3

18 (Ogden et al., 2012) - - - 1
19 (Pliner & Hobden, 1992) 4
20 (Pliner, 1994) - - - 1

21 (Laureati et al., 2015) - 3
22 (Popkess-Vawter et al., 2000)
- - 2
23 (Rollins et al., 2014)
- - 2

24 (Schacht et al., 2006)
- - - 1
25 (Stunkard & Messick, 1985) (Gormally, Black,
81
(Ganley, 1988) Daston, & Rardin,

4
1982)
Three Factor Eating Questionnaire revised version TFEQ-R18
26 (Karlsson et al., 2000)
- 3

Three Factor Eating Questionnaire revised version TFEQ-R21 _TFEQ-R18-V2
27 (Cappelleri, Bushmakin, Gerber,
- - - 1
Leidy, Sexton, Lowe, et al., 2009)
28 (Tanofsky-Kraff et al., 2008)
4
29 (Tylka, 2006) (Avalos & Tylka,
2006) 4
Intuitive Eating Scale–2

30 (Tylka & Kroon Van Diest, 2013)
4
Dutch Eating Behaviour Questionnaire (DEBQ)
31 (van Strein et al., 1986) (Banasiak, (Cebolla, Barrada,
Wertheim, Van Strein, Oliver, &
(van Strein, 2002)
Koerner, & Baños, 2014; J 4
Voudouris, 2001) Wardle, 1987a)
Children’s Dutch Eating Behaviour Questionnaire (DEBQ - C)

32 (van Strein & Oosterveld, 2008) (Baños et al., 2011)
- 3
33 (Braet & van Strein, 1997) (Caccialanza et al.,
2004) - - 2

34 (Walker et al., 2015) 4
35 (Wardle, Guthrie, et al., 2001) (Carnell & Wardle,
(Carnell & Wardle, 2007)
2007) 4

36 (Llewellyn, van Jaarsveld, et al.,
- - 2
2011)
82
Self-report measure of the CEBQ

37 (Loh et al., 2013)
- - 2
38 (Westenhoefer, 1991) (Westenhoefer et (Westenhoefer et al.,

al., 1999) 1999) 4
+ve: Positive significant associations with BMI or BMI-SDS

-ve: Negative significant associations with BMI or BMI-SDS
PA: Partial associations with BMI or BMI-SDS and only some of the sub-scales within the measure.
NA: No associations with any sub-scales within the measure.
BEBQ: Baby Eating Behaviour Questionnaire; CEBQ: Child Eating Behaviour Questionnaire; CEBQ-self-report: Self-report measure of the CEBQ; CPEBQ: Chinese
Pre-schoolers’ Eating Behaviour Questionnaire; CoEQ: Control of Eating Questionnaire; DEBQ: Dutch Eating Behaviour Questionnaire; DEBQ-C: Children’s Dutch
Eating Behaviour Questionnaire; DEBQ-P: Dutch Eating Behaviour Questionnaire parent version; EAH-C: Eating in the Absence of Hunger; ecSI: ecSatter
Inventory; EES: Emotional Eating Scale; EESQ: Eating in Emotional Situations Questionnaire; EES-C: Emotional Eating Scale; EITI: Eating Identity Type Inventory;
EMAQ: Emotional Appetite Questionnaire; EPI-C: Eating Pattern Inventory for Children; FCQ-S and FCQ-T: State and Trait Food-Cravings Questionnaires; FCQ-T-r
and FCQ-T-r: Brief version of the Food Craving Questionnaire-Trait; FSQ: food Situations Questionnaire; FNS: Food Neophobia Scale; FNS-C: Food Neophobia
Scale for children; G-FCQ-T and G-FCQ-S: General index of food craving; HSS: Hunger Sensitivity Scale; ICFNS: Italian Food Neophobia Scale for children; IES:
Intuitive Eating Scale; IES-2: Intuitive Eating Scale-2; IES-H: Intuitive Eating Scale-H; K-PEMS: Palatable Eating Motives Scale for kids; MES: Mindful Eating Scale;
MOF: Meaning of Food Questionnaire; OTS: Overeating Tension Scales; PEMS: Palatable Eating Motives Scale; PFS: Power of Food Scale; MFES: Motivation for
Eating Scale; TFEQ: Three Factor Eating Questionnaire; TFEQ-R18: Three Factor Eating Questionnaire revised version; Three Factor Eating Questionnaire revised
version TFEQ-R21 _TFEQ-R18-V2.
83
4.4.4 Most commonly measured appetitive traits by age group
Google Scholar screening of the 17 most robust questionnaires showed a range of citations
from zero for the ‘Hunger Sensitivity Scale’ (HSS) to 3020 publications for the TFEQ. The
three most highly cited questionnaires were the TFEQ (3020 citations), the DEBQ (1700
citations) and the CEBQ (460 citations). All of the traits measured fall within one of three
theory based categories: ‘restraint’, ‘emotional’ and ‘food and eating/externality’.
The TFEQ has been used in adolescents as well as adults (Gallant et al., 2010); and revised
versions of the TFEQ, the TFEQ-R18 and the TFEQ-R21 or TFEQ-R18-V2 have also been used,
although they are not fully robust. The DEBQ has also been used in children and
adolescents, either reported by the young person themselves using the DEBQ-C, or using a
parent-report version, the DEBQ-P (Braet & van Strein, 1997; van Strein & Oosterveld,
2008); although neither of these obtained a 4-point score for robustness within the
previous section. The CEBQ has only been used in children, the infant version, the BEBQ,
used in babies (Llewellyn, van Jaarsveld, et al., 2011), and in 13 year old Malay adolescents
as a self-report version (Loh et al., 2013), though again the psychometric properties of
these versions have not been fully tested. The sub-scales of these three measures and their
use in adult and child eating behaviour research is shown below in Table 4.4.
84
Table 4.4 The three most commonly used psychometric measures of appetite and the
traits they measure by age group: The TFEQ (3020 citations), the DEBQ (1700 citations)
and the CEBQ (460 citations) a
Theory-based Children and Psychometric

Traits Adults
categories adolescents questionnaires
Cognitive
TFEQ + DEBQ,
Restraint restraint or
DEBQ-C, DEBQ-P
Restraint
Emotional Disinhibition TFEQ
DEBQ, DEBQ-C,
Emotional
DEBQ-P, TFEQ-
eating
R18, TFEQ-R21
Emotional over
CEBQ, CEBQ self-
and under
report
eating
Food and DEBQ, DEBQ-C,

External eating
eating/Externality DEBQ-P
Food CEBQ, BEBQ,

responsiveness CEBQ self-report
Hunger TFEQ
Satiety CEBQ, BEBQ,

responsiveness CEBQ self-report
Enjoyment of CEBQ, BEBQ,

food CEBQ self-report
Slowness in CEBQ, BEBQ,

eating CEBQ self-report
CEBQ, CEBQ self-

Food fussiness
reportb
CEBQ, CEBQ self-

Desire to drink
report
a
Based on Google Scholar citations.
BEBQ: Baby Eating Behaviour Questionnaire; CEBQ: Child Eating Behaviour Questionnaire; CEBQ
self-report: Self-report version of the Child Eating Behaviour Questionnaire; DEBQ: Dutch Eating
Behaviour Questionnaire; TFEQ: Three factor Eating Questionnaire; TFEQ-R18: revised version of the
TFEQ, TFEQ-R21: revised version of the TFEQ.
b
‘Food fussiness’ is split into two factors ‘food fussiness-1’ and ‘food fussiness-2’
4.5 Discussion
This systematic review identified 38 existing psychometric questionnaire measures of

appetite. Of these, 14 were measures developed for use in children (including four in
adolescents and one in infants), and 24 in adults. Nine different countries were
85
represented by the questionnaires and associated research. Several different response

options were found in the 38 appetite measures within this study. Likert-like response
formats from “never”, to “always”, or “strongly disagree”, to “strongly agree”, were the
most common. Seventeen questionnaires were found to have a high level of robustness,
shown by a maximum 4-point score based on one point given each for: good internal
reliability, test-retest reliability, convergent validity and discriminant validity (Arnow et al.,
1995; Burgess et al., 2014; Cappelleri, Bushmakin, Gerber, Leidy, Sexton, Karlsson, et al.,
2009; Cepeda-Benito, Gleaves, Williams, et al., 2000; Geliebter & Aversa, 2003; Hawks et
al., 2004; Pliner & Hobden, 1992; Stunkard & Messick, 1985; Tanofsky-Kraff et al., 2008,
2007; Tylka & Kroon Van Diest, 2013; Tylka, 2006; van Strein et al., 1986; Walker et al.,
2015; Wardle, Guthrie, et al., 2001).
Twenty-one measures did not receive the full score for psychometric strength. Some of
these measures, which were deemed non-robust according to the point system
implemented, have not been widely used, such as the ‘Over-eating Tension Scales’ (OTS),
which was developed in 2000 to report tension and motivation-specific tension surrounding
eating and was only found to be cited on four occasions (n=4 citations) (Popkess-Vawter,
Gerkovich, & Wendel, 2000). In a few cases, low citation counts may be due to the
measures being developed recently, and further validation of the scales may still be under
way. This could apply for the ‘Italian Food Neophobia Scale’ for children (ICFNS) a self-
report adaptation of the ‘Food Neophobia Scale’ for children and adolescents (Laureati et
al., 2015) or the ‘Palatable Eating Motives Scale’ for kids (K-PEMS) that attempts to identify
individual motives for eating tasty foods in adolescents (Boggiano, Wenger, Mrug, Burgess,
& Morgan, 2015) to cite only a few examples.
Only 11 measures were examined for convergent or discriminant validity using other
measures of appetite, and the majority of comparisons (n=7) were in relation to the TFEQ
(Arnow et al., 1995; Cepeda-Benito, Gleaves, Williams, et al., 2000; Dalton et al., 2014;
Hawks et al., 2004; Lohse et al., 2007; Walker et al., 2015; Westenhoefer, 1991). Two
robust measures used experimental validation of their questionnaires (Carnell & Wardle,
2007; Pliner & Hobden, 1992; Wardle, Guthrie, et al., 2001), the ‘Food Neophobia Scale
(FNS) and the CEBQ.
As expected, the most commonly used measures of appetite were the TFEQ, DEBQ and the
CEBQ. The traits captured by these measures were ‘restraint’, ‘disinhibition’ and ‘hunger’
measured using the TFEQ in adults and adolescents ages 12 to 17 years old (Gallant et al.,
86
2010; Stunkard & Messick, 1985); ‘external eating’, ‘emotional eating’ and ‘restraint’
measured using the DEBQ in adults (van Strein et al., 1986), the DEBQ-P as a parent report
of these measures in nine to 12 year old children (Braet & van Strein, 1997), and as self-
report in seven to 12 year old children using the DEBQ-C (van Strein & Oosterveld, 2008);
‘emotional eating’ has also been measured in adults using the adapted versions of the
TFEQ, the TFEQ-R18 and the TFEQ-R21 or TFEQ-R18-V2 (Cappelleri, Bushmakin, Gerber,
Leidy, Sexton, Lowe, et al., 2009; Karlsson et al., 2000); and ‘food responsiveness’,
‘enjoyment of food’, ‘emotional over-eating’, ‘desire to drink’, ‘satiety responsiveness’,
‘emotional under-eating’, ‘food fussiness’ and ‘slowness in eating’ in three to 12-year-old
children using the CEBQ, and ‘food responsiveness’, ‘enjoyment of food’, ‘satiety
responsiveness’, and ‘slowness in eating’ using the BEBQ in infants aged four to 20 months.
The same traits measured in the CEBQ have also been adapted for use in 13 year old Malay
adolescents (Loh et al., 2013). However, the confirmatory factor analysis revealed a
different factor structure to the CEBQ (9 sub-scales vs. 8 sub-scales), separating the ‘food
fussiness’ sub-scale into two measures: ‘food fussiness-1’ and ‘food fussiness-2’, and adding
sugar-sweetened beverages (SSBs) to the ‘desire to drink’ items to express what types of
drinks the items referred to when assessing liquid consumption. This self-report version of
the CEBQ has not been validated in adult samples.
Overall, the traits captured by these three most commonly used measures, relate
predominantly to three different aspects of eating; ‘emotional’, ‘restraint’ and ‘food and
eating/externality’. These aspects of appetite, derive from three of the main theories of
obesity which have been posited to date: (1) the “Psychosomatic” theory, which proposes
that dysphoric mood is part of the aetiology of obesity (Kaplan & Kaplan, 1957); (2) the
“Restraint” theory, which posits that pathological aspects ‘external’ and ‘emotional eating’
are consequences of dieting (Herman, Polivy, Pliner, Threlkeld, & Munic, 1978; Herman &
Polivy, 1975; Polivy & Herman, 1976b); and (3) the “Externality” theory which suggest that
individuals over-eat based on external and lack of internal satiety cues (Schachter & Gross,
1968; Schachter, 1968). These three aspects of appetite are captured not only by those
measures which were identified as being the most common, but in general appeared to
inform all 17 of the robust measures of appetitive traits.
4.5.1 Measures that emcompass ‘emotional’ aspects of appetite
‘Emotional eating’ has been the specific focus of individual questionnaires (Arnow et al.,
1995; Geliebter & Aversa, 2003; Tanofsky-Kraff et al., 2007), as well as being measured by
87
sub-scales within broader questionnaires (Burgess et al., 2014; Cepeda-Benito, Gleaves,

Williams, et al., 2000; Nijs et al., 2007; Stunkard & Messick, 1985; van Strein et al., 1986;
Wardle, Guthrie, et al., 2001). These measures have been developed for both children
(Tanofsky-Kraff et al., 2007; Wardle, Guthrie, et al., 2001) and adults (Arnow et al., 1995;
Burgess et al., 2014; Cepeda-Benito et al., 2000; Geliebter & Aversa, 2003; Nijs et al., 2007;
Stunkard & Messick, 1985; van Strein et al., 1986). Overall, the ‘Emotional Eating Scale’
(EES), and the ‘emotional eating’ (EE) sub-scale of the DEBQ-EE, together with the
‘emotional eating’ sub-scales of the revised measures of the TFEQ, the TFEQ-R18, and the
TFEQ-R21 or TFEQ-R18-V2 (which did not fall within the most robust measures in this
systematic review), measure ‘emotional eating’ in adults across a wide range of emotions.
These measures have been mainly used to study participants with disordered eating or
binge-related obesity spectrums, and differences in associations between ‘emotional
eating’ and BMI obtained using different TFEQ versions, suggest that multiple factors
related to dieting issues and not only emotions are still at play (Cappelleri, Bushmakin,
Gerber, Leidy, Sexton, Lowe, et al., 2009; Karlsson, Persson, Sjöström, & Sullivan, 2000).
Thus, the interaction between these different measures of ‘emotional eating’ still require
further studies, particularly in different populations.
In children, positive associations between BMI and ‘emotional eating’ have not always been
found using measures such as the ‘Emotional Eating Scale’ adapted for use in children and
adolescents (EES-C), the DEBQ-C, the DEBQ-P, and the CEBQ (Baños et al., 2011; Braet &
van Strein, 1997; Croker et al., 2011; Santos et al., 2011; Tanofsky-Kraff et al., 2007; van
Strein & Oosterveld, 2008; Wardle, 1987a). These conflicting results have led to the
suggestion that adults have a greater capacity than children to discriminate between their
emotions (Braet & van Strein, 1997). Overall, there is very little research on whether these
constructs interact with each other, to measure similar aspects of ‘emotional eating’. For
example, significant positive associations were found between the EES ‘Anger/Frustration
and Depression’ sub-scales and the ‘disinhibition’ sub-scale of the TFEQ (Arnow et al.,
1995). Given the majority of these measure were developed in the light of the
“Psychosomatic” theory (Kaplan & Kaplan, 1957), the issue is still argued as to whether the
‘emotional eating’ is brought on by dieting or disordered eating, and whether it is a cause
or consequence of excess weight. Interestingly, the CEBQ is the only questionnaire to
measure ‘emotional eating’ in the light of research surrounding variation in the eating
styles hypothesised to predispose one to weight gain and obesity (Wardle, Guthrie, et al.,
88
2001), rather than in the context of disordered eating. The CEBQ also measures ‘emotional
under-eating’, a sub-scale that is currently not measured in adults.
4.5.2 Measures that emcompass ‘restraint’ aspects of appetite
A total of three questionnaires which measure ‘restraint’ were found to be robust: the
TFEQ, the DEBQ and the ‘Flexible Control’ and ‘Rigid Control’ sub-scales of the TFEQ
‘restraint’ scale (Stunkard & Messick, 1985; van Strein et al., 1986; Westenhoefer, 1991).
All measures of ‘restraint’ originate from the ‘Restraint Scale’ (RS). The RS was initially
developed as a self-report measure of chronic dieting for the purpose of weight control
(Herman & Mack, 1975; Herman & Polivy, 1975), and therefore did not meet the inclusion
criteria to be included in this study. The RS is not considered a valid measure of ‘restrained
eating’, as it also contains items that measure dieting strategies, weight fluctuation and
‘disinhibition’ (Cappelleri, Bushmakin, Gerber, Leidy, Sexton, Lowe, et al., 2009; Herman &
Polivy, 1975; Stunkard & Messick, 1985). The measurements of aspects of ‘disinhibition’
and ‘restraint’ are confounded with each other in the RS (Johnson et al., 2012; Meule,
Papies, & Kübler, 2012). The TFEQ and the DEBQ were developed to try and eliminate this
confounding from the RS (Stunkard & Messick, 1985; van Strein et al., 1986). The TFEQ
‘cognitive restraint’ and the ‘Flexible Control’ and ‘Rigid Control’ dimensions of the
‘restraint’ assess the relationships between ‘restraint’, ‘disinhibition’ and disordered eating
(Stunkard & Messick, 1985; Westenhoefer et al., 1999; Westenhoefer, 1991). The DEBQ, on
the other hand, measures only ‘dietary restraint’ (van Strein et al., 1986). Other scales that
were not found to be sufficiently robust in this review have also been used to measure
‘restraint’ (Cappelleri, Bushmakin, Gerber, Leidy, Sexton, Lowe, et al., 2009; Jiang et al.,
2014; Karlsson et al., 2000).
A number of conflicting results surround the measurement of ‘restraint’ and its relationship
with BMI (Anglé et al., 2009; de Lauzon-Guillain et al., 2006; Williamson et al., 1995)
(Chapter 2, Section 2.5.3). However, it is clear from citations of the TFEQ (3020 citations)
and the DEBQ (1700 citations), that ‘restraint’ has received great attention and research
into ‘restraint’ covers all age ranges from childhood and adolescence to adulthood.
‘Restraint’ has also been studied longitudinally in adults and adolescents (Drapeau et al.,
2003; Johnson & Wardle, 2005; Svensson et al., 2014). Comparisons between different
measures of ‘restraint’ (such as differences between the RS, the TFEQ, the DEBQ and the
Flexible and Rigid control dimensions of ‘restraint’) have also been carried out (Laessle,
Tuschl, Kotthaus, & Pirke, 1989; Williamson et al., 2007) which suggests that further
89
development of ‘restraint’ scales are unwarranted. ‘Restraint’ is not captured by the CEBQ
because the CEBQ was designed to capture those aspects of eating thought to have a
biological basis, whereas ‘restraint’ is thought to be psychologically driven (Wardle,
Guthrie, et al., 2001).
4.5.3 Measures that emcompass ‘food and eating/externality’ aspects of

appetite
The majority of robust questionnaires (n=14), fall into those which measure aspects of
‘food and eating/externality’ (Burgess et al., 2014; Cappelleri, Bushmakin, Gerber, Leidy,
Sexton, Karlsson, et al., 2009; Cepeda-Benito et al., 2000; Geliebter & Aversa, 2003; Hawks,
Merrill, & Madanat, 2004; Nijs et al., 2007; Pliner & Hobden, 1992; Stunkard & Messick,
1985; Tanofsky-Kraff et al., 2008; Tylka & Kroon Van Diest, 2013; Tylka, 2006; van Strein et
al., 1986; Walker et al., 2015). Only two of these measures were specifically developed for
children, the ‘Eating in the Absence of Hunger’ questionnaire (EAH-C) and the CEBQ
(Tanofsky-Kraff et al., 2008; Wardle, Guthrie, et al., 2001). All robust questionnaires except
the CEBQ measure only one or two specific aspect of ‘food and eating/externality’. For
example, the ‘Palatable Eating Motives scale’ (PEMS), assesses motivations for eating tasty
foods through the ‘conformity’ and ‘social’ motives scales (Burgess et al., 2014). Neither of
these sub-scales have been found to be associated with BMI in 169 college students
(Boggiano et al., 2015). The newly developed ‘Hunger Sensitivity Scale’ (HSS) assesses
emotional aspects of hunger, but it measures behavioural changes such as snacking and
eating around others who are eating, as external triggers to internal satiety sensitivity
(Walker et al., 2015). The HSS has been validated against the ‘hunger’ sub-scale of the
TFEQ (Stunkard & Messick, 1985; Walker et al., 2015), and it shows promise as a new
measure assessing ‘hunger sensitivity’; but BMI was not associated with HSS scores in a
sample of 556 university students (Walker et al., 2015). The ‘hunger’ sub-scale of the TFEQ
on the other hand, has shown inconsistent relationship with BMI (Lindroos et al., 1997;
Stunkard & Messick, 1985). Other multi-faceted measures of appetite, such as the DEBQ
and the CEBQ, do not contain measures of ‘hunger’; consequently, psychometric measures
of ‘hunger’ still rely on the TFEQ (Stunkard & Messick, 1985). Previous research suggests
that those who struggle with hunger could also experience cravings and disordered eating
(Elfhag & Rössner, 2005; Finlayson et al., 2007).
A further measure which was found to be robust was the ‘Power of Food Scale’ (PFS)
(Cappelleri, Bushmakin, Gerber, Leidy, Sexton, Karlsson, et al., 2009; Lowe et al., 2009). The
90
PFS is one of the most recently developed and widely used instruments (74 citations) that
measures appetite related to living in a food-abundant environment and it has been
validated against the TFEQ-R21 or TFEQ-R18-V2 (Cappelleri, Bushmakin, Gerber, Leidy,
Sexton, Lowe, et al., 2009). The PFS has shown no significant associations between any of
its three sub-scales and BMI (Cappelleri, Bushmakin, Gerber, Leidy, Sexton, Karlsson, et al.,
2009). Also not associated to BMI, the ‘Food Neophobia Scale’ (FNS) in adults (Pliner &
Hobden, 1992) also obtained a 4-point score for robustness. These results differ from a
similar concept related to neophobia, such as ‘food fussiness’ measured using the CEBQ in
children which has been negatively associated with weight in a few studies (Hill et al., 2009;
Loh et al., 2013; Mallan et al., 2013; Rodenburg et al., 2012; Spence et al., 2011; Viana et
al., 2008) (Chapter 2, Section 2.5.3). No other measure of ‘food fussiness’ was seen in these
robust questionnaires in adults.
‘External eating’ was first measured using the DEBQ (10 items), and its associations with
weight have been inconsistent (Schachter, 1968; van Strein, 1986; Wardle, 1987; Wardle et
al., 1992) (Chapter 2, Section 2.5.3). ‘External eating’ measured using the DEBQ-P (the
parent-report version of the DEBQ) was significantly associated with BMI in one study
(Braet & van Strein, 1997), although it did not significantly differ between obese,
overweight and normal weight groups of children in another study (Caccialanza et al.,
2004). The DEBQ has also been developed as a self-report measure, modified to be
answered by nine to 12 year olds in a version known as the DEBQ-C (Baños et al., 2011; van
Strein & Oosterveld, 2008). The CEBQ construct ‘food responsiveness’, which contains five
items related to response to external food cues, has been consistently positively associated
with BMI-SDS scores in children (Croker et al., 2011; Sleddens, Kremers, & Thijs, 2008;
Viana et al., 2008). Some of the inconsistencies in the associations observed between
‘external eating’ and weight may be driven by ‘emotional eating’ (as a predictor of over-
eating), rather than by eating in response to food cues (Koenders & Van Strein, 2011).
Recently there has been interest in the measurement of ‘intuitive eating’; four of the 38
scales reviewed measured this aspect of eating (Hawks, Merrill, & Madanat, 2004; Hulbert-
Williams, Nicholls, Joy, & Hulbert-Williams, 2013; Tylka & Kroon Van Diest, 2013; Tylka,
2006). ‘Intuitive eating’ has been associated to a tendency to eat following physical hunger
and internal satiety cues to help determine what and how much you eat and is said to be an
aspect of eating which relies on internal sensations. Three measures of ‘intuitive eating’
were found to be robust, the IES and IES-2 and the IES-H scales (Hawks, Merrill, & Madanat,
91
2004; Tylka & Kroon Van Diest, 2013; Tylka, 2006). The IES has been cited in 161
publications to date since it was developed in 2006 (Tylka & Kroon Van Diest, 2013; Tylka,
2006), making it one of the preferred measures in the present literature, although it has yet
to reach the number of citations achieved by older measures. All three ‘intuitive eating’
scales have been found to be negatively associated to BMI in predominantly white college
students (Hawks, Merrill, & Madanat, 2004; Tylka & Kroon Van Diest, 2013; Tylka, 2006).
Intuitive eating shows some resemblance to the ‘satiety responsiveness’ scale in the CEBQ,
which measures eating in response to internal satiety cues. However, a closer look at the
items reveals the CEBQ measures a sensation of fullness (e.g. “I often leave food on my
plate at the end of a meal”, “I often get full before my meal is finished”, “I get full up
easily”), compared to a reliance on intuitive measures of satiety in the IES (“I trust my body
to tell me when….”, “I trust my body to tell me what…”, “I trust my body to tell me how
much”), in children vs adults. Given the possible similarities between ‘intuitive eating’ and
constructs such as ‘satiety responsiveness’ measured by the CEBQ, differences between
these measurements should be identified through convergent/discriminant validity studies
and in similar age ranges. This would help determine if future scale development would
benefit from inclusion of measures of both intuitive internal satiety cues and
responsiveness to satiety.
4.5.4 Limitations
Several limitations of the present study should be acknowledged. The robustness indicator
of psychometric measures was based on both reliability and validity studies. In the last few
decades validity studies have changed from a focus on whether the test measures what it is
intended to measure to the study of participants’ characteristics and what scores they
achieved (Cronbach, 1951; Streiner & Norman, 2015). Thus, although authors might
suggest that the measure which they obtained is valid, this might only refer to the use it
serves in a particular group of people and the context in which it was tested (Streiner &
Norman, 2015). Consequently, older measures, which have been tested in numerous
settings and under different conditions (such as the TFEQ, the DEBQ, or the EES) are
considered to be psychometrically sound measures and are used in different studies to
assess convergent and discriminant validity.
Validation of measures against other similar instruments, although the most common
method of convergent validity, is sometimes difficult to justify if a particular set of
measures already exists (Streiner & Norman, 2015). Other forms of psychometric
92
validation could be carried out through behavioural/laboratory validations, providing

objective measures of specific traits that serve to validate psychometric measures (Carnell
& Wardle, 2007). Five measures have been tested using these types of studies, the ‘Food
Situations Questionnaire’ (FSQ) (Loewen & Pliner, 2000), the Food Neophobia Scale in
children (FNS-C), the Italian version of the FNS, a self-report measure in children (ICFNS),
and the CEBQ (Carnell & Wardle, 2007; Laureati et al., 2015; Loewen & Pliner, 2000; Pliner
& Hobden, 1992; Pliner, 1994). Of these, the CEBQ and the FNS were two measures which
were found to score 4-points for robustness (Carnell & Wardle, 2007). However, when
behavioural measures are correlated against psychometric measures for validation
purposes, it is unclear what size of correlation should be obtained to support the scale
being a valid measure of a trait (Carnell & Wardle, 2007). Contemporary questionnaires
tend to include reliability and validity measures in their publications, obtained from
exploratory and confirmatory factor analysis on different samples.
Given the results from this search are based on published studies, the review is prone to
publication bias. It is unknown how many studies reporting associations between
questionnaire measures of appetite and other factors such as BMI with non-significant
results have remained unpublished. Sampling bias from convenience sampling used to
obtain participants for questionnaire validation is also an issue. Studies using psychometric
questionnaires require the use of validation studies in different samples to assess their
generalisability (Allison & Baskin, 2009; Streiner & Norman, 2015). Where studies used
clinical samples, individuals in obesity clinics and treatment centres for disordered eating
are those most likely to have participated, leaving out participants with these conditions
but no access to treatment. In studies with university or college students, those who were
included generally gained credits for participating in the study, possibly excluding students
who were not interested in the course. All of these scenarios could result in the inclusion of
poorly designed studies that are unreliable due to samples that do not represent the
populations they are supposed to. Further studies using the measures in different samples,
would provide further reliability and validity results that could potentially eliminate this
problem, however this is costly and time consuming (Streiner & Norman, 2015).
Given the wide definition of appetite (Section 2.2), it is impossible to include all of its
broader elements. Only ‘trait’ aspects of appetite were included in this review, which were
related to weight and thought to be present across the whole weight spectrum. For
example, a newly developed ‘Culturally-based Communication about Health, Eating, and
93
Food’ (CHEF) scale, was removed from the search on the basis that these constructs cannot
be viewed as ‘traits’ (Hubbard et al., 2015). Measures of parental feeding practices such as
the ‘Child Feeding Questionnaire’ (CFQ) were also excluded (Birch et al., 2001).
Questionnaires that specifically measured dimensions of eating disorders, were also not
included in the study. Samples of participants with eating disorders were included only if
the scale had subsequently been used in non-pathological conditions. It is therefore
possible that appetite measures developed in clinical samples, but which might be
beneficial for use in non-pathological populations were omitted.
4.5.5 Conclusions
From this review, 38 existing psychometric questionnaires used to measure different

aspects of appetite related to weight were identified. Of these, a total of 17 had high
robustness scores, assessed using a 4-point scoring system. These 17 questionnaires
measure different traits which broadly describe ‘emotional’, ‘restraint’ and ‘food and
eating/externality’ aspects of appetite. Of the most robust measures, the CEBQ used
laboratory-based measures to demonstrate the validity of different aspects of food and
eating of the questionnaire. The three most frequently cited questionnaires were the TFEQ
and the DEBQ, which are used mainly in adults, and the CEBQ, which is used in children.
These three questionnaires have been used extensively in many countries and have allowed
an improved understanding of different appetitive traits.
This review identified several traits measured in children that have no parallel psychometric
measure for adults. There is currently no psychometric measure of ‘satiety
responsiveness’, assessing responsiveness to fullness sensations unrelated to intuition for
adults. ‘Emotional under-eating’ captured by the CEBQ has also not been measured in
adults. ‘External eating’ has been measured in adults using the DEBQ, but this measure has
not been consistently associated with BMI, in contrast to the similar construct ‘food
responsiveness’ from the CEBQ, which has been consistently associated with a degree of
overweight in children. Therefore, adult measures of ‘satiety responsiveness’, ‘emotional
under-eating’ and ‘food responsiveness’ would be useful to allow for exploration of the
impact of these traits on weight into adulthood. Presently, these traits as well as others
such as ‘enjoyment of food’, ‘slowness in eating’, ‘food fussiness’, and ‘desire to drink’ may
be validly measured in children using the CEBQ. An adult version of the CEBQ would extend
the applicability of this measure to another life-stage, allowing longitudinal analysis in
future with older age groups. This review supports the need to develop a measure of
94
appetitive traits in adults that encompasses broader measures of appetitive traits not
related to ‘restraint’ and ‘disinhibition’, including measurements of sensitivity to internal
and external food cues.
95
Chapter 5. Development of the ‘Adult Eating Behaviour Questionnaire’
Chapter 5. Study 2: Development of the ‘Adult Eating

Behaviour Questionnaire’20
5.1 Background Appendix 5.1, Appendix 5.2
The results from the systematic review in Study 1, Chapter 4, highlighted a number of
existing valid and reliable questionnaires that measure appetite. The most commonly used
measures of appetite in adults are the TFEQ (Stunkard & Messick, 1985), the DEBQ (van
Strein et al., 1986). In children, the CEBQ is the most comprehensive measure and was
originally developed for use in the UK. The CEBQ encompasses aspects of food and eating,
as well as emotional aspects of appetite. It also includes other constructs not addressed in
the adult literature, such as ‘satiety responsiveness’ and ‘enjoyment of food’, ‘emotional
under-eating’, ‘slowness in eating’, and ‘food responsiveness’ which is unrelated to
‘restraint’ or ‘disinhibition’.
The CEBQ has been validated for use in children from different ethnic and cultural
backgrounds, including Australian children aged one to five-year-old (Mallan et al., 2013),
and low-income Hispanic and African American children aged two to five years in the USA
(Sparks & Radnitz, 2012). The CEBQ has been used extensively to assess the relationship
between appetitive traits and weight at different ages (Ashcroft, Semmler, Carnell, van
Jaarsveld, & Wardle, 2008; Soussignan, Schaal, Boulanger, Gaillet, & Jiang, 2012; Webber,
Hill, Saxton, Van Jaarsveld, & Wardle, 2009), in different populations, and in different
languages (Santos et al., 2011; Sleddens et al., 2008; Soussignan et al., 2012; Viana et al.,
2008). It has also been used to assess differences in appetitive traits in obese populations
and in clinical settings (Croker et al., 2011).
20
A version of this chapter has been accepted for publication: Hunot, C., Fildes, A., Croker, H.,
Llewellyn, C. H., Wardle, J., & Beeken, R. J. (2016). Appetitive traits and relationships with BMI in
adults: Development of the Adult Eating Behaviour Questionnaire. Appetite.
http://dx.doi.org/10.1016/j.appet.2016.05.024. A copy of this paper is presented in Appendix 5.1.
Versions of this chapter were also presented at conferences (Appendix 5.2).
96
Some of the most important work carried out using the CEBQ, has focused on the
relationship between ‘food responsiveness’ and BMI-SDS, finding positive associations
between this trait and weight in multiple samples (Carnell & Wardle, 2008a; Croker et al.,
2011; Santos et al., 2011; Viana et al., 2008). ‘External eating’ (a similar construct to ‘food
responsiveness’) has also been measured using the DEBQ (van Strein et al., 1986) and its
child report (van Strein & Oosterveld, 2008) and parent report (Braet & van Strein, 1997)
versions. However, reported associations between this trait and BMI have been mixed,
including positive associations (van Strein et al., 1986), negative associations (Baños et al.,
2011), or no associations at all (Caccialanza et al., 2004). These inconsistent findings might
reflect the fact that the DEBQ was designed to assess clinically disordered eating behaviour
and may therefore be less applicable to non-clinical samples, in contrast to the CEBQ, which
was designed to capture a normal range of eating styles.
There is some evidence from studies using the CEBQ that appetitive traits vary with age
(Ashcroft et al., 2008). However, studies exploring changes in appetitive traits across the
life course have been limited by the lack of a comparable self-report measure of appetitive
traits for adults. There has been increased interest from clinicians and researchers who
would like to use the CEBQ in adult populations as weight gain is more common at older
ages and appetitive traits may influence this.
The systematic review of the existing psychometric measures of appetite and appetitive
traits in the previous chapter, shows that there is no measure in adults that encompasses
the aspects of appetite captured by the CEBQ (Wardle, Guthrie, et al., 2001). In particular,
there is no comparable measure of ‘food responsiveness’ and ‘satiety responsiveness’,
neither of which have been adequately captured by existing measures of appetite in
adulthood. Measurement of these traits in adults would contribute to our understanding
of how these specific traits influence weight gain at older ages (French et al., 2012).
Together with the BEBQ (Llewellyn, van Jaarsveld, Johnson, Carnell, & Wardle, 2011), an
infant version of the CEBQ, the addition of the ‘Adult Eating Behaviour Questionnaire’
(AEBQ) would enable these eating traits to be measured across the life course using three
life-stage appropriate instruments. This would make it possible to longitudinally track
appetitive traits from infancy (BEBQ) and childhood (CEBQ) into adulthood (AEBQ), to give a
better picture of the association between appetitive traits and weight across the life-
course. As mentioned previously in Chapter 1, Section 1.4, appetitive trait scores (using
97
the AEBQ) could also serve to inform interventions, tailoring weight loss and weight
management advice to an individual’s appetitive trait profile.
5.2 Aims
The aim of this study was to develop an internally reliable self-report ‘Adult Eating
Behaviour Questionnaire’. The specific objectives were to: (1) adapt a prototype self-report
AEBQ from the parent-report ‘Child Eating Behaviour Questionnaire’ (CEBQ); (2) pilot the
AEBQ in samples of adults; and (3) assess the factor structure of the AEBQ to ascertain the
appetitive traits measured by the questionnaire.
5.3 Methods
Preliminary work was carried out to adapt the CEBQ into a self-report measure for adults.
The decision over which items were selected to be adapted and included in the AEBQ was
based on: (1) The translation of parent-report items into self-report items; (2) findings from
the systematic review (Chapter 4), and input from experts on eating behaviour to develop
new items to measure ‘hunger’ in a self-report format, as ‘hunger’ was not assessed in the
CEBQ; and (3) piloting in a sample of adults.
5.3.1 Translation of the CEBQ into the AEBQ
Initially, the wording of all 35 CEBQ items was changed from the parent-report “My child....
” format to a self-report “I …..” format (e.g. “My child loves food” was changed into “I love
food”). The original response format of the CEBQ (‘never’, ‘rarely’, ‘sometimes’, ‘often’ and
‘always’) was kept (see Appendix 5.3). The CEBQ item, “My child eats more when s/he is
happy” loads onto the ‘emotional under-eating’ scale (Wardle, Guthrie, et al., 2001). It was
of interest that a CEBQ item describing eating more in response to a positive emotion
loaded onto a construct or scale for ‘emotional under-eating’, so questions were added
denoting both directions of emotional responses for items on the ‘emotional under-eating’
and ‘emotional over-eating’ scales, in order to confirm which AEBQ constructs they would
load onto. This meant an additional four ‘emotional over-eating’ items (“I eat more when I
am angry”; “I eat more when I am upset”; “I eat more when I am tired” and “I eat more
when I am bored”) and an additional six ‘emotional under-eating’ items (“I eat less when I
am happy”; “I eat less when I am annoyed”; “I eat less when I am anxious”; “I eat less when
98
I am worried”; “I eat less when I am bored”; and “I eat less when I have nothing else to do”)
were added, in addition to those already contained in the CEBQ. This increased the original
questionnaire from 35 to 45 items.
This early version of the AEBQ with the literal translation of each CEBQ item into a self-
report format was given to 10 adults (eight females and two males; mean age 31.9 ± 7.8) to
complete and provide initial feedback. Further refining of the questionnaire took place in
group discussions with psychologists, dieticians, and experts in eating behaviour (n=4).
Based on these discussions and feedback, the three items from the ‘desire to drink’ scale of
the CEBQ were eliminated because this scale was deemed difficult to complete. Questions
such as “My child is always asking for a drink”, which in the AEBQ became “I am always
asking for a drink”, were also considered difficult for adults to answer as became unclear
what type of drink was being referred to (e.g. alcoholic or non-alcoholic). Furthermore, the
item “my child is always asking for food” from the ‘food responsiveness’ scale in the CEBQ,
which became “I am always asking for food” in the AEBQ, was also perceived to be
inappropriate for an adult to respond to. Therefore, the three items from the ‘desire to
drink’ scale, and the “I am always asking for food” item from the ‘food responsiveness’
scale were eliminated, leaving 41 remaining items.
5.3.2 Review of items from other questionnaires on appetite from existing

literature
Following examination of the main appetite dimensions measured in the systematic review
in Study 1, Chapter 4, Section 4.4.4, it became apparent that the self-report format of the
CEBQ did not contain a measure of ‘hunger’ experience. ‘Hunger’ is an important aspect of
appetite that could not be measured in the CEBQ, as parents are unable to determine their
child’s experienced level of physical hunger and would only be able to report on their
behaviours in relation to food (Wardle et al., 2013).
‘Hunger’ is measured by 14 items in the TFEQ-R18, a shortened version of the TFEQ (de
Lauzon et al., 2004). However, these items fall within the ‘uncontrolled eating’ construct of
the TFEQ-R18, with items such as, “I am usually so hungry that I eat more than three times
a day” and “Dieting is so hard for me because I just get too hungry”. Other items such as, “I
often get so hungry that my stomach feels like it will never be full up” and “I am always
hungry enough to eat at any time” are items that relate to ‘restraint’ and ‘disinhibition’
which, as discussed in the review (Chapter 4, Section 4.5.2), are not the purpose of the
99
AEBQ. Therefore, none of the TFEQ-R18 ‘hunger’ items were deemed appropriate for use
in the AEBQ.
‘Hunger’ measurements were also examined in the ‘ecSatter Inventory’ (ecSI) (Lohse et al.,
2007; Stotts & Lohse, 2007), including the ‘experiential process of hunger’ sub-scale with
one item, “I eat as much as I am hungry for”, which seems inadequate as a measure of
‘hunger’, as it does not attempt to quantify the level or frequency of physical ‘hunger’
(Wardle, 1987b). A further measure examined was the ‘Intuitive Eating Scale’, which
measures ‘reliance on internal hunger and satiety cues’ to determine when and how much
to eat (Avalos & Tylka, 2006; Tylka & Kroon Van Diest, 2013; Tylka, 2006). Here, the ability
to interpret internal signals of satiety are measured with items such as “I can tell when I am
slightly full/slightly hungry”, “I trust my body to tell me when…(to eat)”. Again, this scale
does not attempt to measure the level or frequency with which physical ‘hunger’ is
experienced.
As none of these questionnaires capture differences in experienced levels of physical

‘hunger’ unrelated to emotional or restraining situations and after discussion with a panel
of clinical psychologists, behavioural scientists, dieticians and authors of the original CEBQ, I
felt it was important to find questions to reflect the physical experience of hunger that
could be incorporated into the AEBQ. A set of questions used in the Weight Concern
‘Shape-Up’ manual to help participants distinguish between ‘hunger’ or ‘craving’ appeared
to capture this physical ‘hunger’ (Wardle et al., 2013). Weight Concern is a registered
charity, set up in 1997 to tackle the rising problem of obesity in the UK (Weight Concern,
2016a). Part of the work it does is through ‘Shape-Up’, a lifestyle programme that helps
individuals to manage their weight and improve their health and quality of life. A clinical
psychologist with considerable experience of working with obese patients developed the
five items on ‘hunger’. These items were therefore added to the AEBQ to measure the level
of physical hunger that a person experiences: “I often notice my stomach rumbling”’; “I
often feel so hungry that I have to eat something right away”; “If I miss a meal I get
irritable”; “I am always hungry at certain times of the day”; and “If my meals are delayed I
get light-headed” (Appendix 5.4).
Measures related to desire to eat when in the presence of palatable food (which relate to
‘food responsiveness’) were also not included in the CEBQ, again because parents would be
unable to answer about their children’s eating. Items from other questionnaires were
therefore considered for inclusion. This included the ‘external eating’ construct of the
100
DEBQ (“If food smells and looks good, do you eat more than usual?”; “If you see or smell
something delicious, do you have a desire to eat it?”; “If you walk past the baker, do you
have the desire to buy something delicious?”) (van Strein et al., 1986), the ‘uncontrolled
eating’ scale from the TFEQ-R18 (“When I see or smell really delicious foods, I find it very
difficult to keep from eating - even if I´ve just finished a meal”) (de Lauzon et al., 2004), and
one item from the ‘Power of Food Scale’ (PFS) (“If I see or smell a food I like, I get a
powerful urge to have some”) (Lowe et al., 2009). Although these items were thought to
be clearly worded measures of a person’s interest in food and drive to eat, the experts on
eating behaviour recommended that specific items should be developed for the AEBQ that
captured ‘food responsiveness’ more specifically, as defined by Schachter (1968) (such as
wanting to eat in the presence of others eating, or wanting to eat when seeing or smelling
food). Thus, three ‘food responsiveness’ items were developed and added for piloting: “I
am always thinking about food”, “When I see or smell food that I like, it makes me want to
eat” and “I feel hungry when I am with someone who is eating”. Finally, the panel reviewed
all included and excluded items to ensure no further additions/removals were felt to be
required. The total number of items obtained from this process was 49.
5.3.3 Piloting in a sample of adults
The extended 49-item version of the AEBQ was loaded onto Survey Monkey for piloting
(Appendix 5.5). Survey Monkey is a web-based provider of survey solutions which enables
the researcher to obtain secure data from participants, who are given a direct link to the
previously up-loaded questionnaire (“Survey Monkey,” 2016). The aims of the pilot were to
test the understanding of the questionnaire and to establish if the items and response
options generated by the 49-item AEBQ made sense. The AEBQ was given to a sample of
30 adults, recruited opportunistically through personal contacts and a snowballing
technique was used to increase the response rate with the aim of obtaining a minimum of
40 responses. Colleagues at University College London were asked to circulate a link to the
questionnaire to their friends and family from a range of professional backgrounds.
Anyone aged 18 or older could answer the questionnaire. Participants were asked to
respond to each individual item and the questionnaire as a whole and give feedback on
their experience of completing the AEBQ in an open answer section at the end of the
questionnaire. Open-ended answers were obtained in an Excel spreadsheet from Survey
Monkey. Following completion of data collection, a scoring system was developed that
101
calculated the most frequent comments made and which items they mapped onto, to help
identify which items were problematic (Willig, 2008).
Piloting with 49 adults (21 to 73 years old), 36 women (79.6%) and 13 men (20.4%), led to
the deletion of two items “Given the choice, I would always have food in my mouth”
because several participants commented that it “sounded a bit odd” or was “over the top”;
and a second item (“I am interested in food”) was eliminated because participants reported
they found the meaning ambiguous. Also, the response options ‘never’, ‘rarely’,
‘sometimes’, ‘often’ and ‘always’ were described by participants as unclear and were felt to
not fit the questions adequately (Appendix 5.6). Response options were therefore changed
from ‘never’, ‘rarely’, ‘sometimes’, ‘often’ and ‘always’, to ‘strongly disagree’, ‘disagree’,
‘neither agree not disagree’, ‘agree’ and ‘strongly agree’. The new response options were
tested with a further small convenience sample (two females and three males, aged 31 ± 7
years). This answer format appeared to be more meaningful and better understood by this
sample. The remaining 47 item version of the AEBQ was used to assess its factor structure.
5.3.4 Assessing the factor structure of the AEBQ
5.3.4.1 Design and study population
Adults 18 years and over, were invited to complete a cross-sectional survey collected
between the months of August and September 2013, where the AEBQ was answered via an
on-line questionnaire. Participants were invited to take part by a provider of sampling and
data collection for survey research called Research Now, who hold a panel of over 200,000
UK residents that have consented to answer on-line questionnaires (“Research Now,”
2014). The aim was to recruit at least 500 adults (the minimum sample size for Principal
Component Analysis (PCA) is n=10 participants for each of 47 items – see Section 5.3.4.5
below), with quotas set for 100 participants in each of the following age strata: 18 to 19
years, 20 to 24 years, 25 to 29 years, 30 to 39 years, 40 to 49 years, 50 to 59 years, and 60
plus years.
5.3.4.2 Measures
5.3.4.2.1 Demographic
Participants provided demographic information including ethnicity (data was collapsed into
three categories: ‘White’ and ‘Non-white’ [‘Black’, ‘Asian’ or ‘Mixed’]) (Office for National
Statistics, 2012) and ‘Preferred not to answer’; education (data was collapsed into three
categories for analyses: ‘School’ [‘Primary school/Secondary school/O-level/GCSE’],
102
‘College’ [‘A levels/Technical or trade certificate/Diploma’], and ‘University’

[‘Undergraduate degree/Postgraduate degree’]); employment status (was grouped as;
‘employed’ [‘Employed full-time/ Employed part-time/Self-employed’], ‘not employed’
[‘Unemployed/Full-time homemaker/Unpaid voluntary work/Student’] and ‘disabled or
retired’ [‘Disabled or too ill to work/Retired’]); and, current living arrangement (was
categorized as: ‘Home owner’ [‘Own home outright/Own home with mortgage’], ‘renting’
[‘Rent from local authority/Housing association/Rent privately’] or ‘other’ [‘Living with
parents/Living in University/College residential accommodation’]) (Wardle, Robb, &
Johnson, 2002) (Appendix 5.7).
5.3.4.2.2 Anthropometric
Participants self-reported their weight and height (Appendix 5.7). BMI was calculated used
to categorise the sample into: Underweight (<18.5), normal weight (18.5 to 24.9),
overweight (25.0 to 29.9) and overweight (≥30).
5.3.4.2.3 Appetitive traits
Participants completed the 47 item AEBQ (Appendix 5.7).
5.3.4.3 Statistical analysis
PCA was carried out using SPSS version 22.0 (IBM, 2013b). Descriptive information was
based on frequency tables and cross-tabulation.
5.3.4.4 Principal component analysis
Two similar techniques are commonly used to explore the properties of newly created
scales: Factor Analysis and PCA (Field, 2013). PCA is considered the simplest theoretically
and the soundest mathematically to assess psychometric data (Stevens, 2009), as it
transforms the data set into linear components without estimating components from
unmeasured variables, and accounts for most of the variance of observed variables (Field,
2013).
In order to verify the structure of the AEBQ and to ascertain whether it was similar to the
original CEBQ (Wardle, Guthrie, Sanderson, & Rapoport, 2001), PCA was therefore chosen
to explore the factor structure of the AEBQ.
103
5.3.4.5 Sample size calculation
At least 10–15 participants per variable are commonly recommended to test the factor
structure of a questionnaire (Oppenheim, 2003). However, Tinsley (1987) recommends
having between five and 10 participants per variable up to a total of 300 (beyond which test
parameters tend to be stable regardless of the participant to variable ratio) (Tinsley &
Tinsley, 1987). Comrey and Lee (1992) class 300 as a good sample size for PCA, 100 as poor
and 1000 as excellent. I therefore aimed to collect a sample size of at least 500 participants
to test the 47 item questionnaire.
In order to evaluate the sampling adequacy of a data set, SPSS provides two outputs: The
Kaiser-Meyer-Olkin (KMO) statistic and Bartlett’s test of sphericity. The KMO statistic is
used as a measure of sampling adequacy and should be greater than 0.6. Kaiser (1958)
recommends accepting KMO values less than 0.5 as ‘barely acceptable’, values between 0.5
and 0.7 as ‘mediocre’, values between 0.7 and 0.8 as ‘good’, values between 0.8 and 0.9 as
‘great’ and values above 0.9 as ‘superb’. Values less than 0.5 should be removed (Kaiser,
1958). Bartlett’s test of sphericity, which examines whether the covariances are zero and
the variances are roughly equal, should be statistically significant (Field, 2013).
5.3.4.6 Data extraction
The main aim of PCA is to reduce a large number of correlated variables to a few common
components (factors) that explain the greatest proportion of variance in the data, while
losing the least amount of data. This is initially achieved by calculating the common
variance shared between the variables in a correlation matrix. PCA calculates all possible
linear combinations of variables and extracts the component that describes the variable
combinations explaining the largest amount of sample variance. This first component is
called the first principal component. The subsequent components are then extracted in the
same manner, and each is expressed as the variable combination explaining the greatest
amount of residual variance. The amount of variance explained is known as the eigenvalue.
The larger the eigenvalue, the greater the percentage of variance explained.
Following this procedure, several methods can be used to select which components should
be retained. Firstly, the size of the eigenvalues is an important determinant. According to
Kaiser (1958), all components with eigenvalues above one should be retained (Kaiser’s
criterion). Jolliffe (2002), on the other hand, suggests that Kaiser’s criterion is too strict,
and proposes retaining factors or components with eigenvalues greater than 0.7. Field
(2013), suggests that Kaiser’s criterion could be accurate when the number of items is less
104
than 30 and the communalities21 that result from the extraction are greater than 0.7 or
when the sample size is greater than 250 and the average communality is greater than 0.6.
Secondly, eigenvalues can be plotted on the Y-axis against the factor or component with
which they are associated (X-axis), giving a graphical interpretation of eigenvalues in order
of magnitude known as the scree plot. The cut-off point for selecting the number of factors
or components is known as the ‘point of inflection of the curve’, where there is a significant
change in the slope of the curve. Factors to the left of the point of inflection are retained
(Field, 2013). In the present analysis, all eigenvalues >1 were retained, because the sample
size was greater than 250 and the average communality was also greater than 0.6 (Field,
2013).
5.3.4.7 Rotation methods
When PCA is run, most items tend to load onto the first component. Rotation methods are
then used to maximise the loadings of individual variables or items onto individual factors
and equalise the importance of each component, without disrupting the underlying
solutions (Field, 2013). The choice of rotation depends on the relationship that the factors
are known to have with each other. The ‘orthogonal’ or ‘Varimax’ method allows the
components to be uncorrelated, while the ‘oblique’ or ‘oblimin’ method allows the
components to correlate (Field, 2013).
The ‘oblimin’ rotation method was selected because the components were expected to be
correlated (Field, 2013). This results in two different sets of component matrices: (1) The
‘structure matrix’, which shows the correlations between each variable and factor, and; (2)
The ‘pattern matrix’, which calculates the regression coefficient between each variable on
each component, and shows the unique contribution to each component from each
variable (Field, 2013). The factors were read from the pattern matrix because the
regression values for each items onto each component are taken into account, stabilizing
differences between measurement units and variable variances (Dugard, Todman, &
Staines, 2010).
21
Communality is the proportion of common variance present in a variable - a variable that has no
specific variance would have a communality of 1; a variable that shares none of its variance with any
other variable would have a communality of 0.
105
5.3.4.8 Factor loadings
Factor loadings are a measure of the regression coefficients between items and they tell us
about the relative contribution that a particular item makes to a component (Field, 2013).
Stevens’ (2009) recommends that items with factor loadings of 0.40 or above are retained
for samples greater than 200 participants. A factor loading value greater than 0.40 explains
16% of the variance. However, a value of 0.30 can also be used given the large sample size
collected in this study (Field, 2013). Items with factor loadings greater than 0.30 were
therefore retained because these are considered statistically meaningful with a large
sample size (Field, 2013).
SPSS allows for a set number of factors to be introduced in the analysis, if an underlying
theory exists. In this case however, no set number of factors was introduced in the analysis
in order to observe whether the same structure as the CEBQ emerged for the AEBQ. The
data presented only shows factor loadings above 0.30 (Table 5.3).
5.3.4.9 Missing data
There were no missing data, given participants were forced to respond to each item
through the design of the Survey Monkey questionnaire (i.e. all questions had to be
completed in the on-line response form, otherwise the participant could not click through
to the next page of questions).
5.3.4.10 Assumptions
PCA makes a number of assumptions (Field, 2013), and these were all tested:
(1) The sample size should be adequate. A sample size of n=708 was suitable according to
the parameters discussed in section 5.3.4.1 above.
(2) All variables must show high inter-correlation. Any variables that correlate with no
others should be eliminated, as they do not contribute to the factor structure and any
variables that are perfectly correlated (1.0) should be eliminated (Comrey & Lee, 1992).
There should be several correlation coefficients greater than 0.3 in the correlation matrix.
(3) Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy should be greater than 0.6
and Bartlett’s test of sphericity should be significant.
(4) Correlations should be linear.
106
(5) The model should fit the data well. Having less than 50% of the residuals above 0.05
represents a good model fit.
(6) Variables should be normally distributed. This assumption of normality is used to allow
generalisation of the results beyond the sample collected.
5.3.4.11 Criteria used for eliminating items
Once the PCA was run, a number of criteria were considered to eliminate items from the
questionnaire. The pattern matrix output revealed an initial ten-factor structure with
eigenvalues greater than one. The criteria used to eliminate the items from the
questionnaire were discussed with a group of experts on eating behaviour, as well as the
original authors from the CEBQ. An iterative process was used to gradually remove items
that were represented by unacceptable factor loadings of less than 0.30 (Field, 2013), items
that loaded onto several components, or items that were found to contribute poorly to the
model fit.
5.3.5 Summary statistics
Skewness and kurtosis statistics were calculated for each individual item (not shown in the
results), to test for assumptions for each obtained scale score to check for normality. Items
falling within the range of 1 and -1 for skewness and kurtosis were considered normal.
Correlations between scales were determined using Pearson’s Product Moment correlation
coefficients for normally distributed scales and Spearman's Rho for non-normally
distributed scales. A Pearson’s correlation of ±0.1 represents a small effect, ±0.3 a medium
effect and ±0.5 a large effect (Field, 2013).
5.3.6 Internal reliability
The internal reliabilities of the components derived from the PCA were assessed using
Cronbach’s alpha (Cronbach, 1951). Internal reliability describes the extent to which all the
items in a questionnaire measure the same concept or construct and hence it is connected
to the interrelatedness of the items within the test (Tavakol & Dennick, 2011). A
Cronbach’s alpha of 0.70 or greater was considered acceptable (Field, 2013).
107
5.4 Ethical approval
This study was part of the project ‘ID number 4378/001: “Validation of the AEBQ” for which
ethical approval was obtained from the UCL Research Ethics Committee (Appendix 5.8).
5.5 Results
The 47-item questionnaire was completed by 49 adults aged 21-73 years (mean age 38.2 ±
14.6), of whom 36 were women (79.6%) and 13 were men (20.4%).
5.5.1 Assessment of the factor structure
5.5.1.1 Characteristics of the sample
The 47 item AEBQ was then completed by 708 participants via the on-line questionnaire.
The participants ranged in age from 18 to 81 years (mean age 38.7±17.0). Approximately
half were men (336; 47.5%) and the mean BMI was 26.1±5.8. Most were of white ethnicity
(635; 89.7%). One-third of the sample had completed higher education (196/; 31.9%)
(Table 5.1).
Table 5.1 Characteristics of the sample used to carry out PCA of the 47-item AEBQ
(n=708)
Characteristic n (%)
Age
18 to 29 301 (42.5%)
30 to 59 300 (42.4%)
60 + 107 (15.1%)
Gender
M 336 (47.5%)
F 372 (52.5%)
BMI* n=674
Underweight 30 (4.4%)
Normal weight 328 (48.7%)
Overweight 173 (25.6%)
Obese 143 (21.2%)
Ethnicity
White 635 (89.7%)
Non-white 68 (9.6%)
Preferred not to answer 5 (0.7%)
Education n=700
School 179 (25.6%)
College 242 (34.6%)
University 279 (39.9%)
*674 (95.2% of the sample) participants had a BMI range of 15.34 to 49.87. Participants who
reported a BMI <14 or >50 were excluded as these values were felt to be unrealistic.
108
5.5.1.2 Testing for assumptions
With 47 variables resulting from the AEBQ, having 10 participants per variable would give a
sample of 470, well above the threshold of 300 required for a sufficiently stable analysis
(Field, 2013). Therefore, the sample size of n=708 was considered adequate.
All individual items on the 47-item AEBQ fell within the range of 1 and -1 for skewness and
kurtosis and were normally distributed. There was a majority of inter-correlations above
0.3 between the items, with no multi-collinearity.
The KMO measure of sampling adequacy was 0.878 which is classified as ‘great’
(Hutchenson & Sofroniou, 1999), and Bartlett’s test of sphericity was statistically significant
with χ2 (595)=12558.321, p<0.0001. No absolute residual values were above 0.05,
indicating that the model fits the data well. The retained eigenvalues were all greater than
one and the communalities ranged from 0.6 to 0.8, satisfying Kaiser’s criterion.
All assumptions outlined in the methods were met (Section 5.3.4.10).
5.5.2 Criteria used for eliminating items
PCA was run on the 47 item AEBQ (Appendix 5.7), and the criteria used for the removal of
items after each PCA run is shown in Table 5.2.
109
Table 5.2 Criteria used for the removal of items in the AEBQ factor analysis
Number of
factors
Items in order of removal Reasons for the removal of items
remaining after
PCA
Q29 “I eat less when I am tired” Q29 and Q12 loaded on a single
Q12 “I eat more when I am tired” 9 components factor, which did not make
theoretical sense
Q27 “I eat less when I'm bored” Q27 and Q32 also loaded alone on a
Q32 “I eat less when I have nothing 8 components single factor (as above).
else to do
Q15 and Q45 loaded together with
three ‘food responsiveness’ items
(Q25 “If I allowed myself, I would
eat too much”, Q18 “Even if I am
full up I find room to eat my
Q15 “I eat more when I'm bored” favourite food” and Q44 “When I
Q45 “I eat more when I have nothing 8 components see or, smell food that I like, it
else to do” makes me want to eat”) and a
‘satiety responsiveness’ item (Q40
“I cannot eat a meal if I have had a
snack just before”) onto one factor,
therefore conceptually this was an
issue.
Q43 loaded onto the same factor as
Q43 “I have a big appetite” 8 components ‘food responsiveness’ and ‘hunger’
items.
Q6 “I eat less when I'm happy” was
removed as it loaded onto the
Q6 “I eat less when I'm happy” and ‘emotional over-eating’ construct.
8 components
Q38 “I eat more when I'm happy” Q38 “I eat more when I'm happy”
was removed as it loaded onto the
‘emotional under-eating’ construct.
Q25 “If I allowed myself I would eat Q25 was considered an item that
too much” related more to ‘restraint’ than to
‘food responsiveness’, where it
loaded.
Q20 “I am difficult to please with
meals” 7 components Q20 was removed as an individual
could be considered to be difficult
to please not only with meals but
with many other things as well,
even though it loaded onto the
‘food fussiness’ construct.
Q18 loaded onto both ‘satiety
Q18 “Even if I am full up I find room to responsiveness’ and ‘food
7 components
eat my favourite food” responsiveness’ with the lowest
factor loading
A final re-run of the PCA was carried out, identifying a 35-item questionnaire. Factor
loadings for each item were obtained after ‘oblimin’ rotation and seven components were
110
retained (Table 5.3). The items that clustered onto the same factors corresponded to
similar CEBQ scales and were: Component 1 clustered the newly added ‘hunger’ (H) items
and ‘food responsiveness’ (FR) items, which loaded onto a single component [‘hunger and
food responsiveness’ (HFR)]; Component 2 clustered ‘satiety responsiveness’ (SR) items;
Component 3 clustered ‘emotional under-eating’ (EUE) items; Component 4 clustered ‘food
fussiness’ (FF) items; Component 5 clustered ‘emotional over-eating’ (EOE) items;
Component 6 clustered ‘enjoyment of food’ (EF) items; and Component 7 clustered
‘slowness in eating’ (SE) items. All of the individual items had factor loadings above 0.4,
except for “When I see or smell food that I like, it makes me want to eat”, with a factor
loading of 0.355 on component 1. The item was still retained, because it added to the
meaning of the construct, and is considered statistically meaningful with a large sample size
(Field, 2013). The item, “I often feel hungry when I am with someone who is eating” was
also retained although it loaded onto both component 1 (0.401) and component 6 (0.307);
it was retained as part of component 1 due to its higher factor loading on this component.
A comparison of the items in this final version of the AEBQ and those in the original CEBQ
items is shown in Appendix 5.9. The thirty-five items had an average communality of 0.642
and seven factors explained 64.3% of the variance in the items.
The final 35-item AEBQ with its scoring system can be seen in Appendix 5.10.
111
Table 5.3 Factor loadings of a 35-item AEBQ
Components determined through PCAa
1
Eigenvalue 2 3 4 5 6 7
(% variance explained) H + FR SR EUE FF EOE EF SE
Q26-I often notice my stomach rumbling 0.752

Q37-I often feel so hungry that I have to eat
something right away 0.737
Q42-I often feel hungry 0.69
Q46-If my meals are delayed I get light-
headed 0.66
Q9-If I miss a meal I get irritable 6.638 0.545
Q30-I am always thinking about food (19%) 0.56
Q21-Given the choice, I would eat most of the
time 0.477
Q14-I often feel hungry when I am with
someone who is eating 0.401 0.307
Q44-When I see or smell food that I like, it
makes me want to eat 0.355
Q41-I get full up easily 0.778
Q40-I cannot eat a meal if I have had a snack
just before 5.301 0.753
Q11-I often leave food on my plate at the end (15.2%)
of a meal 0.612
Q31-I often get full before my meal is finished 0.611
Q36-I eat less when I'm annoyed 0.836
3.264
Q17-I eat less when I'm worried (9.3%) 0.835
112
Q47-I eat less when I'm anxious 0.827

Q24-I eat less when I'm upset 0.825
Q22-I eat less when I'm angry 0.756
Q7-I refuse new foods at first -0.826

Q23-I am interested in tasting new food I
haven’t tasted before* 0.815
Q2-I often decide that I don’t like a food, 2.868

before tasting it (8.2%) -0.791
Q13-I enjoy tasting new foods* 0.787
Q33-I enjoy a wide variety of foods* 0.692
Q10-I eat more when I'm upset -0.871
Q8-I eat more when I'm worried -0.86
1.829
Q19-I eat more when I´m anxious (5.2%) -0.83
Q4-I eat more when I'm annoyed -0.814
Q28-I eat more when I'm angry -0.717
Q3-I enjoy eating 0.854
1.368
Q1-I love food (3.9%) 0.831
Q5-I look forward to mealtimes 0.814
Q39-I eat slowly -0.899
Q34-I am often last at finishing a meal 1.206 -0.869
Q16-I often finish my meals quickly* (3.5%) 0.775
Q35-I eat more and more slowly during the
course of a meal -0.672
*Items were reverse scored when calculating scale means and Cronbach’s alphas.
a
Factor loadings above 0.3 are presented; H: ‘hunger’; FR: ‘food responsiveness’; EOE: ‘emotional over-eating’; EF: ‘enjoyment of food’; SR: ‘satiety responsiveness’;
EUE: ‘emotional under-eating’; FF: ‘food fussiness’; SE: ‘slowness in eating’.
113
Chapter 5. Development of the ‘Adult Eating Behaviour Questionnaire
5.5.3 Summary statistics
Descriptive statistics are shown in Table 5.4 for each component (now referred to as
appetitive traits). Although the ‘hunger’ and ‘food responsiveness’ items loaded onto the
same component, they were treated as two separate constructs, because they are
considered to be distinct from a theoretical basis (Schachter, 1968; Stunkard & Messick,
1985). The data for all of the scales were normally distributed, with the exception of
‘enjoyment of food’. Appetitive traits can be grouped into ‘food approach’ scales, which
include ‘hunger’, ‘food responsiveness’, ‘emotional over-eating’ and ‘enjoyment of food’
and ‘food avoidance’ scales, which include ‘satiety responsiveness’, ‘emotional under-
eating’, ‘food fussiness’ and ‘slowness in eating’.
114
Table 5.4 Descriptive statistics of appetitive traits in the 35-item AEBQ (n=708)
Food approach scales Food avoidance scales
Food responsive- Emotional over- Enjoyment of Satiety Emotional Food
Hunger Slowness in eating
ness eating food responsiveness under-eating fussiness
Mean 3.02 3.07 2.74 3.96 2.71 2.96 2.35 2.68
Median 3.00 3.00 2.80 4.00 2.75 3.00 2.40 2.75

SD 0.75 0.78 0.88 0.74 0.79 0.85 0.79 0.87
Skewness -0.10 0.03 0.17 -0.71 0.35 -0.03 0.40 0.25
Kurtosis -0.12 -0.05 -0.23 0.89 0.08 0.01 -0.06 -0.34
SD = standard deviation
115
Correlations between the appetitive traits are shown in Table 5.5. All correlations are in the
direction that was expected, with the ‘food approach’ traits correlating positively with other ‘food
approach’ traits and negatively with the ‘food avoidance’ traits. Similarly, ‘food avoidance’ traits
correlated positively with one another.
116
Table 5.5 Correlations between appetitive traits (n=708)
Food approach traits Food avoidance traits
Food Emotional over- Satiety Emotional under-
Enjoyment of food Food fussiness Slowness in eating
responsiveness eating responsiveness eating
Hunger 0.64** 0.39** 0.33** 0.07 0.22** 0.01 -0.03
Food approach
Food
- 0.49** 0.51** -0.16** 0.10** 0.06 -0.09*

responsiveness
traits
Emotional over-
- 0.21** 0.01 -0.09* -0.01 0.02
eating
Enjoyment of
- -0.29** -0.07* -0.34** -0.16**
food
Satiety
- 0.37** 0.22** 0.43**
avoidance
responsiveness
traits
Food
Emotional
- 0.12* 0.12**
under-eating
Food fussiness - 0.11**
a
Pearson’s correlation was used for normally distributed mean scores, except for ‘enjoyment of food’ where Spearman’s
rho was used.
** Correlation is significant at the 0.01 level (2-tailed)
* Correlation is significant at the 0.05 level (2-tailed)
117
5.5.4 Reliability
Cronbach’s alphas for each appetitive trait were all above 0.7 (α range = 0.762 to 0.881),
indicating that the scales have good internal reliability (Table 5.6). Results show that the
elimination of any one item did not increase the reliability of any trait, with the exception
of the ‘slowness in eating’ scale (α=0.842). Here, the Cronbach’s alpha increased following
elimination of item Q16 “I often finish my meals quickly” (α increased to 0.846) and item
Q35 “I eat more and more slowly during the course of a meal” (α increased to 0.846).
However, both items were retained, as the Cronbach’s alpha value for the scale was still
high with the retention of both these items (α = 0.842) (Table 5.6).
118
Table 5.6 Internal reliability of appetitive trait scales for the AEBQ
Cronbach's
alpha - If
Cron-
AEBQ individual
AEBQ Items bach's
trait items
alpha
were
eliminated
Q26-I often notice my stomach rumbling 0.751
Q37-I often feel so hungry that I have to eat something right away 0.686
H Q42-I often feel hungry 0.762 0.705
Q46-If my meals are delayed I get light-headed 0.719
Q9-If I miss a meal I get irritable 0.732
Q30-I am always thinking about food 0.687
Q21-Given the choice, I would eat most of the time 0.666
FR 0.766
Q14-I often feel hungry when I am with someone who is eating 0.723
Q44-When I see or smell food that I like, it makes me want to eat 0.754
Q10-I eat more when I'm upset 0.837
Q8-I eat more when I'm worried 0.846
EOE Q19-I eat more when I´m anxious 0.877 0.845
Q4-I eat more when I'm annoyed 0.851
Q28-I eat more when I'm angry 0.87
Q3-I enjoy eating 0.767
EF Q1-I love food 0.859 0.782
Q5-I look forward to mealtimes 0.855
Q41-I get full up easily 0.644
Q40-I cannot eat a meal if I have had a snack just before 0.73
SR 0.765
Q11-I often leave food on my plate at the end of a meal 0.73
Q31-I often get full before my meal is finished 0.727
Q36-I eat less when I'm annoyed 0.855
Q17-I eat less when I'm worried 0.846
EUE Q47-I eat less when I'm anxious 0.881 0.85
Q24-I eat less when I'm upset 0.848
Q22-I eat less when I'm angry 0.876
Q7-I refuse new foods at first 0.815
Q23-I am interested in tasting new food I haven’t tasted before* 0.807
FF Q2-I often decide that I don’t like a food, before tasting it 0.852 0.844
Q13-I enjoy tasting new foods* 0.812
Q33-I enjoy a wide variety of foods* 0.831
Q39-I eat slowly 0.738
Q34-I am often last at finishing a meal 0.752
SE 0.842
Q16-I often finish my meals quickly 0.846
Q35-I eat more and more slowly during the course of a meal 0.846
* Items were reversed scored when calculating scale means and Cronbach’s alphas.
H: ‘hunger’; FR: ‘food responsiveness’; EOE: ‘emotional over-eating’; EF: ‘enjoyment of food’; SR:
‘satiety responsiveness’; EUE: ‘emotional under-eating’; FF: ‘food fussiness’; SE: ‘slowness in
eating’
119
5.6 Discussion
This chapter describes the successful development of a new self-report version of the CEBQ
for adults; the AEBQ. The AEBQ has 35 items, loading onto seven scales broadly equivalent
to the CEBQ, which capture four ‘food approach’ traits and three ‘food avoidance’ traits.
The ‘food approach’ traits are: ‘hunger’ which loaded onto the same component as ‘food
responsiveness’, ‘emotional over-eating’, and ‘enjoyment of food’. The ‘food avoidance’
traits are: ‘satiety responsiveness’, ‘emotional under-eating’, ‘food fussiness’ and ‘slowness
in eating’. The AEBQ differs from the CEBQ in that responses are in an agree/disagree
format to suit self-reporting, an additional ‘hunger’ scale is captured, and the ‘desire to
drink’ scale has been removed. The questions in the AEBQ are appropriate for adults in a
self-report format. The AEBQ shows high internal reliability, and should be useful to assess
dimensions of adult appetite that are not captured by existing questionnaires.
5.6.1 Factor structure
The new items on ‘hunger’ and ‘food responsiveness’ that were added to the AEBQ provide
additional information on appetitive traits that could only be obtained through self-report.
The AEBQ measures some overlapping qualities between these constructs, which would
explain why items load onto the same component. However, these scales were kept as
separate theoretical and empirical entities (as described in section 5.5.3), because there
appears to be enough literature to support distinguishing them as separate dimensions of
eating, i.e. ‘hunger’ (Stunkard & Messick, 1985) and ‘externality’ (Schachter, 1968). It is
worth noting however, that the ‘food responsiveness’ items have the lowest factor
loadings, explaining a smaller percentage of the variance than the ‘hunger’ items.
The ‘desire to drink’ construct was eliminated from the AEBQ, as it did not make conceptual
sense for adults and ‘drink’ could be misinterpreted as meaning ‘alcohol’ by adult samples.
The desire for soft drinks or sugar-sweetened beverages (SSBs), which has been associated
with weight gain in children and adolescents (Piernas, Barquera, & Popkin, 2014), had
previously been added to a self-report version of the CEBQ in a sample of 13 year old
Malaysian adolescents (Loh et al., 2013). However, no association between ‘desire to drink’
(with items such as “I always want soft drinks”) and weight in this age group was observed
(Loh et al., 2013). These results could be due to the fact that associations with
consumption of SSBs and BMI or fatness have not always been consistent (Johnson,
120
Mander, Jones, Emmett, & Jebb, 2007; Sweetman, Wardle, & Cooke, 2008). Evidence for
an association between the CEBQ ‘desire to drink’ scale and BMI in children is conflicting,
with some studies showing no association (Powers, Chamberlin, van Schaick, Sherman, &
Whitaker, 2006; Santos et al., 2011; Sleddens et al., 2008; Sweetman et al., 2008; Viana et
al., 2008) and others showing positive associations (Croker et al., 2011; Rodenburg et al.,
2012; Soussignan et al., 2012). Due to the overall balance of null associations with BMI, and
the lack of relevance for adults ‘desire to drink’ was not considered an appropriate scale to
include in the AEBQ. The lack of its inclusion should not diminish the predictive capability
of the AEBQ to assess appetitive traits that are salient for weight management.
Comparison of the traits assessed in the CEBQ versus the AEBQ indicates that appetitive
traits, although stable, may have different relationships with weight across the life course.
In the CEBQ, items representing ‘satiety responsiveness’ and ‘slowness in eating’ load onto
to a single construct, although they are considered to be separate theoretical entities
(Wardle, Guthrie, Sanderson, & Rapoport, 2001). By contrast, in the AEBQ, ‘satiety
responsiveness’ and ‘slowness in eating’ did load onto two separate factors, implying they
are not so closely associated in adults. However, the newly created ‘hunger’ items and
‘food responsiveness’ items did load onto a single component in the AEBQ. A similar
relationship was observed in the TFEQ-R18 (de Lauzon et al., 2004), where ‘hunger’ items
were closely related to the ‘uncontrolled eating’ factor, which although different, is
somewhat comparable to ‘food responsiveness’ as it relates to a disinhibited tendency to
eat opportunistically, such as eating in the presence of others eating, and being responsive
to the palatability of food and eating in response to negative mood (Polivy et al., 1979).
5.6.2 Correlations between sub-scales
Positive correlations were observed between individual constructs within each of the ‘food
approach’ (‘hunger’, ‘food responsiveness’, ‘emotional over-eating’ and ‘enjoyment of
food’) and ‘food avoidance’ (‘satiety responsiveness’, ‘slowness in eating’, ‘food fussiness’
and ‘emotional under-eating’) trait dimensions. Negative correlations were observed
between the two dimensions, suggesting that each dimensions measures a different set of
traits. These results are consistent with those previously shown in the CEBQ and BEBQ
(Llewellyn, van Jaarsveld, et al., 2011; Wardle, Guthrie, et al., 2001), and highlight the
validity of the measure.
121
5.6.3 Limitations
A limitation of this study relates to the selection of piloting methods used. To improve face
(content) validity of the study, adaptation of the AEBQ may have benefited from structured
cognitive testing (Banna, Vera Becerra, Kaiser, & Townsend, 2010). A more qualitative
approach such as ‘Think-Aloud’ interviews could be have been used as an efficient method
to gain insight into participants’ understanding of questionnaire items (Fox, Ericsson, &
Best, 2011). It is possible that this would have led to wordings different from those
selected for use in the final AEBQ. Also, data collection through a survey sampling company
tends to draw demographically homogeneous people to answer the questionnaire, not
allowing for significant ethnic or social differences to be observed from the sample; even
though a good mix of educational levels were represented, the sample was predominantly
white. Given that the data collection was on-line, input from those without internet access
was not possible. This also results in a lack of information available about the number of
questionnaire invitations that were sent and the number of people opting out. A study
with data collection on appetite might also encourage those with an interest in eating and
weight management to participate in the study.
Although the results show the AEBQ to be a reliable measure of seven clear constructs of
appetite (and eight theoretical dimensions), the way in which items were selected for
elimination has its limitations. For example, before reducing the number of items because
of their conceptual overlap, items could have been removed from a quantitative standpoint
according to the effect of their deletion on the Cronbach’s alphas. However, the method I
used, with the help of a panel of experts on eating behaviour, allowed for the ‘food
responsiveness’ construct to be retained. This dimension of appetite is positively
associated with weight in children (Carnell & Wardle, 2008a; Sleddens et al., 2008; Viana et
al., 2008) and infants (Llewellyn, van Jaarsveld, et al., 2011), underscoring its potential
relevance for predicting weight in adults, independently from the effects of ‘restraint’.
Although a systematic way to improve the psychometric value of the questionnaire was
carried out, the removal of different items will always affect factor loadings and measures
of reliability (Field, 2013).
Although the AEBQ appears to be a reliable measure of appetitive traits in adults,

questionnaires should also be validated, preferably against experimental measures (See
Section 2.4.1, Chapter 2). Experimental measures serve as indicators of psychometrically
defined dimensions and ensure the questionnaire is measuring what it should be measuring
122
(Allison & Baskin, 2009; Carnell & Wardle, 2007). The CEBQ was validated in an
experimental setting to demonstrate EAH as a lack of ‘satiety responsiveness’ and a
measure of a higher degree of ‘food responsiveness’ and ‘enjoyment of food’ (Carnell &
Wardle, 2007). Other ways of validating questionnaires is to apply previously validated
questionnaires and assess their convergence with the measure in question (Cepeda-Benito
et al., 2000; Krall & Lohse, 2011; van Strein, Herman, Engels, Larsen, & van Leeuwe, 2007).
Confirmation of the factor structure using Confirmatory Factor Analysis (CFA) can also serve
as a form of construct validity and will be the purpose of Study 3 in Chapter 6.
Conclusions drawn from PCA are necessarily restricted to the specific sample from which
they arise (Field, 2013; Thompson, 1951). Therefore, replication of the component
structure in different groups of adults should be obtained (Streiner & Norman, 2015).
Reliability of a scale can also vary according to the sample used, therefore it becomes
important to repeat the analysis with other samples (Field, 2013). Thus, the objective of
Study 3 in Chapter 6 will be to replicate the component structure of the AEBQ in a
validation sample, to provide evidence for the validity of the AEBQ and to provide further
evidence of its reliability.
5.6.4 Conclusions
The findings from this study demonstrates the underlying structure of the AEBQ, a self-
report measure of appetitive traits in adults, and confirms its internal reliability. The
appetitive traits identified are mostly the same as in children, but with the addition of
‘hunger’, which becomes measurable in a self-report format, and without ‘desire to drink’,
which is difficult to interpret for adults. The AEBQ therefore provides a comprehensive,
convenient, and easy-to-use measure of adult appetitive traits. The AEBQ could allow
large-scale research into those appetitive traits currently not covered by existing adult
measures, but which are strongly related to weight in infant and child populations, and
improve the understanding of the contribution of these traits to weight gain in adulthood.
The following chapter will confirm the structure of the AEBQ in a second sample of adults
and assess the associations between the appetitive traits captured by the AEBQ and weight.
123
Chapter 6. Confirmation of the factor structure of the ‘Adult Eating Behaviour Questionnaire’
Chapter 6. Study 3: Confirmation of the factor structure of

the ‘Adult Eating Behaviour Questionnaire’22
6.1 Background
Study 2 in Chapter 5 used PCA on 47 items developed from the CEBQ (Wardle, Guthrie,
Sanderson, & Rapoport, 2001) in a sample of 708 adults to develop a 35-item AEBQ, which
encompasses three ‘food approach’ scales and four ‘food avoidance’ scales. The ‘food
approach’ scales assess ‘hunger and food responsiveness’ (nine items), ‘emotional over-
eating’ (five items), and ‘enjoyment of food’ (three items). The four ‘food avoidance’ scales
assess ‘satiety responsiveness’ (four items), ‘emotional under-eating’ (five items), ‘food
fussiness’ (five items), and ‘slowness in eating’ (four items).
Having demonstrated that the AEBQ is a reliable instrument, it is important to test the
questionnaire in a different sample to ensure reproducibility and to test for construct
validity (Cole, 1987; Field, 2013). Confirmatory Factor Analysis (CFA) is a method for testing
the construct validity of a questionnaire developed through PCA (Thompson, 1951). CFA
tests the hypothesis that exists between the constructs obtained from PCA, and tests any
previous relationships which exist between the items. Also, given that the reliability of a
scale can vary according to the sample used, it is imperative to repeat the analysis with
other samples to ensure reproducibility (Field, 2013; Streiner & Norman, 2015).
In addition, it is important to explore associations between the appetitive traits measured by

the AEBQ and weight. As discussed in Chapter 2, Section 2.5.2, individual differences in
appetite are thought to help explain variation in weight across the population (Carnell &
Wardle, 2008a; Wardle, 2007), as described by the “Behavioural Susceptibility Theory”
22
A version of this chapter has been accepted for publication: Hunot, C., Fildes, A., Croker, H.,
Llewellyn, C. H., Wardle, J., & Beeken, R. J. (2016). Appetitive traits and relationships with BMI in
adults: Development of the Adult Eating Behaviour Questionnaire. Appetite.
http://dx.doi.org/10.1016/j.appet.2016.05.024. A copy of this paper is presented in Appendix 5.1.
A version of this chapter was also presented at conferences (Appendix 5.2).
124
(BST) of obesity (Llewellyn & Wardle, 2015). A large number of studies in children have
shown that appetitive traits, as measured by the CEBQ, are associated with weight across a
range of geographic locations, cultures, and ethnic groups. ‘Food approach’ scales have
been consistently positively associated with weight (Carnell & Wardle, 2008a; Santos et al.,
2011; Spence et al., 2011; Viana et al., 2008; Webber et al., 2009), while ‘food avoidance’
scales have been negatively associated with weight (Carnell & Wardle, 2008a; Santos et al.,
2011; Spence et al., 2011; Viana et al., 2008; Webber et al., 2009). Whether the
relationship between these appetitive traits and weight holds into adulthood is unknown.
The primary reason for developing the AEBQ was to enable exploration of these
relationships within adult samples.
6.2 Aim and hypothesis
The main aims of this study were to confirm the factor structure of the AEBQ in a new
sample of adults and to investigate the associations between appetitive traits measured by
the AEBQ and BMI. The study also aimed to test the reliability of the questionnaire in this
sample (internal and test-retest). Two hypotheses were tested: (1) that the structure of the
AEBQ would remain the same as that obtained from the PCA in the previous Study 2; and
(2) the relationships between appetitive traits and BMI would be similar to those found in
children using the CEBQ, i.e. that ‘food approach’ traits such as ‘hunger, ‘food
responsiveness’, ‘emotional over-eating’ and ‘enjoyment of food’, would be positively
associated with BMI; and, ‘food avoidance’ traits such as ‘satiety responsiveness’,
‘emotional under-eating’, ‘food fussiness’ and ‘slowness in eating’ would be negatively
associated with BMI.
6.3 Methods
6.3.1 Design and study population
Following a similar method to Study 2, Chapter 5, Section 5.3.4.1 (Sample 1), a second
cross-sectional survey was conducted in November 2014 (Sample 2). Adults aged 18 years
and over, who were members of an on-line survey panel (Research Now) were invited to
take part in the study, through a Survey Monkey link. By responding to the questionnaire,
participants consented to their participation in the study (Appendix 6.1).
125
6.3.1.1 Sample size calculations
Quotas for the data collection were set to ensure the sample matched the general
population with respect to the proportion of overweight and obese participants. Based on
the Health Survey for England 2013 (Health and Social Care Information Centre, 2014),
targets were set for 41% overweight and 24% obese men, and 33% overweight and 26%
obese women. The planned sample size was n=1000 participants and results from Study 2
(Sample 1), were used to calculate age quotas in order to obtain a representative sample of
overweight and obese participants similar to English obesity trends. The aim was to recruit
approximately, 200 (20%) 20 to 29 year olds; 200 (20%) 30 to 39 year olds; 250 (25%) 40 to
49 year olds; 250 (25%) 50 to 59 year olds; and 100 (10%) 60+ year olds. This would result
in approximately 100, 80, 100, 100 and 35 overweight or obese participants respectively
per age group.
Questionnaires were checked individually for the time it took to complete them. On
average, participants spent 20 to 25 minutes answering the questionnaire. As a quality
control measure, all participants who took less than 14 minutes to complete the
questionnaire were excluded, as this would not have allowed sufficient time for
participants to read and respond to the questionnaire with full comprehension after testing
the fastest answering time in different individuals. Forty-three questionnaires were
removed due to acquiescence and extreme responses in greater than 50% of the replied
questionnaires (Allison & Baskin, 2009), and 17 participants did not complete the
questionnaire (43 + 17 = 60 questionnaires removed). The majority of the participants who
were eliminated were men and women under age 30 years and men under age 40 years
old. Thus, a final sample with 954 participants was obtained (94% of those who began the
on-line questionnaire).
6.3.2 Measures
6.3.2.1 Demographic
Participants provided the same demographic information as for Study 2 (Chapter 5, Section
5.3.4.2). Briefly: ethnicity (‘White’ and ‘Non-white’); education (‘School’, ‘College’ and
‘University’); employment status (‘Employed’, ‘not employed’, and ‘disabled or retired’);
current living arrangement (‘Home owner’, ‘renting’, ‘other’). They were additionally asked
about their marital status (‘Single, ‘Co-habiting’ [married, living as married], and ‘other’
[separated, divorced, widowed]) (Appendix 6.1).
126
6.3.2.2 Anthropometric
Participants self-reported height and weight and BMI was calculated and used to categorise
weight status (Appendix 6.1). BMI was categorised into: Underweight (<18.5), normal
weight (18.5 to 24.9), overweight (25.0 to 29.9) and obese (≥30).
6.3.2.3 Appetitive traits
Participants completed the 35-item AEBQ (Sample 2), which was developed in Study 2,
Chapter 5 (Appendix 6.1).
6.3.3 Statistical analysis
Sociodemographic variables for the sample in Study 2 (Chapter 5, Section 5.3.4.2, whom I
will refer to as Sample 1), and the sample in this study (whom I will refer to as Sample 2)
were compared using Chi- squared tests for categorical variables. Means and standard
deviations were calculated for each of the AEBQ scales for Sample 2, and correlations
between scales were determined using Pearson’s Product Moment correlation coefficients
for normally distributed scales and Spearman's Rho for non-normally distributed scales. All
statistical analysis was performed using IBM SPSS Statistics version 22.0 (IBM, 2013b).
6.3.3.1 Confirmatory factor analysis
All analyses were performed using SPSS AMOS version 22.0 (IBM, 2013a). The 35 AEBQ
items23 were entered into a seven factor CFA24 (‘hunger and food responsiveness’ which
loaded onto the same component in the PCA, ‘emotional over-eating’, ‘enjoyment of food’,
‘satiety responsiveness’, ‘food fussiness’ and ‘slowness in eating’). The indicators were
loaded onto the a priori-determined corresponding factors, based on the results from the
PCA of the AEBQ in Study 2 in Chapter 5.
23
In CFA, items are termed indicators, which I will continue to use throughout this chapter, except in
the discussion.
24
In CFA, the term factor corresponds to PCA components. Factors are also known as latent
variables. So the components obtained from PCA will now be referred to as factors.
127
6.3.3.2 Input diagrams
CFA produces an input diagram (output), where single-headed arrows connect the
hypothesised factors (represented by ovals) and the measured indicators (represented by
rectangles). The regression coefficients (β- values) are shown above the arrows. Since each
measured indicator has residual variance not explained by the latent factor, each indicator
is associated with a residual (represented by the smaller circles containing an ‘e’). The
curved two-headed arrows indicate covariance between two factors. In general,
measurement errors between the indicators are assumed to be uncorrelated, but factors
are allowed to correlate between each other (Dugard et al., 2010).
The CFA model must be identified, where the ‘just-model’ represents the model where the
number of data points equals the number of parameters that must be estimated. The
number of parameters that can be calculated while maintaining an identifiable model is
k(k+1)/2, where k=observed variables. In this study, the number of observed variables was
35, therefore the number of identifiable parameters for this sample was 35(35 +1)/2=630.
Given that the model requires 77 parameters to be estimated (35 β- values, 35 residuals
and 7 covariances), this model is therefore over-identified (i.e. it contains fewer parameters
than data points and can therefore be used to test a given theory), resulting in 630-77=553
degrees of freedom, when the model fit is tested (Dugard et al., 2010).
6.3.3.3 Model fit statistics
In order to show whether the proposed model fits the data, correlations between the
variables must be correctly accounted for (Dugard et al., 2010). It is recommended to
consult several fit statistics when running CFA, to assess whether they are consistent
(Thompson, 1951). The normed fit index (NFI) indicates the degree to which the defined
model improves fit over the null model; for example, a NFI of 0.90 means the defined
model improves the fit by 90% relative to the null model (Hu & Bentler, 1999). A
comparative fit index (CFI) of 0.90 to 0.95 suggests a good model fit, as does a Root-Mean-
Square Error Approximation (RMSEA) ≥0.06 (Hu & Bentler, 1999; Thompson, 1951). The
Chi-square test is a measure of the difference between observed and expected covariance
matrices and should be non-significant. However, the Chi-square test readily reaches
significance with large sample sizes even when all other indices indicate a good fit (Dugard
et al, 2010). As in PCA, factor loadings, which tell us about the relative contribution that a
particular item makes to a factor (Field, 2013), should be greater than 0.40 (Stevens, 2009)
(Section 5.3.4.8). Out of several competing models, the model with the lowest AIC
128
(Akaike’s Information Criteria) and BIC (Bayesian Information Criterion) values is considered
the best fit to the data (Dugard et al., 2010; Field, 2013). All these model fit statistics are
presented in the results, however, the AIC and BIC were used as model selection criteria in
this study. Given the AIC is a model statistic which penalizes a model for having a greater
amount off variables by giving it a higher score, the lowest values for AIC was used to
represent the best model fit (Field, 2013).
6.3.3.4 Post-hoc modifications to the model
If the initial output from the CFA does not result in a good model fit, SPSS AMOS provides
two useful diagnostic statistics: (1) standardised residuals, and (2) modification indices.
High standardised residual values for the covariance between two variables, point towards
the relationship between these variables not being well accounted for by the model. A high
modification index, indicated by a high value of the parameter change between variables in
the model, is an indication that co-varying the error terms or residuals between these
variables (part of the same factor) should improve the model fit (Dugard et al., 2010).
Generally, error terms should not be co-varied with observed or latent variables, or with
other error terms that are not part of the same factor. Thus, the most appropriate
modification available is to co-vary error terms that are part of the same factor (Gaskin,
2016).
6.3.3.5 Reliability
Cronbach’s alpha was used to test internal reliability for each appetitive trait, with a value
greater than 0.70 indicating good reliability (Field, 2013). A sub-sample of respondents
from Sample 2, completed the AEBQ again two weeks later to assess test-retest reliability.
Test-retest reliability was assessed using intra-class correlation coefficients (ICC) (McGraw
& Fleiss, 1996) using Cronbach’s alpha model based on the average inter-item correlation
(i.e. every split-half reliability), with results presented as an average measure of the two
correlation scores. Again, values greater than 0.70 indicate good reliability. This method is
considered the best to test test-retest reliability and has been used in the development and
validation of many questionnaires (Bartle, Hill, Webber, van Jaarsveld, & Wardle, 2013;
Carnell & Wardle, 2007; Loh et al., 2013; Tanofsky-Kraff et al., 2007).
6.3.3.6 Relationships with BMI
Correlations between appetitive traits and BMI were determined using Pearson’s
correlation coefficient for normally distributed scales and Spearman's rho for the non-
129
normally distributed scales (‘enjoyment of food’). The linear associations between

appetitive traits (predictor variables) and BMI (outcome variable) were estimated using
linear regression analysis in Sample 2 with realistic BMI’s (>14 and <50). The obtained β-
values represent the slope of the regression line, whereby the greater the slope, the
stronger the relationship between the predictor and the outcome variable (Field, 2013).
The R2 value indicates the proportion of the variance in the outcome variable that is
explained by the model (Field, 2013). The model was adjusted for age and gender.
Respondents with plausible BMI values (>14 and <50) were included in the model (n=940).
The results were checked so that all the assumptions for linear regression analyses were
met:
(1) Linearity of the relationships between the predictor and outcome variables. This was
assessed visually using scatterplots.
(2) Independence of the errors (residuals). This was assessed using the Durbin-Watson test.
Values >2 indicate a negative correlation between adjacent residuals, and a positive
correlation when <2. Values <1 or >3 are considered problematic (Field, 2013).
(3) Homogeneity of variance (homoscedasticity). This was assessed visually using a

scatterplot.
(4) Normality of the errors (residuals). This was assessed using a normality plot of the
residuals.
(5) Multicollinearity of the predictors. No predictor variables should correlate too highly
with one another, e.g. above 0.95 (Field, 2013).
Ethical approval was obtained from the UCL Research Ethics Committee, and contained
within the Project ID number 5766/002: Development and validation of the self-regulation
of eating behaviour questionnaire (Appendix 6.2).
130
6.5 Results
6.5.1 Sample
Results from the on-line responses to the AEBQ were obtained from 954 adults aged 18 to
79, with a mean age of 44.5±12.9 and a mean BMI of 26.1±5.8 (Sample 2). Descriptive
characteristics of Sample 1, recruited more than one year prior to Sample 2 to allow for
comparison between these two groups, are also presented. Results are also shown for a
sub-sample of 93 respondents from Sample 2 (20 to 64 years old, mean age 48.6±12.8),
who completed the AEBQ again two weeks later to assess test-retest reliability. The
descriptive characteristics of all three samples are shown in Table 6.1. No differences in
age group, gender, BMI category, ethnicity or education were found between Samples 1
and 2.
131
Table 6.1 Descriptive statistics of adult samples used to carry out PCA (Sample 1), and
CFA and re-test sample (Sample 2)
Sample 1 Sample 2
PCA CFA Re-test

(n=708) (n=954) (n=93)
n (%) n (%) n (%)
Age
18 to 29 301 (42.5%) 166 (17.4%) 9 (9.7%)
30 to 59 300 (42.4%) 654 (68.6%) 59 (63.4%)
60 + 107 (15.1%) 134 (14.0%) 25 (26.9%)
Gender
M 336 (47.5%) 407 (42.7%) 19 (20.4%)
F 372 (52.5%) 547 (57.3%) 74 (79.6%)
BMI n=674* n=940** n=90
Underweight 30 (4.4%) 25 (2.7%) 2(2.2%)
Normal weight 328 (48.7%) 380 (39.8%) 40 (44.4%)
Overweight 173 (25.6%) 278 (29.1%) 25 (27.8%)
Obese 143 (21.2%) 257 (26.9%) 23 (24.7%)
Ethnicity n=703
White 635 (90.3%) 863 (90.5%) 91 (97.8%)
Non-white 68 (9.7%) 91 (9.5%) 2 (2.2%)
Education n=700
School 179 (25.6%) 243 (25.5%) 28 (30.1%)
College 242 (34.6%) 359 (37.6%) 29 (31.2%)
University 279 (39.9%) 352 (36.9%) 36 (38.7%)
* See Section 5.5.1.1, Table 5.1, Chapter 5.
** 940 (98.5% of the sample) participants had a BMI range of 14.99 to 48.01
6.5.2 Confirmatory Factor Analysis
The 35 AEBQ indicators, loaded onto their hypothesized underlying factors, resulting in the
input diagram from SPSS AMOS seen in Figure 6.1 (Model 1). As seen in Figure 6.1 below,
the ranges of loadings obtained for each factor were: ‘Hunger and food responsiveness’
(HFR), from 0.39 to 0.76; ‘emotional over-eating’ (EOE), from 0.70 to 0.88; ‘enjoyment of
food’ (EF), from 0.72 to 0.89; ‘satiety responsiveness’, from 0.57 to 0.83; ‘emotional under-
eating’ (EUE), from 0.65 to 0.84; ‘food fussiness’ (FF), from 0.71 to 0.89; and ‘slowness in
eating’ (SE), from 0.71 to 0.90, suggesting they were adequate, as they were above the
required value of 0.40 (except for HFR from 0.39) (Stevens, 2009) (Figure 6.1). The ‘food
approach’ traits (HFR, EOE, EF) and ‘food avoidance’ traits (SR, EUE, FF, SE) were positively
correlated within the two domains and negatively correlated between the two domains,
indicating that each group of scales measures different sets of traits (Figure 6.1).
132
Table 6.2 shows the AEBQ components with indicator names and numbers for Model 1,
which correspond to the final 35 item AEBQ obtained from the PCA in Study 2, Chapter 5
(Appendix 5.10).
133
Figure 6.1 CFA model for a 35 item, 7-factor AEBQ (Model 1)
134
Table 6.2 AEBQ components with indicator names and numbers (Model 1)
Compo- Factor
Number Indicator
nent loading
Q8 If I miss a meal I get irritable 0.69
Q20 I often notice my stomach rumbling 0.44
Q28 I often feel so hungry that I have to eat something right away 0.70
Q32 I often feel hungry 0.66
HFR Q34 If my meals are delayed I get light-headed 0.39
Q12 I often feel hungry when I am with someone who is eating 0.68
Q16 Given the choice, I would eat most of the time 0.69
Q22 I am always thinking about food 0.69
Q33 When I see or smell food that I like, it makes me want to eat 0.53
Q4 I eat more when I'm annoyed 0.78
Q7 I eat more when I'm worried 0.87
EOE Q9 I eat more when I'm upset 0.88
Q15 I eat more when I´m anxious 0.81
Q21 I eat more when I'm angry 0.70
Q1 I love food 0.87
EF Q3 I enjoy eating 0.89
Q5 I look forward to mealtimes 0.72
Q23 I often get full before my meal is finished 0.61
Q10 I often leave food on my plate at the end of a meal 0.66
SR
Q30 I cannot eat a meal if I have had a snack just before 0.57
Q31 I get full up easily 0.83
Q14 I eat less when I'm worried 0.78
Q17 I eat less when I'm angry 0.75
EUE Q19 I eat less when I'm upset 0.82
Q27 I eat less when I'm annoyed 0.80
Q35 I eat less when I'm anxious 0.84
Q6 I refuse new foods at first 0.65
Q18 I am interested in tasting new food I haven’t tasted before* 0.75
FF Q2 I often decide that I don’t like a food, before tasting it 0.89
Q11 I enjoy tasting new foods* 0.87
Q24 I enjoy a wide variety of foods* 0.71
Q13 I often finish my meals quickly* 0.76
Q25 I am often last at finishing a meal 0.88
SE
Q26 I eat more and more slowly during the course of a meal 0.71
Q29 I eat slowly 0.90
* Items were reverse scored when calculating scale means and Cronbach’s alphas.
Food approach scales: HFR: ‘hunger and food responsiveness’; EOE: ‘emotional over-eating’; EF:
‘enjoyment of food’.
Food avoidance scales: SR: ‘satiety responsiveness’; EUE: ‘emotional under-eating’; FF: ‘food
fussiness’; SE, ‘slowness in eating’.
135
Model 1 resulted in reasonable, but not good model fit: RMSEA = 0.061, NFI=0.871,
CFI=.0896, χ2(df=539) =2431.345, p<0.001 (Hu & Bentler, 1999) (Table 6.3). The CFI was
below 0.90 and the RMSEA above 0.06, which are the cut-offs that indicate a good model
fit. After looking at the modification indices and the co-varied error terms with the largest
parameter changes that were part of the same factor (Dugard et al., 2010), too many
unexplained correlations were found between the errors of the indicators on the ‘hunger
and food responsiveness’ factor. These results are seen in Figure 6.2, still as part of Model
1.
136
Figure 6.2 CFA model for a 35 item, 7-factor AEBQ with covariances between errors
(Model 1)
The modification indices and co-variances of error terms on the same factors in competing
models were examined. In the model with the lowest AIC and BIC values (i.e. the model
with the best fit to the data), too many unexplained correlations were found between error
terms. To correct this issue, the ‘hunger and food responsiveness’ factor was split into two
137
separate factors: ‘hunger’ and ‘food responsiveness’; each indicator was allowed to load on
to their respective factor (Figure 6.3) (Model 2). The ranges of factor loadings obtained for
these two new factors were: ‘hunger’ (H) from 0.44 to 0.79 and ‘food responsiveness’ (FR)
from 0.55 to 0.72, all above the minimum 0.40 (Stevens, 2009). ‘Hunger’ and ‘food
responsiveness’ were strongly correlated (0.86) (Figure 6.3). Correlations between ‘food
approach’ traits and ‘food avoidance’ traits remained the same as those observed in Figure
6.1 and Figure 6.2 for Model 1 (with and without covariances for error terms).
Table 6.3 shows the AEBQ components with indicator names and numbers for Model 2,
taken from the original AEBQ (Appendix 5.10).
138
Figure 6.3 CFA for model a 35 item, 8-factor AEBQ, with ‘hunger’ and ‘food
responsiveness’ separated (Model 2)
139
Table 6.3 AEBQ components with indicator names and numbers (Model 2)
Compo Factor
Number Indicator
nent loading
Q8 If I miss a meal I get irritable 0.57
Q20 I often notice my stomach rumbling 0.50
I often feel so hungry that I have to eat something right 0.75
H Q28
away
Q32 I often feel hungry 0.79
Q34 If my meals are delayed I get light-headed 0.44
Q12 I often feel hungry when I am with someone who is eating 0.67
Q16 Given the choice, I would eat most of the time 0.72
FR Q22 I am always thinking about food 0.72
When I see or smell food that I like, it makes me want to 0.55
Q33
eat
Q4 I eat more when I'm annoyed 0.78
Q7 I eat more when I'm worried 0.87
EOE Q9 I eat more when I'm upset 0.88
Q15 I eat more when I´m anxious 0.81
Q21 I eat more when I'm angry 0.70
Q1 I love food 0.87
EF Q3 I enjoy eating 0.89
Q5 I look forward to mealtimes 0.72
Q23 I often get full before my meal is finished 0.61
Q10 I often leave food on my plate at the end of a meal 0.66
SR
Q30 I cannot eat a meal if I have had a snack just before 0.57
Q31 I get full up easily 0.83
Q14 I eat less when I'm worried 0.78
Q17 I eat less when I'm angry 0.75
EUE Q19 I eat less when I'm upset 0.82
Q27 I eat less when I'm annoyed 0.80
Q35 I eat less when I'm anxious 0.84
Q6 I refuse new foods at first 0.65
I am interested in tasting new food I haven’t tasted 0.75
Q18
before*
FF Q2 I often decide that I don’t like a food, before tasting it 0.89
Q11 I enjoy tasting new foods* 0.87
Q24 I enjoy a wide variety of foods* 0.71
Q13 I often finish my meals quickly* 0.76
Q25 I am often last at finishing a meal 0.88
SE
Q26 I eat more and more slowly during the course of a meal 0.71
Q29 I eat slowly 0.90
* Items were reversed for scoring.
Food approach scales: H: ‘hunger’; FR: ‘food responsiveness’; EOE: ‘emotional over-eating’; EF:
fussiness’; SE: ‘slowness in eating’.
140
The 8-factor model (Model 2) produced a better model fit than Model 1: RMSEA = 0.058,
NFI=0.880, CFI=0.905, χ2(df=532) =2254.657, p<0.001. The AIC and BIC measures were
lower for Model 2 than Model 1 (Table 6.4) indicating Model 2 best fits the data.
141
Table 6.4 Model fit indices for Models 1 and 2 in CFA of the AEBQ
Exogenous
Model Items Factors NFI CFI RMSEA χ2 df AIC BIC
variables
7
Model (H+FR on a
35 42 0.871 0.896 0.061 2431.345 539 2613.345 3055.665
1 single
factor)
8
Model (H + FR as
35 43 0.880 0.905 0.058 2254.657 532 2450.657 2927.002
2 separate
factors)
FR: ‘Food Responsiveness’; H: ‘Hunger’.
AIC: ‘Akaike’s Information Criteria’; BIC: ‘Bayesian Information Criterion’; CFI: ‘Comparative Fixed Index’; χ2: ‘Chi-square’; df: ‘degrees of freedom’; NFI: ‘Normed Fixed
Index’; RMSEA: ‘Root Mean Square Error of Approximation’.
142
6.5.3 Internal and external reliability
Table 6.5 shows the Cronbach’s alphas for internal reliability and ICC for test-retest
reliability (Field, 2013). Cronbach’s alphas were all above 0.70 (α range = 0.751 to 0.904)
for internal reliability, indicating that the scales for each appetitive trait are reliable. Test-
retest reliability was also good, with all Cronbach’s alphas greater than 0.70 (α range =
0.732 to 0.910).
Table 6.5 Internal and test-retest reliability measures for the AEBQ in an adult sample
Internal reliability Test re-test reliability
Factor (Cronbach’s alphas) (ICC, 95% CI)
(n=954) (n=93)
Hunger 0.751 0.821 (0.730 to 0.881)
Food responsiveness 0.753 0.871 (0.805 to 0.914)
Emotional over-eating 0.904 0.732 (0.596 to 0.823)
Enjoyment of food 0.859 0.860 (0.789 to 0.907)
Satiety responsiveness 0.753 0.865 (0.797 to 0.911)
Emotional under-eating 0.896 0.772 (0.656 to 0.849)
Food fussiness 0.877 0.907 (0.860 to 0.939)
Slowness in eating 0.884 0.910 (0.864 to 0.940)
Food approach scales
Food avoidance scales
ICC: Intra-Class Correlation Coefficient; CI: Confidence Interval.
6.5.4 Descriptive statistics of the appetitive trait
Similarly to results in Study 2, Chapter 5, all appetitive traits were normally distributed
except for ‘enjoyment of food’, which was skewed to the right. Descriptive statistics for
each appetitive trait are presented in Table 6.6.
Table 6.6 Descriptive statistics of appetitive trait mean scores (n = 954)

H FR EOE EF SR EUE FF SE
Mean 2.92 2.98 2.74 4.00 2.61 2.83 2.29 2.62
SD 0.78 0.78 0.98 0.74 0.81 0.92 0.84 0.97
Median 3.00 3.00 2.80 4.00 2.50 2.80 2.20 2.50
Skewness -0.00 0.16 0.18 -0.69 0.30 0.12 0.40 0.34
Kurtosis -0.22 -0.28 -0.60 0.60 -0.34 -0.26 -0.35 -0.54
SD = standard deviation
143
Correlations between the appetitive traits are shown in Table 6.7. All correlations are in
the expected directions, with the ‘food approach’ traits correlating positively with each
other and negatively with the ‘food avoidance’ traits, and vice versa.
Table 6.7 Correlations between appetitive traits (n=954)

FR EOE EF SREUE FF SE
H 0.62** 0.36** 0.34** -0.04
0.12** -0.03 -0.05
avoidance approach
traits
Food
FR - 0.44** 0.55** -0.23**

-0.03 -0.10** -0.21**
EOE - 0.19** -0.14**
-0.32** 0.09** -0.14**
EF - -0.28**
-0.10** -0.36** -0.20**
SR -
0.30** 0.20** 0.47**
traits
Food
EUE - 0.03 0.21**

FF - 0.06
a
Pearson’s correlation was used for normally distributed mean scores, except for ‘enjoyment of
food’ where Spearman’s rho was used.
** Correlation is significant at the 0.01 level (2-tailed)
* Correlation is significant at the 0.05 level (2-tailed)
6.5.5 Relationships between BMI and appetitive traits
BMI was positively correlated with ‘food responsiveness’(r=0.07; p<0.05) ‘emotional over-
eating’ (r=0.26; p<0.01) and ‘enjoyment of food’ (r=0.07; p<0.05) (‘food approach’ traits),
and negatively correlated with ‘satiety responsiveness’ (r=-0.13; p<0.01), ‘emotional under-
eating’ (r=-0.20; p<0.01) and ‘slowness in eating’(r=-0.11; p<0.01) (‘food avoidance’ traits).
No relationships were found between BMI and ‘hunger’ or ‘food fussiness’ (Table 6.8).
These correlations were carried out in the complete sample with realistic BMI values
(n=940).
Table 6.8 Correlations between BMI and appetitive traits in the total adult sample
(n=940)
Food avoidance
Food approach traits
traits
H FR EOE EF SR EUE FF SE
BMI -0.03 0.07* 0.26** 0.07* -0.13** -0.20** 0.03 -0.11**
* Correlation is significant at the 0.05 level (2-tailed).

** Correlation is significant at the 0.01 level (2-tailed).
144
Table 6.9 shows the results from the multiple linear regression model predicting BMI from
the appetitive traits measured in the AEBQ, after controlling for age and gender. All
assumptions for linear regressions were met. Significant associations remained between
higher ‘food responsiveness’ (β=1.208; 95% CI: 0.710 to1.706; p<0.001), ‘emotional over-
eating’ (β=1.903; 95% CI: 1.530 to 2.777; p<0.001), ‘enjoyment of food’ (β=-1.277; 95% CI:
0.306 to1.329; p=0.002) and BMI. Significant associations also remained between lower
‘satiety responsiveness’ (β=-0.934; 95% CI: -1.405 to -0.462; p<0.001), ‘emotional under-
eating’ (β=-0.195; 95% CI: -1.689 to -0.866; p<0.001), and ‘slowness in eating’ (β=-0.672;
95% CI: -1.060 to -0.283; p=0.001) and BMI. No associations were found between either
‘hunger’ or ‘food fussiness’ and BMI. Thus, for example, an increase in one point for ‘food
responsiveness’, resulted in an increase in 1.208 points in BMI (p<0.001), which explains
5.6% of the variance.
Table 6.9 Multiple linear regression for BMI and appetitive traits (n=940)
p
Appetitive traits β coefficient (SE) 95% CI for β R2
value
‘Food approach’
Hunger 0.346 (0.259) -0.163 to 0.854 0.182 0.034
Food responsiveness 1.208 (0.254) 0.710 to 1.706 0.000 0.056
Emotional over-eating 1.903 (0.191) 1.530 to 2.277 0.000 0.125
Enjoyment of food 0.817 (0.261) 0.306 to 1.329 0.002 0.043
‘Food avoidance’
Satiety responsiveness -0.934 (0.240) -1.405 to -0.462 0.000 0.048
Emotional under-eating -1.277 (0.210) -1.689 to -0.866 0.000 0.070
Food fussiness 0.285 (0.231) -0.167 to 0.738 0.216 0.034
Slowness in eating -0.672 (0.198) -1.060 to -0.283 0.001 0.044
Note: Age and gender as covariates.
β coefficient: Un-standardised values of β; CI: Confidence Intervals; SE: Standard Error; R2:
Coefficient of determination.
6.6 Discussion
The CFA revealed that the best model fit for the AEBQ was to separate the ‘hunger’ and
‘food responsiveness’ traits, resulting in an eight factor model of the AEBQ. CFA confirmed
that although ‘hunger’ and ‘food responsiveness’ have some overlapping qualities, they
stand alone as separate dimensions of appetite in adults (Meyer & Pudel, 1972; Schachter
& Gross, 1968; Schachter, 1968; Stunkard & Fox, 1971). The final eight scales have good
internal reliability (all Cronbach’s alphas > 0.7) (Field, 2013), consistent with the results
obtained in Study 2, as well as previous studies of the CEBQ (Ashcroft et al., 2008) and the
BEBQ (Llewellyn, van Jaarsveld, et al., 2011). The AEBQ also showed good test-retest
145
reliability (all Cronbach’s alphas > 0.70). This is consistent with test-retest reliability results
reported for the CEBQ, with the exception of ‘emotional over-eating’ and ‘emotional under-
eating’ which showed lower test-retest reliability in the original children’s version of the
questionnaire (Wardle, Guthrie, et al., 2001). The increased stability of emotional eating
traits measured by the AEBQ may reflect a better ability to capture ‘emotional eating’
behaviours in adults through self-report, as opposed to parent-report for the CEBQ which
allows for potential parental bias.
Consistent with results from Study 2 (Sample 1), correlations between scores for the
appetitive traits were positively correlated with one another, while the correlations
between traits in the ‘food avoidance’ and ‘food approach’ dimensions were generally
negative. These correlations between traits are consistent with those seen in the CEBQ and
the BEBQ (Llewellyn, van Jaarsveld, et al., 2011; Wardle, Guthrie, et al., 2001).
6.6.1 Relationships with BMI and appetitive traits
Regardless of age and gender, adults with a higher BMI had higher scores for ‘food
responsiveness’, ‘emotional over-eating’ and ‘enjoyment of food’ and lower scores for
‘satiety responsiveness’, ‘emotional under-eating’ and ‘slowness in eating’. However, no
significant associations were found between BMI and the newly added construct ‘hunger’
or ‘food fussiness’. The new AEBQ ‘hunger’ scale is a measure of physical hunger (e.g.
stomach rumbles) unrelated to emotional or restraining situations as measured in the
TFEQ-R18 (Karlsson, Persson, Sjöström, & Sullivan, 2000; Stunkard & Messick, 1985). It is
possible that people find it difficult to assess their level of physical hunger, perhaps due to
its relationship to forms of ‘disinhibition’ and issues with eating regulation (Karlsson et al.,
2000). However, these null findings may also indicate that people become overweight for
reasons other than having an increased level of hunger. It is also likely that individuals
differ in their perception and interpretation of what hunger actually means (Wardle, 1987).
As seen in the factor loadings, the relationship between ‘hunger’ and ‘food responsiveness’
was very high, although the CFA ultimately revealed separating these scales provided the
best model fit. Future studies using the AEBQ will determine if it is necessary to retain the
‘hunger’ scale as an important appetitive trait in adults.
The lack of association between BMI and ‘food fussiness’ in adults might reflect the fact
that ‘food avoidance’ resulting from ‘food fussiness’ in adults could be directed towards a
much smaller number of foods, while greater variation exists in relation to children’s ‘food
146
fussiness’ (Croker et al., 2011; Spence et al., 2011; Webber et al., 2009). In adults, picky
eating which is sometimes interchangeably used with ‘food fussiness’, has been associated
with a series of anomalous eating behaviours and attitudes towards food, particularly
rejecting food based on sensory and olfactory characteristics, as well as from contact with
other food or because the food was touched by another person (Kauer, Pelchat, Rozin, &
Zickgraf, 2015). This study conducted in the USA by Kauer et al., showed that over a third
of adults reported being a ‘picky’ eater, which is higher than has been observed in children.
However, relationships between ‘food fussiness’ measured using the CEBQ and BMI have
also been inconsistent in children, with some studies finding negative associations with
weight (Loh et al., 2013; Mackenbach et al., 2012; Rodenburg et al., 2012; Spence et al.,
2011; Svensson et al., 2011), while others report no association between ‘food fussiness’
and child BMI-SDS (Cao, Svensson, Marcus, Zhang, & Sobko, 2012; Santos et al., 2011;
Soussignan et al., 2012; Sparks & Radnitz, 2012). Once again, future studies will be needed
using the AEBQ to establish if ‘food fussiness’ remains as a useful sub-scale in this measure.
Overall, ‘food responsiveness’, ‘enjoyment of food’ and ‘emotional over-eating’ are the
most common ‘food approach’ scales that show positive associations with weight in
children (Croker et al., 2011; Santos et al., 2011; Sleddens et al., 2008; Soussignan et al.,
2012; Spence et al., 2011; Viana et al., 2008; Webber et al., 2009). The most common
negative associations with weight in childhood are seen with ‘satiety responsiveness’ and
‘slowness in eating’ (Croker et al., 2011; Parkinson et al., 2010; Santos et al., 2011;
Soussignan et al., 2012; Webber et al., 2009). The relationships between appetitive traits
and BMI observed in this adult sample are consistent with findings from the child literature,
although the correlations are slightly smaller in magnitude (‘satiety responsiveness’ -0.13
[p<0.05] in this study, vs -0.19 in three to five year olds and -0.23 in eight to 11 year olds,
both p<0.001) (Carnell & Wardle, 2008a)]; which may be indicative of appetitive traits
exerting a differential influence on weight across the life course. Overall, these results
suggest that the relationships between appetitive traits and weight previously observed in
children still stand in adulthood. Any discrepancies may be a consequence of adults’
reporting their own appetite as opposed to parents’ reporting on behalf of their child. For
example, social desirability might influence adult reporting of appetite to a greater extent,
as discussed previously in Section 2.4.2.1, Chapter 2. Adults are also more likely to engage
in weight loss and weight maintenance practices than children and adolescents (Neumark-
Sztainer, Wall, Larson, Eisenberg, & Loth, 2011), and this could suppress the impact of
certain traits on BMI, whereas children typically do not exert such control over their eating.
147
6.6.2 Limitations
There are limitations to this study. As discussed in Study 2 in Chapter 6, data collection
through a survey sampling company tends to draw similar people to the questionnaire,
which prevented the investigation of ethnic differences in appetitive traits in this sample.
Demographic similarities between Sample 1 in Study 2 and Sample 2 in the present study
may be due to them being recruited from the same company, even though the two samples
were recruited over a year apart. Other limitations of using a survey sampling company
include that it might attract more health conscious participants, which could in turn bias
the results. Participants might also respond by altering their habitual behaviour due to
their heightened awareness of their behaviour from completing the questionnaire
(Hawthorne effect) (Lanigan, Wells, Lawson, & Lucas, 2001).
Weight and height were self-reported which is likely to have resulted in weight being
under-estimated and height over-estimated, leading to an under-estimation of BMI
(Gorber, Tremblay, Moher, & Gorber, 2007). Data collection through self-report could also
potentially exclude participants due to under-estimation of BMI, if the self-reports are
under-estimates to the degree that participants are incorrectly classified as underweight
(Cameron & Evers, 1990; Johnson, Beeken, Croker, & Wardle, 2014; Nawaz, Chan,
Abdulrahman, Larson, & Katz, 2001). Compared to the results for the most recent Health
Survey for England 2013 (Health and Social Care Information Centre, 2014), where 41% of
men were overweight and 24% were obese, this study obtained a sample of 36.4%
overweight and 26.8% obese men. In the case of women, the Health Survey for England
2013 results were 33% overweight and 26% obese, compared to 25.8% overweight and
28.9% of obese women in this study. Therefore, although age quotas were selected to
obtain the most representative sample, these percentages were not quite obtained, falling
mostly short in overweight representation for both men and women.
This mis-reporting of height and weight could further explain why the associations between
BMI and appetitive traits were smaller than those found in children. The cross-sectional
nature of the study precludes any inferences about causation of appetitive traits on BMI
and intra-individual continuity of appetitive traits into adulthood. Finally, because the
questions referred to eating behaviours and there is general awareness that these may be
related to weight, participants may have responded in a socially desirable way, possibly
under-reporting ‘food approach’ behaviours and over-reporting ‘food avoidance’
148
behaviours. If this occurred, it could contribute to the fact that the correlations were
smaller than in children (Carnell et al., 2013; Carnell & Wardle, 2008a).
In order to further externally validate the AEBQ for use in different adult samples, it must
be determined whether it measures what it is intending to measure (Thompson, 1951).
Although CFA is a form of validation, it is preferable to validate newly created
questionnaires against other validated appetite measurement instruments (Hyland, Irvine,
Thacker, Dawn, & Dennis, 1989) or against laboratory measures of appetite, as was done
with the CEBQ (Carnell & Wardle, 2007).
6.6.3 Conclusions
The findings from this study confirm that the structure of the AEBQ, as a self-report
measure of appetitive traits in adults, holds true in a different sample of adults. The AEBQ
is a 35-item questionnaire, which measures eight appetitive traits. The relationships
between appetitive traits and BMI in adulthood in this study were comparable to those
observed in children, indicating that approach-related and avoidance-related appetitive
traits are systematically (and oppositely) associated with BMI across the life-course. Traits
such as ‘food responsiveness’, ‘emotional over-eating’ and ‘enjoyment of food’ were
positively associated with BMI, and ‘satiety responsiveness’, ‘emotional under-eating’ and
‘slowness in eating’ were negatively associated with BMI. No associations were found
between either ‘hunger’ or ‘food fussiness’ and weight, suggesting these traits may not
relate to weight in adulthood. Future research should seek to replicate these findings in
more diverse samples and using longitudinal designs. Given the associations between the
AEBQ and BMI, the AEBQ could also be used to inform weight control interventions, by
tailoring advice based appetitive trait scores of overweight and obese individuals. This will
be the focus of Study 4 in the following chapter.
149
Chapter 7. Development of a brief appetitive trait tailored intervention in a sample of
overweight and obese adults.
Chapter 7. Study 4: Development of a brief appetitive trait

tailored intervention in a sample of overweight and obese
adults 25
7.1 Background
A key reason for exploring the relationship between appetitive traits and BMI in adulthood
is to identify potential targets for intervention. This thesis has so far demonstrated positive
correlations between BMI and ‘food approach’ traits (‘food responsiveness’, ‘emotional
over-eating’, ‘enjoyment of food’), and negative correlations between BMI and ‘food
avoidance’ traits (‘satiety responsiveness’, ‘emotional under-eating’, ‘slowness in eating’) in
a large sample of UK adults. If these traits are modifiable, they could represent targets for
weight management interventions.
The AEBQ, developed as part of this thesis, enables identification of an individual’s

appetitive trait profile. Providing individuals with feedback on their AEBQ scores and
tailored weight management advice specific to their individual trait profile, may help them
lose weight or maintain their weight. Providing individuals who are overweight or obese
with an explanation for their tendency to gain weight that isn’t routed in low willpower or
poor choices, may also help to remove some of the stigma and blame associated with
obesity, and help them to feel more confident about managing their weight (Meisel &
Wardle, 2014b).
This approach is also supported by research on the benefits of tailoring advice. Tailoring
capitalises on people’s desire to receive personalised advice and tailored information is
considered to be more relevant than generic communications, and can enhance the effects
of health-promoting messages (Kreuter, Strecher, & Glassman, 1999). A study which
tailored an individual’s weight management advice according to their genetic risk of obesity,
was previously conducted with 18 to 30-year-old university students. Students who received
25
A version of this Study and Study 5 were accepted as an abstract to present in November 2016 at
The Obesity Society in New Orleans, USA.
150
genetic feedback plus corresponding tailored weight management advice were significantly
more likely to report being ready to control their weight than students who received general
advice alone. This effect was stronger in those who received feedback stating they were at
comparatively ‘high risk’ of obesity according to their genetic test results. However this did
not translate to any difference in weight between the groups (Meisel, Beeken, Jaarsveld, &
Wardle, 2015).
A review of psychosocial predictors of successful weight loss and weight loss maintenance
highlighted a need for more research into individualized approaches to weight
management (Teixeira et al., 2005). Few studies have explored the potential to tailor
weight management advice based on a person’s appetitive traits. Previously, ‘eating in the
absence of hunger’ (EAH) measurements have been used to assign appetite awareness
training or cue exposure treatment among overweight and obese children and their
parents, in relation to eating disorders (Boutelle et al., 2011; Fisher & Birch, 2002; Tanofsky-
Kraff et al., 2008). Both treatments resulted in significant decreases in children’s binge
eating, with the food exposure treatment also resulting in significant decreases in EAH,
while the awareness training program produced no change in EAH (Boutelle et al., 2011).
Several authors have also proposed matched obesity treatments based on the eating
behaviour traits captured by the DEBQ, the ‘disinhibition’ scale of the TFEQ, the ‘cognitive
restraint’ scale of the TFEQ-R18, and the ‘Power of Food Scale’ (PFS) (Finlayson, Cecil, Higgs,
Hill, & Hetherington, 2012; van Strein, van de Laar, et al., 2007). However, it appears that
no such studies have been conducted to date and similarly no previous work has explored
the provision of tailored weight management advice based on the appetitive traits
measured by the CEBQ, and now the AEBQ.
7.2 Aims and objectives
The aim of the present study was to develop and test a brief appetitive trait feedback
intervention, to help with weight management in a group of overweight and obese adults.
The study falls under the ‘development’ phase of the Medical Research Council (MRC)
framework for developing and evaluating complex interventions (Craig et al., 2008) and
corresponds to the first five steps within the Six Steps for Quality Intervention Development
(6SQuID; Wight, Wimbush, Jepson, & Doi, 2015). The specific objectives were to:
151
1. Develop an intervention with tips corresponding to the appetitive traits measured

by the AEBQ; and,
2. Test the intervention on a small scale to determine:
a. Interest in and acceptability of the intervention, including information on
the number of tips followed, response rates, and loss to follow-up;
b. Potential impact of the intervention on weight over eight weeks following
receipt of the tips; and,
c. Participants’ experience of the intervention, including compliance with the
tips, the perceived usefulness of the tips and barriers to use of the tips.
7.3 Intervention development
The development of a low intensity, internet-based, ‘Appetitive Trait Tailored Intervention’

(ATTI), was conducted in line with the Six Essential Steps for Quality Intervention
Development (6SQuID) outlined in Table 7.1 (Wight et al., 2015). It follows Steps 1 through
5 of the 6SQuID:
Table 7.1 Six steps in public health intervention development

1. Define and understand the problem and its causes.
2. Clarify which causal or contextual factors are malleable and have greatest scope for change.
3. Identify how to bring about change: the change mechanism.
4. Identify how to deliver the change mechanism.
5. Test and refine on small scale.
6. Collect sufficient evidence of effectiveness to justify rigorous evaluation/implementation.
Source: (Wight et al., 2015)
7.3.1 Step 1. Define and understand the problem and its causes
Chapters 1 and 2 summarised the need for weight management and outlined the various
causes of obesity. The “behavioural susceptibility theory” (BST) of obesity proposes that an
individual’s appetitive traits make them more or less susceptible to certain obesogenic
environmental exposures and excess weight gain (Carnell & Wardle, 2008a). Evidence to
date has primarily come from studies of children (Carnell & Wardle, 2008a; Sleddens et al.,
2008; Spence et al., 2011; van Jaarsveld et al., 2011; Viana et al., 2008). However, the
development of the AEBQ (Study 2, Chapter 5), and the finding that appetitive traits
measured by the AEBQ are related to BMI in adulthood (Study 3, Chapter 6), suggest that
appetitive traits may also play a role in excess weight gain in adulthood.
152
7.3.2 Step 2. Clarify which causal or contextual factors are malleable and
have greatest scope for change
Research exploring the modification of appetitive traits is limited. Twin studies in children
have shown a strong genetic contribution to appetitive traits measured by the CEBQ
(Llewellyn, Trzaskowski, Plomin, & Wardle, 2013; Llewellyn et al., 2014). However, there is
also a significant environmental contribution to variation in these traits, suggesting they
would be susceptible to environmental intervention (Llewellyn & Wardle, 2015).
Furthermore, genetic contribution to phenotypic traits does not mean they cannot be
modified. The CEBQ ‘food fussiness’ trait has been shown to be highly heritable in young
children and is closely connected with the rejection of certain foods such as vegetables
(Fildes, van Jaarsveld, Cooke, Wardle, & Llewellyn, 2016). However, a large body of
evidence shows simple repeated exposure intervention can work to decrease the
behavioural expression of ‘food fussiness’ in early childhood (Daniels et al., 2015; Howard,
Mallan, Byrne, Magarey, & Daniels, 2012).
It is possible that advice targeting the modification of weight-related appetitive traits could
provide a simple personalised weight management intervention for adults. Therefore,
weight management tips were developed for the following AEBQ appetitive traits which
were considered to be potentially modifiable through behavioural or cognitive changes:
high ‘food responsiveness’, high ‘emotional over-eating’, low ‘satiety responsiveness’, and
‘fast eating’ as the inverse of ‘slowness in eating’ (i.e. low scores on ‘slowness in eating’).
No tips were developed for ‘enjoyment of food’ as it is problematic to make
recommendations for ‘not enjoying your food’ and the responses to this item were highly
positively skewed (mean 4.00±0.74), providing limited scope for change. No tips were
developed for ‘emotional under-eating’ or ‘food fussiness’, as these traits may in fact
confer protection against weight gain (Wardle, Guthrie, Sanderson, & Rapoport, 2001), and
‘food fussiness’ was not associated with BMI in Study 3. Finally, no tips were developed for
‘hunger’, as again no relationships were seen between this scale and BMI in Study 3.
The idea behind the ATTI was to provide participants with weight management tips based
on their AEBQ measured appetitive profile. ‘High’ and ‘low’ categories for each appetitive
trait were created based on the AEBQ response scale (1; ‘strongly disagree’ to 5; ‘strongly
agree’). Individuals were classed as having ‘high’ scores for the ‘food approach’ traits (‘food
responsiveness’, and ‘emotional over-eating’), if their mean score for each trait was greater
153
than 3. They were classed as having a ‘low’ score for ‘food avoidance’ traits (‘satiety
responsiveness’, and ‘slowness in eating’), if their mean score for each trait was less than 3.
During the initial development stages of the intervention, I wanted to find out if individuals
wanting to manage their weight would be interested in receiving tailored weight
management tips based on their appetitive trait profile. To achieve this aim, participants
from Sample 2 (Study 3, Chapter 6), who were members of an on-line survey panel were
asked a series of questions after completing the on-line AEBQ. Participants reported their
interest in participating in an intervention that incorporated feedback on their AEBQ
responses and tailored appetitive trait weight management tips. The full list of questions is
provided in Appendix 7.1, questions included:
• “Would you be interested in receiving feedback on your appetitive traits (i.e. styles
of eating that could make you gain or lose weight) and tips on how to manage them
accordingly?” (Response options: ‘yes’, ‘no’, ‘maybe’);
• “Would you be interested in taking part in a study looking at the effect of giving
people feedback on their appetitive traits?” (Response options: ‘very likely to take
part’, ‘likely to take part’, ‘somewhat likely to take part’, ‘probably would not take
part’);
Results were obtained from 954 participants (Sample 2, Study 3, Chapter 6). A total of
243/954 (25.5%) participants replied they would not be interested in receiving feedback on
their appetitive traits, leaving 711/954 (74.5%) participants who responded to the full
feasibility questionnaire. The full descriptive results detailing the target population’s
interest are presented in Appendix 7.1. When responding to the question “Is there any
information you think would be particularly useful for a study on appetitive trait
feedback?”, participants replied they would be interested in receiving information about
‘healthy food options’ (444/628; 46.5%), as well as tips on ‘eating self-awareness’ (373/681;
39.1%). These results were taken into account when developing the tips. When asked “do
you think that knowing about your appetitive traits would change how you eat?”, very few
said ‘no’ (45/711; 4.7%). Overall, there was enthusiasm for taking part in a study to test
these appetitive trait recommendations, with nearly two-thirds (440/711; 61.9%) being
‘very likely to take part’ or ‘likely to take part’.
154
7.3.3 Step 3. Identify how to bring about change: the change mechanism
The ATTI tips were developed using ‘Shape-Up’ as a starting point. ‘Shape-Up’ is a
behavioural weight loss/healthy lifestyle program based on Cognitive Behavioural Therapy
(CBT), that has been used in a number of settings (Beeken et al., 2013; Wardle et al., 2013;
Wardle, Liao, et al., 2001; Weight Concern, n.d.-b). CBT is the backbone for the majority of
successful multi-component weight loss interventions (Kirk et al., 2012; Wardle & Johnson,
2015). Tips were developed for each selected trait, providing simple information that
would help participants to modify their behaviours. This was done using simple language in
an accessible pdf format.
The ATTI tips leaflet was developed in conjunction with Professor Jane Wardle (Professor of
Psychology), Dr Rebecca Beeken (Senior Research Psychologist) and Dr Helen Croker
(Dietician). The leaflet consisted of three sections:
The first section gave information on the importance of a healthy weight, how “losing
weight might improve my health” (NICE Clinical Guideline 189, 2014), and how appetitive
traits may play a role in weight gain. It also contained a feedback section on “your personal
appetite profile”, which provided personalized information about the “traits that could be
making things more difficult for you”, based on participants’ responses to the AEBQ. If a
participant did not have ‘high’ scores for any of the ‘food approach’ traits or ‘low’ scores for
any of the ‘food avoidance’ traits (i.e. was not classified as having an ‘adverse’ trait), they
were told that they did not have any specific problems with these traits.
The second section of the ATTI tips leaflet contained each individual’s tips for managing
their ‘adverse’ traits. A set of weight management tips was developed for each AEBQ-
defined ‘adverse’ trait. I used techniques adapted from ‘Shape-Up’ to inform and refine the
tips. For example, for high ‘food responsiveness’ (a scale which includes the item; “I often
feel hungry when I am with someone who is eating”) the tip “Suggest doing things with
friends that don’t involve food, like going for a walk in the park” was developed. This tip is
based on ‘response substitution’ techniques to avoid external triggers to eat (Wardle &
Johnson, 2015; Wardle et al., 2013). Another example was the use of ‘stimulus control’
techniques (Hartmann-Boyce et al., 2016, 2015; Wardle et al., 2013) utilized in several of
the tips, such as “Serve yourself a meal that is the right amount for you” which was
developed as a ‘satiety responsiveness’ tip to help prevent participants from over-eating.
The ‘emotional over-eating’ tip was the only one to use ‘cognitive restructuring’ (Dalle
155
Grave et al., 2013; Rapoport et al., 2000; Wadden, Webb, Moran, & Bailer, 2012), “A lot of
people find food comforting. When you are feeling upset, annoyed or anxious this may be a
risky time. Eating something when you are feeling this way may make you feel better in the
short term, but in the long run might make you feel worse, especially if you are trying to
manage your weight”.
The newly developed tips were then discussed with other experts in eating behaviour and
the original ‘Shape-Up’ authors, reviewed for clarification, and modified. The tips were
then further refined by health psychologists who had backgrounds in energy balance.
Finally, the ATTI tips were piloted with two individuals who were asked to follow them for a
week, to obtain lay input on whether they were easy to understand and feasible.
The third and final section of the ATTI tips leaflet included information and advice about
behaviour change techniques that have been highlighted as important for successful weight
management interventions in several reviews (Campbell, Johnson, Messina, Guillaume, &
Goyder, 2011; Gupta, 2014; Hartmann-Boyce et al., 2014; Stead et al., 2015). These
techniques included self-monitoring, goal setting, and the need for social support
(Hartmann-Boyce et al., 2016; Michie, Atkins, & West, 2014; Michie et al., 2013). General
advice based on these techniques was incorporated into the information sent to the
intervention participants (Appendix 7.2 and Appendix 7.3).
The final appetitive trait tips are shown in Table 7.2. ‘Food responsiveness’ had six
corresponding tips; ‘emotional over-eating’ one tip; ‘satiety responsiveness’ five tips; and,
fast eating or ‘slowness in eating’ had three tips.
156
Table 7.2 Appetitive trait weight management tips for, ‘food responsiveness’, emotional over-eating’, ‘satiety responsiveness’, and ‘fast eating’
Having high levels of ‘food responsiveness’ • Some people are particularly susceptible to food temptations around them. Avoid buying
Chapter 7. Development of a brief appetitive trait tailored intervention in a sample

means that the sight or smell of food, or even unhealthy foods and don’t have them available in your home. This will help to take away the urge
looking at someone else eating, can make you to eat them a
want to eat
• Try to identify what specific types of food make you want to eat. So if you walk past a bakery or a
particular shop that sells treats you love on your way home, take a different route a b
• If you are with others who are eating and it is not your meal time, try having a low calorie drink
such as water with lime/orange, tea or coffee b
• Suggest doing things with friends that don’t involve food, like going for a walk in the park b
• Avoid going to the supermarket when you are hungry and use a shopping list. This will help stop
you from buying foods you don’t need a
• Some people can train themselves to resist their ‘problem foods’. You could try this. Start with
something easy. If you like salty foods, use a plain cracker, if you like sweet foods, use a plain
biscuit. Wrap it up in cling film and leave it by your desk (or somewhere where you see it often).
See how you feel about this. Repeat for 10 days, and see if your urge to eat it goes down. Then
move to a more desirable cracker or biscuit. Once you’ve done this for several days and have
of overweight and obese adults.

successfully avoided eating the food, repeat the exercise with food on a plate. Remember, sit it
out and avoid the temptation to eat. This will help you train yourself to be less responsive to food
a
• A lot of people find food comforting. When you are feeling upset, annoyed or anxious this may be
a risky time. Eating something when you are feeling this way may make you feel better in the
If you are an emotional eater you tend to eat to
short term, but in the long run might make you feel worse, especially if you are trying to manage
comfort yourself when you feel sad, or worried.
your weight c
157
• Have a plan for another way to comfort yourself that does not involve food. Identify three
alternatives to eating that might help you distract yourself and that you enjoy doing or that feel
like a treat. Talk to a friend about how you feel, play a game, go on social media, read the news,
go for a walk b
Chapter 7. Development of a brief appetitive trait tailored intervention in a sample

If you have low ‘satiety responsiveness’ you are • Some people over-eat because they have trouble recognising when they are full. Half way
less likely to notice when you are full and you through your meal, stop and try to pay attention to how full you are a
may eat more than you need.
• Serve yourself a meal that is the right amount for you. Don’t have second helpings. Put left overs
in the fridge or freezer straight away. If you need help with portion sizes, go to:
http://www.nhs.uk/Livewell/5ADAY/Pages/Portionsizes.aspx a
• You may be used to eating more than you need. Retrain yourself. It takes time to get used to
eating smaller quantities of food and feeling satisfied. Try using a smaller plate than usual a
• If someone else is serving - remember you do not have to clear your plate. Left-overs can be
thrown away or put away to save for the next day a
• Avoid ‘mindless’ eating. Don’t eat while you’re watching the television, writing an e-mail, or
reading. Stop eating if you are doing something else. Try to eat in a designated place and at set
times a
of overweight and obese adults.

If you are a fast eater, you tend not to notice • Eating slowly gives your brain the time to realise that food has entered your body and energy
when you are full, which can make you over-eat. supply is on its way. This will help you feel full. Try to eat slower than those that are eating
around you, and try to be the last one to finish your meal a
• Put your fork/spoon down in between bites. Take the time to enjoy the taste and the texture of
the foods you eat a
• Always sit down to eat your meals if you can. Standing up or rushing from one place to the next
tends to increase speed of eating a
158
a Stimulus control b Response substitution c Cognitive restructuring

The appetitive trait sheet in a pdf format was designed to be visually appealing, including
different colours and imagery, with accessible font and layout. Appendix 7.2 shows an
example of an ATTI leaflet sent to a participant who had adverse scores for all four traits:
‘food responsiveness’ and ‘emotional over-eating’, and low scores for ‘satiety
responsiveness’ and ‘slowness in eating’ (i.e. was a fast eater). A second example of the
tailored ATTI leaflet is provided for a participant who had adverse scores for three traits:
‘food responsiveness’, and low ‘satiety responsiveness’ and ‘slowness in eating’ (Appendix
7.3).
7.3.4 Step 4. Identify how to deliver the change mechanism
In order for weight loss advice to be delivered on a large-scale and cost-effectively, it needs
to be brief. Brief interventions have the potential to be integrated into routines and can
reach a broader audience (Clark, Hampson, Avery, & Simpson, 2004). Simple advice is also
preferred by participants in weight management programmes and has been shown to
improve adherence (Mata, Todd, & Lippke, 2010). The internet is an affordable medium of
delivery for weight management advice that enables greater coverage than face-to-face
intervention delivery (Arem & Irwin, 2011; Hartmann-Boyce et al., 2015).
As part of the feasibility questions completed by participants from Sample 2 (Study 3,

Chapter 6), information was obtained on how participants would like to receive the ATTI
(Appendix 6.1). Just over half (558/954; 58.5 %) responded ‘yes’ to receiving appetitive
trait-based advice for managing weight and tips on how to manage them accordingly;
153/954 (16.0%) replied ‘maybe’. Of the 711 (74.5%) participants who responded to the
feasibility questions, the majority (611/711; 85.9%) said they would prefer to receive this
information ‘via e-mail’ and just under two-thirds (452/711; 63.4%) wanted input/tips
provided ‘weekly’ over the course of the eight-week intervention. Refer to Appendix 7.1
for full descriptive results detailing the target population’s interest.
Together, existing literature and the findings from the feasibility study provide support for a
brief, tailored weight management intervention delivered via the internet. Adults reported
being interested in participating in a weight management intervention delivered via e-mail
and tailored to their individual appetitive traits. Step 5, will therefore seek to test the ATTI
in overweight and obese adults wanting to manage their weight for future refinement.
159
7.4 Methods - Step 5. Testing the intervention on a small scale
The testing of the ATTI was conducted on-line, using a pre-post design with follow-up of
participants at eight weeks. The study started between the months of June-July 2015 and
ended eight weeks later in August-September 2015.
7.4.1 Participants
Overweight and obese participants (BMI ≥25) were recruited via the Weight Concern ‘Big
Panel’ - an on-line panel of approximately 1800 people, who have first-hand experience of
being overweight and weight management attempts (Weight Concern, 2016c) (Appendix
7.4). Once ‘Big Panel’ members were initially contacted, those potentially interested in
participating in a tailored intervention were sent a second link to assess inclusion criteria to
the study (Appendix 7.5).
7.4.2 Inclusion and exclusion criteria
Individuals who were eligible to take part in the study had to be over 18 years of age, be
overweight or obese (BMI ≥25), and had to be willing to take part in the study. Individuals
were excluded if they were unable to give consent, if they were pregnant, or if they had a
terminal illness (Appendix 7.5).
7.4.3 Measures
Participants completed a questionnaire at baseline, along with questions on food

preferences which served for another fellow PhD student and were not part of this thesis
(Section 7.6). Selected questions were repeated at the end of the eight-week intervention.
This is provided in Appendix 7.4 and is described below. The questionnaire was completed
on-line using Survey Monkey.
7.4.3.1 Demographic
Demographic information was collected at baseline. Participants reported their gender,

age (in years), and marital status. Participant responses for marital status were collapsed
into three groups for analysis: ‘Single’ (‘Single’), ‘cohabiting’ (‘Married/Living with partner’),
‘other’ (‘Divorced/Separated/Widowed’) (Appendix 7.4).
Information was collected on participants’ ethnicity, and level of education. Ethnicity data
was collapsed into two categories: ‘White’ and ‘Non-white’ (‘Black’, ‘Asian’ or ‘Mixed’).
160
Education data was collapsed into three categories for analyses: ‘School’ (‘Primary
school/Secondary school/O-level/GCSE’), ‘College’ (‘A levels/Technical or trade
certificate/Diploma’), and ‘University’ (‘Undergraduate degree/Postgraduate degree’).
Participants reported their current employment status which was grouped as; ‘employed’
(‘Employed full-time/ Employed part-time/Self-employed’), ‘not employed’
(‘Unemployed/Full-time homemaker/Unpaid voluntary work/Student’), and ‘disabled or
retired’ (‘Disabled or too ill to work/Retired’). They also reported their current living
arrangement which was categorized as: ‘Home owner’ (‘Own home outright/Own home
with mortgage’), ‘renting’ (‘Rent from local authority/Housing association/Rent privately’)
or ’other’ (‘Living with parents/Living in University/College residential accommodation’)
(Appendix 7.4).
7.4.3.2 Anthropometric
Weight and height were self-reported at the start of the intervention and after the end of
the eight-week period, via e-mail. These measurements were used to calculate initial BMI
and final BMI categories. BMI values between 25 and 29.9 were classified as ‘overweight’,
and BMI values greater or equal to 30 were classified as ‘obese’.
7.4.3.3 Appetitive traits
The 35-item AEBQ was completed at baseline (Appendix 5.10, Appendix 7.4). AEBQ
responses for each participant were scored in the standard way (Study 2 and 3, Chapters 5
and 6), and scale scores were used as the basis for their appetitive trait profile, as described
above in Section 7.3.3.
7.4.3.4 Number of tips followed
The proportion of participants that received each tip was calculated, alongside the
proportion of participants who received all the tips (four tips), three tips, two tips or one
tip.
7.4.3.5 Response rate and loss to follow-up
The response rate was obtained from the Survey Monkey replies to e-mails sent to the ‘Big
Panel’. Those participants who gave their weight at the end of the eight weeks were
classified as ‘completers’. ‘Non-completers’ did not give their weight at the end of the
intervention. Withdrawals were recorded alongside reasons for drop out when provided.
161
7.4.3.6 Effects on weight
Changes in weight over the eight-week period served as preliminary data for the
intervention’s effect on weight. Weight change over the study period was categorized as:
‘kept the same weight’, ‘lost weight’, ‘gained weight’, and ‘don’t know final weight’ (‘non-
completers’ only).
7.4.3.7 Compliance and use of tips
Over the eight-week intervention, the participants were sent a weekly follow-up
questionnaire (WFQ) for assessment of compliance, perceived usefulness of the tips and
barriers to use of the tips, as well as questions on the use of other weight loss programs, via
e-mail (Appendix 7.6).
Each tip was evaluated for compliance (e.g. of a ‘food responsiveness tip: “Have you
avoided buying unhealthy foods and stopped having them in your home, so that you aren’t
tempted to eat them?”). Responses for each tip were collapsed into three categories for
analysis as: ‘All the time/Most of the time’, ‘A bit of the time’, and ‘None of the time’
(Appendix 7.6).
Participants were asked about the perceived usefulness of the tips: “Overall, do you feel
these tips are helping you to manage your ‘food responsiveness’/’emotional over-
eating’/’satiety responsiveness’/fast eating?” with response options: ‘Yes’, ‘No’, and ‘Some
of them’. Also a question on goal setting was included: “Have you made any weekly goals
for yourself to help you follow these tips?” Response options included: ‘Yes’ or ‘No’
(Appendix 7.6).
Barriers to using the tips was assessed with the question: “What has made it difficult for
you to follow this tip? Tick/strike/highlight the answers that have made it the most difficult
for you to follow this tip this week”. Possible response options included: ‘Time’; ‘Self-
motivation’; ‘Lack of support from significant others’; ‘I don’t believe it will help’; ‘This week
has included different activities from my usual routine’; ‘I didn’t find it difficult’; and ‘Other
(please specify)’. Only the tip ‘Have you tried to train yourself to resist ‘problem foods’?’
also included the response option: ‘I don't feel I'm ready to carry out this tip’ (Appendix
7.6).
Use of other weight loss programs followed alongside the tips was also assessed. Within
the WFQ, participants were also asked “are you currently following any other program to
162
help you manage your weight?”. Responses were categorised into: ‘self-directed weight
loss program (e.g. following a low fat, low carbs, counting calories or in general trying to eat
a healthy diet)’; ‘program-led weight loss (e.g. following a weight loss group, website app,
or diet book)’; ‘strict elimination diet (e.g. fasting, using replacement meals)’; ‘increased
physical activity’; ‘not following any other weight loss program’; ‘other (please specify)’
(Appendix 7.6).
7.4.4 Recruitment
In April 2015 members of the ‘Big Panel’ were contacted by e-mail and invited to take part
in an on-line questionnaire (Appendix 7.4). The e-mail contained brief information about
the ATTI study and a Survey Monkey link to the on-line questionnaire described above.
Panel members who completed this questionnaire and reported being “interested in
receiving feedback on their eating behaviour and appetite” were then contacted again via
the e-mail address they provided. In May 2015, interested participants were sent a new
Survey Monkey link containing an information sheet with further details about the study,
and a brief screening questionnaire to determine eligibility (Appendix 7.5). Eligible
participants were asked to sign a consent form and given the opportunity to ask questions
(Appendix 7.7). Recruitment lasted until June 2015, when the intervention started.
7.4.5 The intervention
The intervention lasted for eight weeks. Each participant was e-mailed a tailored ATTI
leaflet which included their personalised appetitive trait profile and corresponding tips (e.g.
in Appendices 7.2 and 7.3). They were sent a weekly reminder to continue following the
tips via e-mail over the eight weeks.
7.5 Analyses
7.5.1 Descriptive statistics
Descriptive statistics were produced to show the demographic and anthropometric

characteristics of the study ‘completers’ versus ‘non-completers’. For categorical variables,
differences were explored using cross-tabulations with Fisher's Exact test, due to the small
size of the sample (Field, 2013). An independent sample t-test was used for to explore
differences between ‘completers’ vs. ‘non-completers’ by age.
163
For participants responding to more than one WFQ, their usefulness, goal setting, barriers,
and use of other weight loss programs alongside the tips, was based on their modal
response. Percentages were calculated from the WFQ to show the proportion of
participants that had followed the tips, had found the tips helpful, had set themselves goals
for each trait, and had reported any barriers to following the tips. These were analysed
based on the total number of participants that followed each tip and who responded to a
WFQ. Frequencies were calculated on the total number of individuals who returned a WFQ
at least once over the intervention period.
To calculate the overall weight change in the sample, a paired samples t-test was used.
Data were checked for normality (Kolmogorov-Smirnov test) and for outliers using boxplots.
Effect size was calculated (Cohen's d), by dividing the mean difference by the standard
deviation of the difference, where a value of 0.2 is considered small, 0.5 medium and 0.8
large (Cohen, 1988).
Cross-tabulations with Fisher's Exact test were used to explore participants’ weight loss
categories by the number of WFQ responses they replied to, due to the small size of the
sample. Statistical analysis was performed using IBM SPSS Statistics version 22.
Ethical approval was obtained from the UCL Research Ethics Committee, and contained
within the Project ID number 4378/003: Development and pilot testing of a brief feedback
intervention concerning appetitive traits and exploratory analysis of food preferences in
relation to weight tendencies in a sample of overweight and obese adults (Appendix 7.8).
All questions regarding food preferences belonged to a fellow PhD student’s research and
were not part of this thesis.
7.7 Results
7.7.1 Number of tips followed
In total, 50/53 (94.3%) participants were provided with tips targeting ‘high food
responsiveness’; 31/53 (58.5%) were given tips for ‘high emotional over-eating’; 29/53
(54.7%) were given tips for ‘low satiety responsiveness’; and 42/53 (79.2%) were given tips
for ‘fast eating’. A third of participants 18/53 (34.0%) received two tips, 17/53 (32.1%)
164
received all four tips, 15/53 (28.3%) received three tips, and three participants received
only one tip (5.7%). For a full list of tips given to each individual participant see (Appendix
7.9).
7.7.2 Response rate and loss to follow-up
The initial e-mail was sent to all members of the ‘Big Panel’ (n~1800), and 138/1800
participants (7.7%); completed an on-line questionnaire and agreed that they would be
interested in receiving feedback on their eating behaviour and appetite. Those interested
were e-mailed and, 100/138 (72.5% response rate) participants completed a second Survey
Monkey questionnaire to establish eligibility. A total of 8/100 (8%) of participants self-
reported that they did not meet the inclusion criteria: due to pregnancy (n=4); due to
terminal illness (n=3); providing no reason (n=1). Of the remaining 92 eligible participants,
22 (23.9%) were excluded due to: going away on holiday/not being available during the
study period (n=20); suffering from severe depression (n=1); or fasting for Ramadan (n=1).
Of the 70 eligible respondents, 53 consented to take part (75% response rate) in the
development intervention study for a period of eight weeks beginning in June 2015. The
flow of participants through the study is shown in Figure 7.1.
Sample demographics are provided in Table 7.3. The majority of participants were women
(49/53; 92.5%) and of white ethnic background (48/53 [90.6%]). Participants were aged
between 27 to 76 years old (mean ± sd: 47.9±11.1), and had a BMI range of 25.4 to 56.8,
(mean ± sd: 35.7±8.11). They were predominantly married or living with a partner (47
[88.7%]), and most were employed full-time, or part-time or self-employed (38 [71.7%]).
A total of 32/53 (60.4%) participants provided their final weight and were classified as
‘completers’. No significant differences were seen between ‘completers’ and ‘non-
completers’ by age, gender, ethnicity, marital status, education, employment or living
arrangements (Table 7.3).
165
Table 7.3 Demographic characteristics and initial BMI of participants (n=53)

Group
Non-
Total difference
completers Completers
n (%), unless stated sample (test
n=21 n=32
n=53 statistic,
p)
Age (years; mean±SD) 47.87±11.14 47.19±11.75 48.31±10.89 t(51)=
-0.356,
p=0.72
Gender
M 4 (7.5%) 2 (9.5%) 2 (6.3%) p=0.521*
F 49 (92.5%) 19 (90.5%) 30 (93.8%)
Initial BMI
Overweight 19 (35.8%) 7 (33.3%) 12 (37.5%) χ2(2)=0.09
Obese 34 (64.2%) 14 (66.7%) 20 (62.5%) 6,
p=0.779**
Ethnicity
White 48 (90.6%) 19 (90.5%) 29 (90.6%) p=0.667*
Non-white 5 (9.4%) 2 (9.5%) 3 (9.4%)
Marital status
Single 4 (7.5%) 2 (9.5%) 2 (6.3%) χ2=3.913,
Co-habiting 47 (88.7%) 19 (90.5%) 28 (87.5%) p=0.205**
Other 2 (3.8%) 0 (0.0%) 2(6.3%)
Education
School 10 (18.9%) 4 (19.0%) 6 (18.8%) χ2=0.585,
College 15 (28.3%) 7 (33.3%) 8 (25.0%) p=0.869**
University 28 (52.8%) 10 (47.6%) 18 (56.3%)
Employment
Employed 38 (71.7%) 15 (71.4%) 23 (71.9%) χ2=0.302,
Not employed 6 (11.3%) 2 (9.5%) 4 (12.5%) p=1.000**
Disabled or retired 9 (17.0%) 4 (44.4%) 5 (55.6%)
Current living arrangements
Home owner 44 (83.0%) 17 (81.0%) 27 (84.4%) χ2=1.518,
Renting 8 (15.1%) 3 (14.3%) 5 (15.6%) p=0.544**
Other 1 (1.9%) 1 (4.8%) 0 (0.0%)
* Fisher’s Exact test was not computed, so p value is reported.
** Fisher’s Exact test.
Of the 21 participants who were lost to follow-up, five (23.8%) withdrew, two cited lack of
time, two cited personal circumstances, and one reported that they could not engage with
the intervention. Seven (33.3%) participants were unable to be contacted and nine (42.9%)
participants gave no reasons for failing to complete the study (Figure 7.1).
166
Initial e-mail to ‘Big Panel’

members (n~1800)
Didn’t respond (n= 1640)
Number of interested members

– replied to first SM
questionnaire (n=160)
Replied to second SM
questionnaire (n=100)
Not eligible (n=8)

No reason=1
Pregnancy=4
Terminal illness=3
Covered inclusion criteria

(n=92)
Not eligible (n=22)

Will be going away and will find it
difficult to follow the tips=18
Gave dates not appropriate to study
times=2
Severe depression=1
Fasting for Ramadan=1
Contacted further via e-mail

(n= 70)
Were sent tailored appetitive

traits tips – based on AEBQ
scores (n= 53)
Loss to follow-up (n=21)

Withdrew=5
Unable to contact=7
No reasons given=9
Participants at 8-week follow-up

(n= 32)
Figure 7.1 Flow chart of participants of the ATTI
167
7.7.3 Effects on weight
Upon checking for assumptions, one outlier was detected in the weight data (152.90 kg)
that was more than 1.5 box-lengths from the edge of the box in a boxplot. Analyses were
repeated excluding the outlier (not shown in the results) but the decision was made to
include this participant as the findings did not change.
Looking at weight change across the group of ‘completers’, differences in initial weight and
final weight were normally distributed, as assessed by Kolmogorov-Smirnov test (p=0.200).
Participant mean weight was significantly lower after the intervention (mean±sd: 90.9 kg ±
19.4 kg) compared to pre-intervention (mean±sd: 92.1 kg ± 19.8 kg; t(31) = 2.727, p=0.01)
with a medium effect size (d=0.48).
Among ‘completers’ (n=32), 20/32 (62.5%) participants lost weight, 6/32 (18.8%) kept the
same weight, and 6/32 (18.8%) gained weight (overall mean weight loss=-1.2kg, sd= 0.44).
Among those who lost weight, 10/20 (50.0%) lost less than 5% of their original weight
(Mean: 2.5 kg [1.0 kg to 3.7 kg]), 6/20 (30.0%) lost between 5 to 10% weight (Mean: 5.9 kg
[5.0 kg to 9.7 kg]) and 4/20 (20.0%) lost more than 10% of their initial weight (Mean: 15.1
kg [10.4 kg to 28.4 kg]). Percentage weight loss for each participant can also be seen in
Appendix 7.10.
The number of WFQ that participants responded to did not differ by weight change
category (χ2=6.825, p=0.109) (Table 7.4).
Table 7.4 Number of replies to weekly follow-up questionnaires by weight change

category in completers
Weight change category χ2, p
Weekly follow up questionnaire replies
n(%)
0 1-4 5-8
n=5 n=11 n=16
Same weight 2 (33.3%) 1 (16.7%) 3 (50.0%)
χ2=6.825,
Lost weight 1 (5.0%) 7 (35.0%) 12 (60.0%)
p=0.109*
Gained weight 2 (33.3%) 3 (50.0%) 1 (16.7%)
* Fisher’s Exact test.
7.7.4 Compliance, perceived usefulness and barriers to use of the tips
Compliance with the tips was assessed based on responses to the WFQ (Table 7.5). WFQ
response rates were similar regardless of the appetitive traits being targeted; 68% for those
168
who received the ‘food responsiveness’ tips and ‘emotional over-eating tips’, 72% for those
who received the ‘satiety responsiveness’ tips, and 64% for those who received the fast
eating tips provided responses for the WFQ.
Some tips appeared to be better received than others. For example, while all participants
reported following the ‘food responsiveness’ tip “train yourself to resist ‘problem foods’?”,
most participants (62.8%) reported that they did not follow the ‘food responsiveness’ tip “If
you are with others who are eating and it is not your meal time, try having a low calorie
drink such as water with lime/orange, tea or coffee”. For those who completed the WFQs
more than once, responses were similar across the questionnaires and there was no
indication that participants might have stopped following the tips as time went on (Table
7.5).
169
Table 7.5 Number of participants that reported following each tip
Number of Reported following the tip N (%)
Tips participants

All /Most of the
A bit of the time None of the time
receiving tipa time
Have you avoided buying unhealthy foods
15 15 4
and stopped having them in your home, so 34
(44.1%) (44.1%) (11.8%)
that you aren’t tempted to eat them?
Have you been able to identify specific
15 14 5
types of food that make you want to eat 34
(44.1%) (41.2%) (14.7%)
and tried to avoid them?
When you have been with others who are
‘Food 8 5 21
eating and it is not your mealtime, have 34
responsiveness’ (23.5%) (14.7%) (61.8%)
you tried having a low calorie drink?
Have you suggested doing things with 5 11 18
34
friends that do not involve eating? (14.7%) (32.4%) (52.9%)
Have you avoided going to the
22 5 7
supermarket when hungry and used a 34
(64.7%) (14.7%) (20.6%)
shopping list?
Have you tried to train yourself to resist 34
34 0 0
‘problem foods’? (100.0%)

Have you made a plan to comfort yourself
‘Emotional 19 1 1
with something other than food when you 21
over-eating’ (90.4%) (4.8%) (4.8%)
are feeling upset, annoyed or anxious?
Have you stopped and paid attention to
8 6 7
how full you feel half-way through your 21
(38.1%) (28.6%) (33.3%)
‘Satiety meal?
responsiveness’ Have you been eating the right portion
10 7 4
170
sizes for you and storing left-overs? 21

(47.6%) (33.3%) (19.1%)
Number of Reported following the tip N (%)
Tips participants All /Most of the
A bit of the time None of the time
receiving tipa time

Have you tried retraining yourself to eat 4 14 3
21
smaller quantities of food? (19.1%) (66.6%) (14.3%)
3 19
Have you stopped clearing your plate? 21 0
(14.5%) (90.5%)
10 9 3
Have you avoided mindless eating? 21
(47.6%) (42.9%) (14.5%)
Have you tried to eat slower than those 12 7 8
27
who are eating around you? (44.5%) (25.9%) (29.6%)
Have you been putting your fork down in 7 11 9
‘Fast eating’ 27
between bites? (25.9%) (40.7%) (33.3%)
Have you been sitting down for your 26 1
27 0
meals? (96.3%) (3.7%)
a
Number of participants who were following a specific tip and who responded to a questionnaire at least once.

171
Perceived usefulness of the tips was assessed from the WFQ. This was assessed with the
questions “Overall, do you feel these tips are helping you to manage your food
responsiveness/ emotional over-eating/ satiety responsiveness/ fast eating?”. Table 7.6
shows that around one quarter of participants did not find any of the ‘food responsiveness’,
‘fast eating’ and ‘satiety responsiveness’ tips helpful (23.5% to 28.6%), and the ‘emotional
over-eating’ tip was not found to be helpful by three quarters of participants (71.4%).
Table 7.6 Perceived helpfulness of the tips

Number of Perceived helpfulness of tips n(%)
participants
Helpful Not helpful Some helpful
receiving a tipa
‘Food 14 8 12
34
responsiveness’ (41.2%) (23.5%) (35.3%)
‘Emotional over- 6 15
21 N/A
eating’ (28.6%) (71.4%)
‘Satiety 8 6 7
21
responsiveness’ (38.1%) (28.6%) (33.3%)
12 7 8
(44.5%) (25.9%) (29.6%)
a
Number of participants who were following a specific tip and who ever responded to a
questionnaire.
When asked if participants had made any weekly goals for themselves for each adverse
trait they had, the majority of participants replied they had not (70.3% to 85.7%) (Table
7.7).
Table 7.7 Reported goal setting by participants receiving tips

Have you made any Replies to goal setting
Number of
weekly goals for yourself
participants
to help you follow these Yes No
receiving a tipa
tips?
9 25
‘Food responsiveness’ 34
(26.5%) (73.5)
4 17
‘Emotional over-eating’ 21
(19.0%) (81.0%)
3 18
‘Satiety responsiveness’ 21
(14.3%) (85.7%)
8 19
(29.6%) (70.3%)
a
Number of participants who were following a specific tip and who responded to a questionnaire
at least once
Barriers to following the tips were also obtained from the WFQ. Table 7.8 shows the
barriers to following the appetitive trait tips that participants reported over the eight-week
intervention period. With the exception of the tip “Have you avoided going to the
172
supermarket when hungry and used a shopping list?”, a substantial proportion (35.3% to
52.9%) of participants reported they did not find the ‘food responsiveness’ tips difficult to
follow. Few participants reported that ‘lack of time’, and lack of ‘self-motivation’ or
‘support from others’ were barriers to following the ‘food responsiveness’ tips. The
majority of participants reported that barriers to following these tips were due to ‘other
reasons’. The most common of these ‘other reasons’ included ‘force of habit’ or ‘forgetting
to carry them out’. Other practical barriers were: ‘My week has not involved eating with
friends’; ‘When I meet friends it is to have a meal’; ‘Low calorie drinks are expensive’; and
‘Family life involves having unhealthy food around’. Very few participants didn’t believe
the tips would help them. A total of eight participants (23.5%), felt they ‘did not feel ready
to carry out’ the “have you tried to train yourself to resist ‘problem foods’?” tip (Table 7.8).
The majority of participants found that ‘other reasons’ were also the barriers to not
following the ‘emotional over-eating’ tip (61.9%). Some of these ‘other reasons’ were ‘I did
not relate to that trait in me’, ‘I forgot’, and ‘time of the month’. Three participants (14.3%)
suggested they ‘didn’t find this tip difficult to follow’, and 14.3% of participants mentioned
they ‘didn’t believe this tip would help’ (Table 7.8).
In the case of the ‘satiety responsiveness’ tips, apart from the “have you been eating the
right portion sizes for you and storing left-overs?”, and the “have you stopped clearing your
plate?” tips, the largest proportion of participants ‘did not find the tips difficult to follow’
(35.0% to 47.6%). For these tips, ‘other reasons’ were also reported to be barriers (37.5%
to 59.1%). Examples of ‘other reasons’ were: ‘I don’t like to waste food’; I don’t eat that
much anyway’; ‘I used a smaller plate, so there is no need to leave food’; ‘I don’t eat more
than I need, all my food is measured’ (Table 7.8).
Finally for the fast eating tips, approximately 70.1% of participants said they ‘didn’t find it
difficult’ to follow the “have you been sitting down for your meals?” tip. For the “have you
tried to eat slower than those who are eating around you?”, and the “have you been
putting your fork down in between bites?” the participants mentioned as most common
‘other reasons’ for not following these tips (48.1% to 51.8%). The most common ‘other
reasons’ they gave were: ‘force of habit’, or ‘forgetting to carry them out’. However, they
also mentioned: ‘I don’t put my fork down between bites, because I don’t like food getting
cold’ and ‘I find it very difficult to put my fork down between bites’ (Table 7.8).
173
Table 7.8 Barriers to following appetitive trait tips
Barriers to following tips

n (%)
This week
Number of Did not
Lack of included
Appetitive participants Lack of Did not feel
Tips Did not support different
trait receiving Lack of self- believe it Other ready
find it from activities
tipsa time motivatio would reasons to
difficult significant from my
n help carry
others usual
out tip
routine
Have you avoided buying
unhealthy foods and
stopped having them in 12 3 4 3 12
34 0 0 N/A
your home, so that you (35.3%) (8.8%) (11.8%) (8.8%) (35.3%)
aren’t tempted to eat
them?
Have you been able to
Food identify specific types of
responsi- food that make you want 13 4 3 1 6 7
34 0 N/A
to eat and tried to avoid (38.2%) (11.8%) (8.8%) (2.9%) (17.6%) (20.6%)

veness
them
When you have been

with others who are
eating and it is not your
14 1 2 14
mealtime, have you tried 34 0 0 0 N/A
(41.2%) (2.9%) (5.9%) (44.1%)
having a low calorie
174
drink?
n (%)

This week
Number of Did not
Lack of included
n help carry
others usual
out tip
routine
Have you suggested
doing things with friends 16 1 1 2 14
34 0 0 N/A
that do not involve (47.1%) (2.9%) (2.9%) (5.9%) (44.1%)
eating?
Have you avoided going
to the supermarket when
3 1 2 22 5
hungry and used a 33 0 0 N/A
(9.1%) (3.0%) (6.1%) (66.7%) (15.2%)
shopping list?
Have you tried to train 8

18 2 2 1 3

yourself to resist 34 0 0 (23.5%)
(52.9%) (5.9%) (5.9%) (2.9%) (8.8%)
‘problem foods’? *
Have you made a plan to

comfort yourself with
Emotional something other than
3 1 3 1 13
over- food when you are 21 0 0 N/A
(14.3%) (4.8%) (14.3%) (4.8%) (61.9%)
eating feeling upset, annoyed or
175
anxious?
n (%)

This week
Number of Did not
Lack of included
n help carry
others usual
out tip
routine
Have you stopped and
paid attention to how full
10 3 2 6
you feel half-way through 21 0 0 0 N/A
(47.6%) (14.3%) (9.5%) (28.6%)
your meal?
Have you been eating the

2 4 4 6
right portion sizes for you 16 0 0 0 N/A
(12.5%) (25.0%) (25.0%) (37.5%)
and storing left-overs?
Satiety Have you tried retraining
7 2 1 1 3 6
responsi- yourself to eat smaller 20 0 N/A
(35.0%) (10.0%) (5.0%) (5.0%) (15.0%) (30.0%)
veness quantities of food?

Have you stopped 3 5 1 13
22 0 0 0 N/A
clearing your plate? (5.7%) (22.7%) (4.5%) (59.1%)
Have you avoided 8 6 1 7
22 0 0 0 N/A
mindless eating? (36.4%) (27.3%) (4.5%) (31.8%)
Have you tried to eat
6 2 3 3 13
slower than those who 27 0 0 N/A
(22.2%) (7.5%) (11.1%) (11.1%) (48.1%)
Fast are eating around you?
eating Have you been putting
176
6 1 5 2 2 14
your fork down in 27 0 N/A
(22.2%) (3.7%) (18.5%) (7.4%) (7.4%) (51.8%)
between bites?
n (%)

This week
Number of Did not
Lack of included
n help carry
others usual
out tip
routine
Have you been sitting 20 3 1 3
27 0 0 0 N/A
down for your meals? (74.1%) (11.1%) (3.7%) (11.1%)
a
Number of participants who were following a specific tip and who responded to a questionnaire at least once.
* This was the only tip that included this response option.

177
7.7.5 Use of other weight loss programs followed alongside the tips
A total of 17/53 (32.1%) participants did not respond to the WFQ and therefore did not
provide information about whether they were following other weight loss programs
alongside the tips. Eleven participants (30.6%) reported they were not following any other
weight loss program. Twelve participants (33.3%) described following a program-led type
diet (e.g. following a weight loss group, website app, or diet book); nine participants (25%)
followed a self-directed type program (e.g. following a low fat, low carbs, counting calories
or in general trying to eat a healthy diet); three participants (8.3%) increased their physical
activity levels; and one participant (2.8%) reported following a strict elimination-type diet
(e.g. fasting, using replacement meals) (Appendix 7.6).
7.8 Discussion
I developed a novel, brief intervention that tailors weight loss tips to individuals’ appetitive
trait scores (the ATTI), based on their AEBQ responses, and tested this in a small population
of overweight and obese adults. Initial interest in this study was low (7.7% of those
contacted) and loss to follow-up was high (40%), which raises some questions about the
acceptability of the intervention and/or study procedures. However, of those who
completed the intervention, the majority lost weight, and just over a quarter of participants
lost more than 5% of their initial body weight, which suggests the intervention holds some
promise. Responses to the WFQ identified specific tips that were more difficult to follow or
that participants found less helpful. Responses also identified difficulties engaging
participants in the goal setting element of the intervention, and specific barriers to
following the tips. This feedback could help to refine the intervention going forward.
7.8.1 Development and testing of the ATTI
The development of the ATTI followed the six steps of Wight et al.’s Quality Intervention
Development (6SQuID) model (Wight et al., 2015). The theoretical basis for this study (Step
1 of the 6SQuID) comes from the finding that appetitive traits are both measurable and
associated with BMI in adults (Study 3, Chapter 6). This study, replicates results from
studies in children showing that ‘food approach’ and ‘food avoidance’ appetitive traits are
oppositely associated with BMI in children (Carnell & Wardle, 2008a; Croker et al., 2011).
Recent research has further shown that appetitive traits are linked with eating patterns in
178
everyday life, whereby children with high ‘food responsiveness’ eat more frequently, and
children with low ‘satiety responsiveness’ eat larger quantities of food at each eating
occasion (Syrad, Johnson, Wardle, & Llewellyn, 2016). Also, appetitive traits such as ‘food
fussiness’, which have been connected with the rejection of certain foods such as
vegetables, can be modified by simple repeated exposure in early childhood (Daniels et al.,
2015; Howard et al., 2012), despite being highly heritable (Fildes et al., 2016). Taken
together this research suggests certain appetitive traits are causally associated with weight
and that tailored strategies may be effective to reduce the potential effects of appetitive
traits on weight gain. ‘Food responsiveness’, ‘satiety responsiveness’, ‘emotional over-
eating’ and ‘slowness in eating’ were the four appetitive traits selected for targeting. These
traits were selected because they were associated with BMI in adults and were thought to
be the most malleable and provide the greatest scope for modification (according to Step2
of the 6SQuID development). No tips were developed for ‘hunger’ and ‘food fussiness’ as
these traits were not found to be associated with BMI in Study 3, Chapter 6. No tips were
developed either for ‘enjoyment of food’ as it was a trait present in the majority of the
participants from Sample 2, Study 3; no ‘emotional under-eating’ tips were developed as
this trait has been mainly associated with lower weight (Wardle, Guthrie, et al., 2001).
Previously, the DEBQ has been used to examine if appetite measures are related to dieting
with the purpose of weight control in patients with Type 2 diabetes. The findings suggested
that matched treatments for obesity could be developed focusing on ‘emotional eating’ and
‘external eating’ (van Strein, van de Laar, et al., 2007). However, to date no studies
including weight management advice targeting appetitive traits measured by the DEBQ
have been published. High EAH scores, measured by laboratory assessment (Fisher & Birch,
2002), have been used to randomly allocate one of two weight management treatments to
eight to 12 year old overweight and obese children. Children who were exposed to food
cues, decreased their EAH post-treatment and six months post-treatment, although
appetite awareness training showed no change in EAH (Boutelle et al., 2011). None of the
treatments tested produced changes in BMI until the 12-month post-treatment
assessment. This study was mainly concerned with how the different treatments effected
EAH, and is therefore not a true example of a personalised treatment. After searching for
weight loss interventions tailored to appetitive traits, I found no other results. The ATTI
represents the first attempt to provide tailored weight management advice targeted at
appetite trait related behaviours.
179
The development of the appetitive trait tips themselves (step 3 of the 6SQuID) utilised
existing weight management advice developed using CBT techniques such as ‘stimulus
control’ and ‘response substitution’, and other behaviour change techniques (Hartmann-
Boyce et al., 2016; Michie, West, Brown, & Gainforth, 2014; Michie et al., 2011; Wardle &
Johnson, 2015; Wardle et al., 2013). The tips were kept simple in order to facilitate
adherence (Mata et al., 2010; Wardle et al., 2013). Simple weight loss advice has been
found to be successful in other studies (Beeken et al., 2012; Lally et al., 2008).
In order to confirm willingness to participate in a tailored appetitive trait focused weight

management intervention and to establish preferred methods of delivery, feasibility
question responses were collected from Sample 2, in Study 3, Chapter 6. The results of the
feasibility study informed the design and delivery of the intervention (Step 4 of the 6SQuI).
This feasibility study revealed participants were interested in receiving simple personalised
information about their appetitive behaviours, which could be used to help them manage
their weight. The majority of participants in the feasibility study reported that they would
like to receive intervention information via e-mail. Currently, there is a need for more cost-
effective and efficacious weight loss interventions (Arem & Irwin, 2011; Jebb et al., 2011),
and the internet has previously been shown to be potentially useful method of weight
management delivery, and provides an adequate medium for the development of
interventions (Arem & Irwin, 2011; Webb, Joseph, Yardley, & Michie, 2010). It was
therefore decided that the ATTI would be primarily delivered via e-mail.
Together Steps 1 through 4 of the 6SQuID, led to small scale testing of the ATTI (Step 5 of
the 6SQuID).
7.8.2 Effects on weight
The majority of ATTI participants who completed the intervention lost weight (-1.2 kg) over
the eight-week period, corresponding to a medium effect size (0.48); and just over a
quarter of participants lost more than 5% of their initial body weight. Although weight loss
was not the primary objective of this study, these results are promising given this was the
first small scale test of the ATTI. This finding suggests the ATTI may be effective as a weight
management intervention.
No statistical differences were seen in participants’ replies to the number of WFQ when
analysed by weight categories. While it appeared those who lost more weight replied to
more WFQ than those who didn’t lose weight or gained weight, these differences were not
180
significant. These results suggest that similar weekly input was made by participants
independent of their weight loss, showing participants’ interest in giving feedback of their
experience. However, the small sample size means caution should be applied when
interpreting these findings.
7.8.3 Compliance, perceived usefulness and barriers to use of the tips
Participants who completed the study provided responses throughout on their experience
of the intervention. They reported using most of the tips provided to them, and found
them helpful. However, certain tips were reported to be more difficult to follow and
participants found some tips less helpful than others. Tips such as “if you are with others
who are eating and it is not your meal time, try having a low calorie drink such as water
with lime/orange, tea or coffee” (‘food responsiveness’), and “suggest doing things with
friends that don’t involve food, like going for a walk in the park” (‘food responsiveness’),
were reported as not used at all by a large proportion of participants. Most participants
also reported following the ‘slowness in eating’ tip, “put your fork/spoon down in between
bites” only ‘a bit of the time’. Participants also reported not relating to the ‘emotional
over-eating’ tip. Refining the intervention might require discussing in more detail with
participants the relevance to them of each tip and potential barriers of following them. For
example, a possible barrier to following the “suggest doing things with friends that don’t
involve food, like going for a walk in the park” tip, might be that seeing friends often
involves invitations to social gatherings centred around eating (e.g. birthday meals etc.),
which was mentioned as a barrier by participants. Therefore, refining the intervention
might not require removal of tips, but possible expansion and the generation of new tips to
be adapted to individual situations (e.g. “If going out with friends involves eating, try to
make healthy food choices, and don’t get carried away by what your friends are eating”). In
general, refining of the tips and other aspects of the intervention will be required in order
to move forward with more rigorous testing of the intervention (Step 6 of the 6SQuID).
Some participants used the ‘emotional over-eating’ tip, however, the majority did not find
this tip useful. A reason might be that only one tip was developed for the ‘emotional over-
eating’ trait, possibly leaving the participant feeling s/he had less options to follow. CBT
techniques to change emotions and negative thoughts around food, present the challenge
of modifying beliefs and feelings (Wardle et al., 2013), and this may be more difficult to
achieve with a simple/single tip. Adding tips related to dealing with unhelpful thoughts
which surround ‘emotional over-eating’, could be used to improve advice related to
181
modifying this trait. However, emotional control training was found to be less effective in
promoting behaviour change than providing stress management techniques in a systematic
review of 85 internet-based studies (Webb et al., 2010). Furthermore, the majority of
participants did not set themselves any goals to follow during the intervention, which could
have hindered following the tips. Goal setting has been implicated as an important
predictor of both weight loss and maintenance, although further studies are required
(Stubbs et al., 2011). Efforts should be made to get participants more involved in following
the tips. Studies report vigilant self-monitoring of eating behaviours and weight, as
essential for long-term weight maintenance (Wing & Phelan, 2005; Wing, Crane, Thomas,
Kumar, & Weinberg, 2010). This might be achieved by increasing the number of reminders
sent. Also participants could be reminded not only to continue following the tips, but also
to set themselves goals and to write down what makes them want to eat when they
shouldn’t.
Participants were also asked what barriers prevented them following the tips. Common
replies included ‘force of habit’, or ‘forgetting to carry them out’. Healthy habits have been
shown to be acquired through repetition, and it is possible that content specific advice such
as habit-based advice could be added to the appetitive trait tips (Beeken et al., 2012; Lally
et al., 2008). This may help individuals to build them into their routines and help them
maintain the tips over the longer term. It might also help provide the motivation needed to
continue following the tips, as lack of ‘self-motivation’ was also mentioned as a barrier by a
small proportion of the participants who followed the tips. Suggesting to participants to
put a photo of themselves when they were slimmer on their fridge, might also be included
in the tips as part of a strategy to improve motivation (Hartmann-Boyce et al., 2016). Other
barriers identified were related to external situations such as having to have unhealthy
foods around the house for the sake of other family members (‘food responsiveness’), and
in a small proportion of participants ‘lack of time’ to follow the tips (for all the tips).
Barriers to following ‘satiety responsiveness’ tips related to already using smaller plates or
portion sizes and therefore being unable to feasibly reduce meal size even further. The
majority of participants found the tip recommending putting the fork down between bites
difficult to implement, again reported as based on habit and not liking to eat cold food.
Refining the tips to address some of these barriers would be beneficial to ensure their
helpfulness and increased usage.
182
7.8.4 Limitations
A number of limitations are present in this study. First of all, as in Studies 2 and 3, the self-
report nature of the information obtained is subject to different types of bias. Heights and
weights were self-reported and under-estimates of BMI calculations are likely (Cameron &
Evers, 1990), particularly at follow up. Accuracy of self-reported measures of height and
weight in older age groups is known to be reduced (Kuczmarski, Kuczmarski, & Najjar,
2001). This would imply further under-estimates might be present, given nearly a quarter
of the participants here were over the age of 60 years. This could have inflated the change
in weight observed in the study. Participants have been known to inflate results, given the
enhanced motivation to lose weight when participating in a trial (Jebb et al., 2011). This
limitation could be reduced by obtaining objective measures of height and weight in future
studies. Distinction between measures of fat mass and fat free mass have also been
correlated with different eating behaviour traits in university students, additional recording
of these measures could also provide further information about intervention weight
change, not related to BMI alone (Finlayson et al., 2012).
The response rate when ‘Big Panel’ members were first contacted was extremely low
(7.7%). First of all, it is unknown how representative of the overall overweight and obese
population in the UK the ‘Big Panel’ is. It is also possible that members’ contact details
were not up-to-date and therefore these results may not be an accurate reflection of how
many people received or opened the initial invitation e-mail. Therefore, it is unlikely that
this is a true reflection of the level of interest in the study. Once eligible participants were
identified, approximately three quarters consented to take part in the study. Possible ways
to increase questionnaire responses involve the use of incentives for participation and this
should be considered if the ATTI is developed for testing within a randomised control trial.
Opt-out methods have also been found to be useful recruitment tools (Treweek et al.,
2013).
Once the intervention started, drop-out rate was high. The lack of personal contact
resulting from the internet-based delivery of the intervention may have contributed to
these drop-out rates (Arem & Irwin, 2011). However, for those who persisted with the
ATTI, response rates to the WFQ used to assess the compliance with the tips was high for all
four traits (64% to 72%). These results reflect high participant engagement with the traits
and the study itself. Some participants mentioned that they did not relate to the profiling
received from their AEBQ answers, so personal contact would enable clarify discussion of
183
these difficulties, the tips themselves or any part of the intervention they did not feel
comfortable with. Personal contact could also increase participant motivation to lose
weight (Jebb et al., 2011). However, the implications of including personal contact would
increase the cost of the intervention, as well as the overall costs of making the intervention
itself more comprehensive. Weight loss programs which include behavioural counselling as
part of their multi-component strategies, have been found to lead to effective weight loss,
however, variable effectiveness has been observed across different studies (Kirk et al.,
2012). The cost of implementing such studies on a larger scale can be prohibitive and
therefore presents an important barrier to scalability (Coons et al., 2012). The simple,
straightforward and low cost design of ATTI means it has the potential for wide scale
dissemination.
Another limitation from this study is the use of participants from the ‘Big Panel’. Panel
members are contacted regularly to take part in research studies and weight management
interventions, potentially resulting in ‘over-use’. This could have a series of effects. First of
all the ‘Big Panel’ attracts people with an interest in weight loss, and those who “know
about the trials and tribulations of trying to lose weight” (Weight Concern, 2016a, 2016c).
Previous weight loss attempts is a known predictor of weight loss failure (Stubbs et al.,
2011; Teixeira et al., 2005). As such it may have been beneficial to have asked participants
about their previous weight loss attempts, greatest weight loss achieved and past highest
BMI. Secondly, members of the ‘Big Panel’ may have become ‘fatigued’ with participating
in previous weight management research and interventions and therefore less willing to
participate on this occasion. Recruitment from primary care settings or the general
population may result in a more representative sample, allowing better generalisation to
the wider population, and potentially higher response rates.
The WFQs also had their limitations. The time it took to answer the WFQ was a cause of
withdrawal for two participants and may also have put others off that did not respond. It
may also have acted as an effective component of the intervention for those who did
complete it as a form of self-monitoring and/or because it was a reminder and additional
contact for participants. These issues will need to be explored further in future studies.
Further limitations are present within this study. Although significant changes in weight
were seen in the ‘completers’, the study was not powered to look at any differences in
weight change. Participants also provided input as to whether they were following other
weight loss programs alongside the tips. This was also self-reported and therefore this data
184
also has its limitations. Nearly a quarter of participants didn’t follow any other program,
and approximately one third did not provide any information on whether they had followed
other recommendations. However, approximately half reported using some other form of
weight loss program at the same time as they followed the tips. It is not possible to know
whether the changes in weight were therefore due to the tips or accompanying weight loss
methods. Also, given the small number of participants that took part in the intervention
(n=53), the significant results in change in weight achieved through following the ATTI for
eight weeks, should be interpreted with caution (Lancaster, Dodd, & Williamson, 2004). A
further limitation was the lack of a control group, which could have been used to show the
differences in weight obtained between a group that received an individualised appetitive
trait profile and corresponding tips, versus another group that didn’t and only received
general weight loss information. However, this study could serve as the basis for a future
randomised controlled trial (Step 6 of the 6SQuID: rigorous implementation of the study),
which is beyond the scope of this thesis.
This study is also limited by the fact that the majority of participants were women [n=49
(92.5%)]. Gender bias is common in weight loss studies (Jebb et al., 2011; Provencher et al.,
2004). It is more common for women to want to lose weight than men (Nicklas, Huskey,
Davis, & Wee, 2012; Provencher et al., 2004; Wardle & Johnson, 2002), even with a greater
proportion of men than women being overweight (Provencher et al., 2004). Also, men tend
to show different patterns of weight loss and be more successful at losing than women, so
findings associated with one gender are not necessarily possible to extrapolate to the other
(Wardle & Johnson, 2015). Future studies should attempt to recruit a more proportioned
sample, and any observed gender- differences could serve to better tailor future weight
management interventions (Kim et al., 2015).
Finally, this study was mainly carried out to develop and test a tailored intervention based
on individualised appetitive trait feedback, to determine compliance with the tips, including
perceived usefulness and barriers to using the appetitive trait tips. The results obtained
were mainly descriptive questionnaire-based reports of participants’ experience of
following the tips. No in-depth information regarding participants’ experiences of following
the tips can be obtained through questionnaire data. Therefore, Study 5 of this thesis will
qualitatively assess participants’ experiences of following ATTI.
185
7.8.5 Conclusions
This study involved the development of a brief intervention, ATTI, designed to provide
individuals with a profile of their appetitive traits (measured by the AEBQ) and
corresponding personalised weight management tips. Appetitive trait tips were developed
based on CBT and behaviour change techniques that serve to help individuals modify the
behavioural expression of appetitive traits. The ATTI was tested in small-scale, internet-
based eight-week study. The majority of participants reported finding most of the tips
helpful, with the exception of the single ‘emotional over-eating’ tip and two of the ‘food
responsiveness’ tips not followed. Improvement to these and other tips are necessary if
this intervention is to be taken forward. The most common barriers identified were related
to ‘force of habit’, or ‘forgetting to carry the tips out’, suggesting that incorporating habit-
based techniques and providing additional reminders could improve future ATTI adherence.
An average weight loss of 1.2 kg was achieved by ‘completers’, suggesting a small effect of
the intervention on weight. This is promising, however the study was not powered to
formally test effectiveness, and lack of a control group prevents generalisability of the
results. The next chapter of this thesis (Study 5) will further explore participants’
experiences of taking part in ATTI through in-depth qualitative interviews.
186
Chapter 8. Participant experiences of a brief appetite-based weight management intervention
Chapter 8. Study 5: Participant experiences of a brief

appetite-based weight management intervention (ATTI)26
8.1 Background
The preliminary findings from Study 4 (Chapter 7), suggested that overweight and obese
participants who completed an eight-week intervention including tailored appetitive trait
feedback (the ATTI), followed most of the tips provided. On average, participants lost 1.2
kg over the intervention period. However, participants also reported a number of barriers
to following these tips. The most common barriers described were ‘force of habit’ or
‘forgetting to carry them out’. There were also specific tips that were not followed by
participants (e.g. ‘food responsiveness’ and ‘emotional over-eating’ tips). The present
chapter will further explore the experiences of participants following the intervention to
obtain a deeper understanding of the challenges they faced, what they liked and why. This
will inform development of the intervention going forward, in line with Step 6 of the
6SQuID (Wight et al., 2015).
8.2 Aim
The aim of this study was to qualitatively explore participants’ experience of the eight-week
ATTI, including barriers and facilitators to compliance.
26
A version of this Study and Study 4 were accepted as an abstract to present in November 2016 at
The Obesity Society in New Orleans, USA.
187
8.3 Methods
8.3.1 Study design
A qualitative methodology was chosen to explore participants’ experiences of the ATTI from
Study 5. Benefits from quantitative questionnaires such as those given to participants in
the previous Chapter 7, include being able to obtain information that can be analysed to
describe the general characteristics of the sample, as well as those questions asked relevant
to the study. However, they are unable to capture the level of detail obtained from
qualitative semi-structured interviews. This in-depth exploration enables a deeper insight
into participants’ involvement with the study, whether they followed the ATTI and the tips,
and whether they found these to be beneficial for weight management purposes.
Semi-structured interviews were selected and considered an appropriate method for

exploring participants views, experiences, beliefs and motivations of following the ATTI
(Gill, Stewart, Treasure, & Chadwick, 2008). Interviews were chosen over focus groups as
the choice methodology, as I was interested in individual’s experiences rather than a group
overview. Plus, it allowed for the interviews to be held via phone, which permitted for the
remainder of the study to be carried out without any personal contact. The interviews
were conducted with a sub-sample of participants at the end of the intervention period. I
conducted all of the interviews in this study, having personal experience of working in
clinical settings with overweight and obese patients and previous qualitative research
experience.
8.3.2 Participants & recruitment
Following the ATTI, all participants (n=53) who started the intervention were contacted via
e-mail (by me; CH) and invited to take part in a qualitative interview, including those who
formally withdrew from the intervention (n=5). From the beginning of the study,
participants were aware that they would be contacted at the end of the eight weeks to be
interviewed if they had agreed to this at the time of consent (Appendix 7.5). I aimed to
obtain a broad range of views and to continue interviewing until data saturation was
achieved (Morse, 1995).
Participants ID number, gender (male = ‘M’; female = ‘F’) and weight loss category (lost
weight = ‘LW’; same weight = ‘SW’; gained weight = ‘GW’); and whether they were
‘engaged’ or ‘non-engaged’ with the intervention (engaged = ‘E’; non-engaged = ‘NE’),
188
identify the participants involved in the study. Participants’ weight loss categories were
calculated from their self-reported weight and height at the beginning and end of the
intervention (Study 4, Section 7.4.3.2, Chapter 7).
8.3.3 Interview topic guide
The interview topic guide was developed to include a series of open-ended questions
(Appendix 8.1). The guide covered areas such as the participant’s motivation for taking part
in the study, their understanding of the information they had been given, level of
agreement with the appetitive trait profiling, questions about each tip they had followed,
and reasons for having been successful or not at following the tips.
The interview guide was piloted with two lay overweight individuals who had followed the
tips for a week (Section 7.3.3, Chapter 7). Interviews were practiced to have minimum
input and to prompt participant replies only when applicable (Oppenheim, 2003). They
were carried out by phone between the months of August and September 2015 and
recorded using an electronic recorder, with only myself and the participant present to
ensure anonymity.
8.4 Analyses
8.4.1 Thematic analysis of interviews
The ConsOlidated Criteria for REporting Qualitative Research (COREQ) checklist was
followed throughout (Tong, Sainsbury, & Craig, 2007)(Appendix 8.2). A professional
transcription company (Devon Transcription) transcribed verbatim 18 of the 21 interviews.
I completed the remaining three transcriptions in order to familiarize myself with the
process and the data (Braun & Clarke, 2006). To maintain anonymity, participant’s ID
numbers rather than names or other identifying information were used throughout the
transcription process.
A thematic analysis approach was used (Braun & Clarke, 2006). Thematic analysis is
independent of theory, and allows themes to emerge from the interviews using an
inductive approach (‘bottom up’ approach) closely linked to the data. The six phases of
Braun and Clark’s (2006) thematic analysis were followed: familiarization, generation of
codes, searching for themes, reviewing themes, defining the themes, and writing the
report.
189
8.4.2 Coding the interviews
Initial coding of the interviews began through the familiarization process of reading and re-
reading the transcripts, as well as listening to the recorded transcripts to check for
transcription mistakes. All transcripts were imported into NVivo (QSR International Pty Ltd,
2012), a platform used with unstructured data to facilitate coding and analysis. Initial
coding was carried out using five interviews to provide examples of the generated codes.
Amending of the codes was then carried out with a group of researchers (RJB), (AF) and
(FJ), and a final list of codes was agreed upon. All transcripts were then coded by selecting
the text which captured the intended context/quote. Initially, one quote in the text could
be assigned to different codes. All quotes were then revisited, until they were assigned to a
single code. A total of five transcripts were selected for coding comparison using kappa for
inter-rater agreement by a second researcher.
8.4.3 Data saturation
Interviews were carried out until saturation of themes in the data was reached. Saturation
was obtained after 18 interviews, however I carried out three further interviews to ensure
no new information was obtained and to increase the richness of the data through detailed
description (Morse, 1995).
8.4.4 Themes
After coding in NVivo had taken place, codes were grouped into themes that related to the
experience of the intervention. Additional themes arose from the data but were excluded
from the current analyses as they did not contribute to the aim of this study. For example,
some participants discussed at length the benefits and downfalls of different types of
weight loss programs, not in relation to this intervention. Other participants described
what meanings they gave to food. Themes were checked across the sample to ensure they
could be applied to those interviewed.
Ethical approval was obtained from the UCL Research Ethics Committee; Project ID number
4378/003 (Appendix 7.8).
190
8.6 Results
8.6.1 Participants
A total of 21 out of 32 (65.6%) ‘completers’ from the ATTI study agreed to take part in semi-
structured telephone interviews. Appendix 8.3 shows a detailed list of participants’
characteristics. The interviews lasted between 20 to 48 minutes (39 minutes on average).
The participants who took part in the semi-structured interviews did not differ from the
overall sample in terms of age, gender, BMI category, ethnicity, marital status, level of
education, employment, or living arrangements. There was no significant difference by
weight categories between those who were interviewed and not interviewed (χ2=3.410;
p=0.270). Those interviewed were significantly more likely to have replied to ‘5 to 8 WFQ’
14/21 (66.7%), than those not interviewed (‘5 to 8 WFQ’ 2/21 [33.3%]) (χ2=12.166;
p=0.001).
Of those interviewed, four participants reported following the tips for a short period of
time (approximately two weeks overall), but still liked the study, though two did not relate
to the tips (Participant 35, F, LW, and Participant 37, F, LW). Two participants did not like
the intervention (Participant 16, M, SW, and Participant 8, F, LW, who followed the
intervention for four to five weeks). Finally, one participant mentioned she did not follow
the tips because she was not ‘ready to follow them’ (Participant 4, F, GW). These five
participants were classified as ‘non-engaged’ participants (NE), and their views are
expressed in the results (See section 8.3.2).
8.6.2 Themes
Three main themes emerged from the data with the influence of their own sub-themes:
(1) Experience of the intervention: (i) Engaging with the tips and materials; (ii) The
importance of tailoring; (iii) Focus on drivers of behaviour change; (iv) Too low intensity: a
desire for more information; and, (v) The role of personal contact;
(2) Consequences of the intervention: (i) Increased self-awareness; (ii) Behaviour change;
(iii) Physical consequences;
(3) Barriers and facilitators to adherence: (i) Routines and habits; and, (ii) Social networks.
191
Themes were identified from thematic analysis of the semi-structured interviews. They are
detailed below, and illustrative quotes are included where appropriate with participants’ ID
number, gender, weight loss category and level of engagement provided in brackets to add
context to the quotes. Very few differences were seen between participants’ responses by
whether they had lost weight or not lost weight during the intervention. When differences
by weight were seen they are reported. More differences were seen between participants
who engaged with the intervention (n=16) vs. those who did not (n=5), and these are
highlighted below.
8.6.2.1 Experience of the intervention
The first theme was related to the experience of the ATTI. This theme was influenced by
five key sub-themes: (i) Engaging with the tips and materials; (ii) The importance of
tailoring; (iii) Focus on drivers of behaviour change; (iv) Too low intensity: a desire for more
information; and, (v) The role of personal contact.
Engaging with the tips and materials
Engaged participants reported that getting involved with the study was in part because the
tips were very clear, simple and easy to understand: “They’re very nicely laid out as well and
they’re very, like, colourful and they grab your attention. That’s what I noticed about them
first of all.” (Participant 10, F, LW, E). Although some of these participants described being
initially put off because of concerns that the tips would be difficult to implement, they
discussed how things became clearer upon engaging with the material: “At first, it was
quite difficult to follow. I looked at it and I thought, ooh that’s going to be really hard, but as
I worked my way through it and kept rereading it, it got easier and easier” (Participant 13, F,
SW, E).
Delivery of the tips via e-mail was seen as a facilitator to initial engagement for all
participants because it was not too intrusive to their daily life. The pdf format in which they
were presented, made them accessible on different devices, such as computer screens,
tablets, and phones. However, suggestions for modification of the delivery of the tips were
also made. They recommended receiving daily tips or reminders via text or within an app
to promote engagement over the longer term: “I’ve seen apps where these kind of things
pop up at regular intervals during the day. So that kind of tends to keep it more focused.”
(Participant 26, F, LW, E); and “Or just a text sort of thing” (Participant 35, F, LW, NE).
192
Engaged participants relied on the weekly reminders as well as the WFQ to keep them
motivated. They liked the frequency with which both were sent, although suggestions were
made to increase the reminders to twice a week, and decrease the time-consuming WFQ to
possibly every couple of weeks in a Survey Monkey format, to encourage a faster response
and delivery to fit into their schedule. This was also recommended by those not engaged.
The WFQ served as a reminder and a tool for reinforcement, helping participants to
maintain focus. They made participants think twice about what they were doing, as well as
prompting them to try to follow the tips: “The reporting back on Tuesday is absolutely
crucial” (Participant 6, M, LW, E). This was also the case for those participants who did not
engage with the tips for the duration of the intervention: “So in a small way, it would give
me a retrospective on the previous week going, “Yes, I didn’t do that, did I? Damn. Okay.”
(Participant 16, M, SW, NE). The majority of participants did not set themselves any goals,
either because they had previously set goals which they already followed, or because they
felt these didn’t work: “I find it’s a real negative actually, to set goals at all.” (Participant 2,
F, LW, E); or they just set themselves an initial goal to keep following the tips without
writing anything down: “I’ve not really set myself any goals, just to keep trying really”
(Participant 10, F, LW, E). Participants who engaged with the study also had difficulties
remembering to use the tips: “I’d forget for a few days and then think so I haven’t done it,
and I’d still… that disturbs me as a person, but it was fine to do it like that.” (Participant 23,
F, GW, E).
The importance of tailoring
The fact that the tips were personalised, or tailored, was considered by most of the
participants, to be one of the key strengths of the intervention. The personalisation
motivated engaged participants to follow the tips, even when the tips did not present them
with new information: “I think I knew these things about me, but it’s the first time I’ve seen
them providing a response and filling out a questionnaire” (Participant 26, F, LW, E). The
individual tailoring using the AEBQ was also seen as a novel tool: “Some of it is reinforcing
advice that I’ve heard before. But it’s nice how it’s all collated into one place and it’s
tailored for me so that I’ve got it all there to hand rather than having to wade through
pages and pages of things that aren’t even related to me.” (Participant 6, M, LW, E).
Some of the engaged participants felt that the tailoring of the tips was fundamental to their
success: “It’s incomparable, to be honest with you, because it’s so personal and it’s so spot
on. And what it has done is it’s asked me first of all what my particular areas of trouble are.
193
And that is absolutely vital, in my opinion” (Participant 28, F, LW, E). The information on
appetitive traits in the pdf allowed participants to recognise their personal appetitive
profile, which they had previously been unaware of: “I wouldn’t have registered the ‘food
responsiveness’ consciously without it being pointed out and explained” (Participant 41, F,
LW, E). And those engaged in the study felt that knowing that there were some traits they
didn’t have and therefore tips they did not need to follow was very comforting: “... so not
only to know the things that I needed to work on, but things that I... like I said before, that I
don't need to worry about so much, that I kind of think, ‘Actually, that’s really useful that I
don't have to... oh goodie, I already know when I’m full or not.’ So I found that really, kind
of... the fact that it was tailored was really helpful. That’s probably the biggest thing, to
know that not everybody would be like that, if that makes sense?” (Participant 46, F, LW, E).
However, not all participants had such a positive experience. Several participants did not
identify with a particular trait. They would have liked to have seen the results from their
AEBQ scores: “I don’t really have much of a problem with emotional eating, I don’t think to
myself, oh I’m getting all screwed up and I must eat something. It just doesn’t happen. It
will be interesting to see what I’d answered on that one, actually.” (Participant 6, M, LW, E);
or they simply did not agree with the AEBQ scores: “What put me off a little was the tips
sheet clearly had come to a conclusion based on my questionnaire, I would have liked to
have seen more… almost a review of the tip sheet – ‘Because you said this, this and this,
we’ve come to this conclusion’.” (Participant 16, M, SW, NE). Not relating to the traits
caused some participants not to engage with the study: “I didn't feel they were as
personalised as maybe I could have done with. I couldn't identify with some of the
descriptions, therefore it was much harder for me to...{follow}” (Participant 8, F, LW, NE);
and “Have a plan for another way to comfort yourself that doesn’t involve food’. Okay. We
can all have a plan, and I’ve had plans for years. And then when it got on to the tips, as I
say, they are not me, and they don’t feel like they will work.” (Participant 16, M, SW, NE).
Focus on drivers of behaviour
Participants acknowledged following many diets throughout their lifetime. However, those
who engaged with the tips compared the ATTI positively to other weight loss programmes
in terms of the focus being on what might be driving certain behaviours, as opposed to
simply asking participants to change these behaviours: “most things say eat less and don’t
eat this and do eat that and then, you know, more exercise, but they don’t spend quite so
much time thinking about how other people affect your weight loss, which I think this does,
194
it makes you think a little bit more” (Participant 2, F, LW, E). As a consequence, participants
felt the tips helped them to retrain their behaviour, which was felt to be beneficial in terms
of maintaining any changes made over the longer term: “Well, I think it is helping because it
is retraining me, and a lot of the other programs don't do that. They're all for weight loss,
and then as soon as you stop the programme, you put it back on again. Whereas what
you’ve got to do is you've got to re-train yourself to eat healthily and to not do the things
that you're doing wrong” (Participant 18, F, LW, E). One non-engaged participant
understood how the intervention was trying to motivate people: “I have found it
interesting, and I’m interested to see a different approach rather than everybody has to eat
cottage cheese and run four miles a day, or whatever. So I can see it’s trying to find how to
motivate people.” (Participant 35, F, LW, NE).
Too low intensity: A desire for more information
For those participants who did not engage with the intervention, the study did not
sufficiently motivate them to become engaged with the study: “So actually, it didn’t bother
me much. And I probably needed it to bother me more” (Participant 16, M, SW, NE). A few
non-engaged participants described feeling bored with having to try just another program
to lose weight, and one participant felt the tips were more for weight maintenance (rather
than weight loss): “I think some of your tips and stuff would actually be better just for
weight management, from my personal perspective.” (Participant 4, F, GW, NE). The tips
were perceived to be too low intensity and this made them easy to forget: “But quite often,
I have to say, I just completely forgot about them” (Participant 37, F, LW, NE). Non-engaged
participants felt the tips were not focused enough and suggested changes to improve the
interventions, such as having specific tips to follow every week: “And it was a bit… given it’s
a tip sheet, I think it was too focused on giving me the background rather than saying, ‘So
here’s your tips for this week, do these three things,’ for example.” (Participant 16, M, SW,
NE). One participant who did engage with the intervention and lost weight, also perceived
the tips to be too low intensity to have an impact on their own: “Yes, I think you’d have to
use it..{in combination with another weight loss method} I don’t think… because you’d have
to have a certain amount of knowledge about what to eat, or I think have to be following
some form of food management, if you like to call it. You couldn’t just eat the whole
spectrum, even if you were following these tips, it wouldn’t work. So you have to use it in
conjunction with some form of diet or other.” (Participant 2, F, LW, E).
195
All participants felt that the ‘emotional over-eating’ tip did not contain enough information
to help them follow this tip. Advice covered the few emotions listed in the AEBQ
questionnaire, but it did not extend to other emotions that participants felt triggered their
‘emotional over-eating’: “It says ‘sad or worried’. I mean, some people can eat when they
are happy, or me when I’m stressed, sort of thing, but you wouldn’t recognise it really unless
it was actually written down.” (Participant 13, F, SW, E). And participants criticised the
inclusion of a single tip related to ‘emotional over-eating’ and suggested having more tips
to help them with this trait: “They were quite sparse though {the ‘emotional over-eating’
tip}, compared to the other, and they didn't give any new information.” (Participant 35, F,
LW, NE).
The website links included in the tips were found to be very useful for some of the engaged
participants, though not everybody accessed these resources. There were, however,
suggestions to add additional links including sites providing information specifically on
appetitive traits, in order to provide a reliable and trustworthy source of relevant on-line
information. Engaged participants did feel that more information on healthy eating, such
as a list of further reading and website references might be useful for increasing the impact
of the intervention: “Yes, maybe a guideline to a weekly what you should eat during the day
sort of thing. Maybe a guideline on that would be helpful for people that weren’t following
a weight loss programme.” (Participant 47, F, LW, E). They also suggested including more
information on healthy snacks and healthy food options.
The role of personal contact
Most participants mentioned that direct personal contact could facilitate adherence with
the intervention, while counteracting feelings of loneliness and isolation of trying to lose
weight: “I think what's useful is finding that somebody other than you cares. I think it's very
lonely, being overweight.” (Participant 8, F, LW, NE). While some of the engaged
participants felt that the internet-based contact was sufficient to keep them following the
tips, others felt that more personal contact was necessary. It was suggested that
professional support and personal contact would have been particularly beneficial during
the initial profiling stage, at the beginning of the intervention: “I think, as I said earlier, a
beginning meeting and then going through this with the tips would be just amazing.”
(Participant 23, F, GW, E). Participants who did not engage with the intervention also felt
the need for more personal contact: “I think if I'd have had this conversation, even via
Skype, at the beginning, yeah, I would have been able to explain a little more; it would have
196
become a little more personal.” (Participant 8, F, LW, NE). Contact with a health
professional, would have been beneficial and thought to introduce an element of further
accountability: “Maybe it might have been nice to have had some telephone contact”
(Participant 30, F, LW, E); “Yes, I think so {would like more personal contact}. I do think that
you have to be accountable to someone else somewhere along the line” (Participant 2, F,
LW, E).
8.6.2.2 Consequences of the intervention
This second theme comprised three sub-themes: (i) Increased self-awareness; (ii) Behaviour
change; (iii) Physical consequences.
Increased self-awareness
Participants who engaged in the study discussed how the tips helped them to become more
self-aware of their behaviours and traits: “I wouldn’t have registered the food
responsiveness consciously without it being pointed out and explained.” (Participant 41, F,
LW, E). A heightened understanding of their traits and their behavioural consequences
increased their self-efficacy for making changes: “I think the most significant thing is just
the knowledge that I’m more responsive. So I just think twice about everything and I know
that it isn’t a genuine want for something.” (Participant 17, F, LW, E); and answering the
AEBQ helped them achieve this: “I've been trying to lose weight for a long time. And by
doing the survey, it came up with some things that maybe I wasn't aware of.” (Participant
18, F, LW, E). Engaged participants discussed how this increased awareness also helped
them to remain focused on their personal goals: “Well, it helped me to focus on what I was
trying to do. I didn’t always go through it step by step, like your slides, but certainly I kept
the gist of it to the top of my mind a lot of the time so that whenever I did think about
eating, I actually thought about these points, and that meant, or does mean, that when I
eat I am more aware of what I eat and why I’m eating it, and what I can expect from eating
it, and that helps.” (Participant 48, F, LW, E).
The tips provided the engaged participants with tools to feel more prepared and confident:
“I don’t know if confidence is the right word to use. I feel better armed, better equipped.
Yes, that’s it. I feel better equipped.” (Participant 28, F, LW, E). However, this self-
awareness was also seen in those who did not engage in the intervention: “I feel different in
myself for the slightly increased level of consciousness about me and my relationship to
food.” (Participant 16, M, SW, NE).
197
Behaviour change
Participants who engaged with the study described how the intervention had motivated
them to make changes to their dietary and physical activity behaviours. For example, they
described finding other things to do, instead of using food as a way to comfort themselves,
as was suggested for the ‘emotional over-eating’ tip: “I would try and do something nice for
myself, like either file my nails, paint my toes, toenails, have a nice shower with nice smelly
things, or something” (Participant 25, F, SW, E).
Specific tips emerged as being difficult to follow by those engaged in the study. The ‘satiety
responsiveness’ tip on not clearing the plate or eating smaller quantities was difficult for
some participants, because they felt that they were already careful to eat very little or
about the right amount. Participants also found it very difficult to not clear their plate
because of perceptions about wasting food, although some did achieve changes related to
this behaviour: “When I was a child, it was always that, ‘No, you empty your plate, there are
people in this world that are starving,’ and I was brought up with that mindset, anyway. So
that’s a very difficult one to get out of. But in saying that, I don’t waste food now, so I don’t
have that problem, because I only buy what I’m going to use and I make sure it’s all used.”
(Participant 28, F, LW, E).
Some participants who followed the intervention failed to identify with the tips within a
given trait, which made for difficulties in behaviour change. For example, there were
participants who did not feel they had a problem with seeing other people eat, or that
social eating situations either did not apply or were not a problem for them: “But we don’t
often have friends… we don’t really have people come around that often, so that’s not a
problem. So it’s only family mostly who actually come” (Participant 2, F, LW, E). One of the
most difficult tips to follow in the ‘slowness in eating’ category was the advice to put your
fork down between bites. Participants complained that this behaviour led to their food
getting cold: “I have tried, but I don’t like cold food. That’s the other thing. I like to eat it
before it gets cold” (Participant 9, F, SW, E).
Physical consequences
Those participants who lost weight were motivated by this weight loss: “I’m feeling more
energetic, I’m feeling good when people have noticed. For example, I’ve been told that my
thighs are smaller, that I’ve lost a bit of weight.” (Participant 41, F, LW, E). Participants who
engaged in the intervention also described changes they experienced as a consequence of
following the study, which extended beyond weight loss: “it’s actually had a massive impact
198
on my training as well, because now my nutrition is so much better because I’m not having
things that aren’t nutritious anymore, because I’m conscious of the fact that I don't want to
be eating a bit of chocolate or some crisps or everything.” (Participant 17, F, LW, E).
8.6.2.3 Barriers and facilitators to adherence
The last theme that emerged related to barriers and facilitators to adherence, which was
underlined by two sub-themes: (i) Routines and habits; and, (ii) social networks.
Routines and habits
Keeping to routines helped those participants who engaged with the intervention to follow
the tips: “When you are in a routine it’s a lot easier, but when you are out of routine or with
other people who don’t seem to care what they eat or whatever, then it gets a bit harder.”
(Participant 9, F, SW, E). In order to fit the tips into their routines, participants either made
changes to the tips to fit with their lifestyles or suggested how the tips could be improved
to increase their chances of success. For example, doing on-line food shopping helped
some participants avoid problems with ‘food responsiveness’ and temptation experienced
when walking along the aisles of a grocery store: “Because I do an on-line shop, so I don’t go
shopping, so that makes it slightly easier because then you are not seeing the food, and I
tend to buy the favourites each week, so I’m not looking down the sweetie aisle or biscuit
aisle, or whatever.” (Participant 25, F, SW, E). For some participants, wrapping a biscuit in
cling film, a tip recommended to train participants in countering their ‘food
responsiveness’, felt the concept of self-training was abstract and vague: “I was supposed
to wrap something up in Clingfilm and leave it by my desk. That’s possibly what it was that’s
put me off doing that, was I just kind of thought oh I can’t see that working.” (Participant 4,
F, GW, NE). However, engaged participants reported leaving out an entire unopened
packet, and attempting not to open it, or simply decided keeping problem foods out of
sight, essentially finding a different way to include the tips into their routines.
The occurrence of unscheduled or unplanned events disrupted routine and became a

barrier to success in engaged participants: “Things that are unplanned, I’ve not anticipated
really {prevent me from following the tips}.” (Participant 48, F, LW, E). Planning was difficult
particularly in the context of being able to find time to fit everything into the day: “Yes,
that’s the biggest thing for me is the time factor” (Participant 25, F, SW, E); this was also
reported by those who did not engage in the study. And even a routine such as work
sometimes kept participants from over-eating because it was a distraction, but for others
unplanned events at work was the cause of further stress: “You know like at work they just
199
bring in cakes and things, and you think [sighs] and that’s when I just succumbed and just
thought, ooh I’m starving, I’m going to have one of these” (Participant 13, F, SW, E).
Participants found that occasions that broke routines such as family visits, would throw off
people’s plans: “my wife loves baking, and she does tend to bake when the family visits, and
things that don’t normally appear in our menu suddenly appear and they are quite
attractive and very tempting” (Participant 6, M, LW, E). Participants that did not engage
with the intervention mentioned that going away on holiday or going out on weekends was
one of the main reasons that prevented them from following the tips: “I think, to be honest,
a lot of it is just I’ve had a lot of weekends away and I’m not in control of what I can eat and
that then, you know, it just makes life so difficult when you are just trying to be controlled,
really.” (Participant 4, F, GW, NE). After these events, resuming the tips was difficult.
Participants wanted to incorporate the tips into their habitual behaviours, so that they
could keep reinforcing the information: “Yes, and it makes me maintain it as well, because I
don’t think it’s got into a habit where I’d be able to do it without thinking about it, so I need
to keep on top of it and keep bringing it to the front of my mind to make sure I’m still doing
it, like a check.” (Participant 10, F, LW, E). Overall participants who followed the tips
seemed to be able to incorporate them into their routines. However participants also
reported that existing ‘bad’ habits sometimes acted as a barrier to adherence: “I do have a
lot of problems with habits. Like if I go to the garage, I just buy a chocolate or something
like that. I think I’m a bit lazy at times.” (Participant 10, F, LW, E). Similarly: “Just, again, not
thinking, just going straight for it.” (Participant 46, F, LW, E). And this was the case for non-
engaged participants also: “My habit of eating fast” (Participant 16, M, SW, NE).
Social networks
Engaged participants mentioned a need to have support from their family and friends to
help them make changes: “{I’d say to my daughter} Remind me to eat slower.” (Participant
25, F, SW, E). Participants suggested that advice about building a support network to
encourage adherence could become part of the tips themselves: “I got my husband
involved, I got a few friends involved … and they provided a support network, encouraging
me to wait or to eat healthy instead” (Participant 41, F, LW, E). Other people’s examples of
successful behaviour change were also reported as important facilitators, such as previous
weight loss from spouses or friends: “my husband has lost weight recently and changed a
lot of things about his diet, so he will be more in tune with my goals” (Participant 9, F, SW,
E); or mimicking of other people’s good behaviour: “My husband, he eats really slow, so I
200
was trying to match his pace.” (Participant 13, F, SW, E). Participants who did not engaged
in the intervention made suggestions of different ways they could be supported by others
to help them make changes: “Having someone else clear away is a big one, but you didn't
really touch on that…. And the other one I would have is have somebody in the kitchen while
you're preparing the food, because I don't pick anywhere near as much when other people
are there.” (Participant 8, F, LW, NE).
One of the main barriers to following the tips were related to personal or family health
issues. An example from one participant was having disabled children, who required time
and attention, creating a stressful environment in which to follow a personal weight loss
intervention. Other family issues also created stressful situations, such as having teenagers
around the house who ate different foods or having other family members who were not
on a diet were also a strain: “I think a difficult one tends to be the buying and having
healthy foods in the home, because there’s not always going to be everyone in the home
wanting to keep an eye or be on a diet or watch their weight.” (Participant 41, F, LW, E).
People who are more susceptible to eating in the presence of others might experience this
barrier more often.
Access to a newsletter, and support through on-line forums were proposed in order to
create a weight management community and provide a space to discuss personal issues by
both engaged and non-engaged participants: “I suppose it would always be interesting to
access a newsletter just so that you know where you fit into the bigger picture, I suppose.”
(Participant 48, F, LW, E); “you could possibly look at a forums and things like that.”
(Participant 8, F, LW, NE).
8.7 Discussion
This study built on findings from Study 4, Chapter 7, with the aim of exploring participants’
experiences of taking part in a brief Appetitive Trait Tailored Intervention (ATTI). Three
themes and their sub-themes emerged from the interviews. First, the experience of the
intervention revealed five sub-themes: Engaging with the tips and materials; the
importance of tailoring; focus on drivers of behaviour change; too low intensity: a desire for
more information; and, the role of personal contact. The second theme, consequences of
the intervention included three sub-themes: Increased self-awareness; behaviour change;
and physical consequences. Finally, two sub-themes emerged from the third theme
201
barriers and facilitators to adherence, which related to routines and habits and social
networks. Overall 21 interviews were obtained, 16 from participants who engaged with the
intervention, and five interviews from participants who engaged only very briefly with the
intervention. The findings from this study yield useful recommendations for refining the
intervention before moving forward with rigorous implementation of the study (step six
steps of Wight et al.’s 6SQuID steps) (Wight et al., 2015).
8.7.1 Experience of the intervention
Overall, differences between engaged and non-engaged participants’ experiences of

following the ATTI and using the tips were seen. Engaged participants found the
intervention to be clear and easy to understand. These results are in line with previous
studies that have shown that weight loss intervention material that is simple and easy to
deliver, can be a beneficial way to obtain significant weight loss (Lally et al., 2008), although
the ATTI differs in that the simple information was delivered via the internet. Engaged
participants also relied on the WFQ as reminders, they saw them as motivational tools that
helped as reinforcement, helping them to maintain focus, although suggestions were made
to change the format and frequency of their delivery. However, they still had difficulties
remembering to follow the tips. Recommendations to receive daily tips via text or within
an app were proposed to help with sustaining engagement, and serve as more frequent
reminders. A systematic review of 85 internet-based health interventions showed that
those providing supplementary delivery modes such as text-messaging were more effective
at promoting health behaviour change (Webb et al., 2010). Webb et al., found that overall,
internet-based weight management programmes had a small effect on weight-related
behaviour change, providing a suitable medium for delivery of an intervention, however
text messaging enhanced the effectiveness of such interventions. This could therefore be
beneficial to include in taking the ATTI forward.
Engaged participants also found the personalization and tailored aspects of the tips
motivating and fundamental to their success, and even non-engaged participants found the
ATTI provided information about their own traits that they didn’t know. A big part of
weight loss interventions is related to motivation (Dalle Grave et al., 2013; Metzgar,
Preston, Miller, & Nickols-Richardson, 2014; Stead et al., 2015). For those engaged
participants who did not have unfavourable scores for all the traits, it was comforting to
know that they did not have to follow all of the tips. Tailoring of diets to specific traits and
personal characteristics (which could include, age, gender, or even factors such as weight
202
loss expectations, or lifestyles) is a known predictor of success at weight loss and weight loss
maintenance (Teixeira et al., 2005) (as discussed in Chapter 1). Personal tailoring of weight
management information, alongside the provision of feedback regarding genetic risk of
obesity was viewed as beneficial among ‘higher-risk’ and ‘average-risk’ students and among
overweight and obese adults (Meisel & Wardle, 2014a). Results from other qualitative
studies also suggest that adapting weight loss interventions to participants’ differing
characteristics is likely to improve outcomes (Stead et al., 2015). To further assess the
value of tailoring to specific traits, future studies could explore if greater success is achieved
with the ATTI compared with a non-tailored intervention providing similar information for
weight management.
This study shows that participants felt that the ATTI compared favourably to other weight
loss programs, and in particular it allowed them to consider their own behaviours and start
to retrain them. This suggests that the development of the tips based on CBT adapted from
‘Shape-Up’ (Study 4, Section 7.3.3) was partly successful and in line with the aim to provide
skills to control over-eating tendencies (Wardle & Johnson, 2015; Wardle et al., 2013).
However, some participants felt that they needed the intervention to be more intensive.
These participants seemed to get bored with having to try just another program to lose
weight, and one participant felt the tips were more for weight maintenance purposes. In a
systematic review of 23 studies, self-help interventions (which are self-directed and do not
require professional input to deliver) produced significant, albeit modest, weight losses at 6
months when compared with minimal interventions (Hartmann-Boyce et al., 2015).
However, this level of input may not be sufficient for everyone and so the ATTI could
potentially be used alongside other programs, as was suggested by one participant. More
intense versions of the ATTI could also be tested, for example providing participants with
more support, which also fits with the desire for increased personal contact from a health
professional.
The recommendation to provide initial contact after the AEBQ profiling suggests this should
be taken into consideration when continuing with the ATTI in future evaluations. Personal
contact with a health professional is known to increase weight loss success and previous
qualitative studies have shown that having some form of nutritional education for weight
maintenance, was necessary to making healthier choices (Jebb et al., 2011; Metzgar et al.,
2014). Personal contact also provides a form of accountability which is known to increase
motivation (Stead et al., 2015). Both engaged and non-engaged participants in this study
203
mentioned that some form of personal contact made them feel more accountable, thereby
increasing their motivation to apply the tips.
Other ways of providing more support would be to include web links to more information,
for example recipes and healthy food choices, and a webpage with more explicit
information on appetitive traits, something that participants suggested. These results are
in line with replies to the feasibility study carried in Study 4, Section 7.3.2, where 46.5% of
participants reported that they would be interested in receiving information about ‘healthy
food options’. However, healthy food options were not included in the tips, as they were
thought to be related mostly to ‘enjoyment of food’. Given that the majority of participants
have high scores of ‘enjoyment of food’ (mean 4.00±0.74), no tips were developed for this
trait, based on the steps for quality intervention development (Wight et al., 2015), there
seemed to provide a limited scope for change. However, consideration to participants’
desire for this input should be taken, when taking the ATTI into future research.
In some cases, participants reported that they found specific tips difficult. In particular, the
‘emotional over-eating’ tip was found to be the least informative and was least endorsed by
the participants. Participants mentioned that different types of emotions set them off.
However, the ATTI only includes feeling ‘upset’, ‘worried’, ‘anxious’, ‘annoyed’ or ‘angry’,
and participants mentioned including other emotions as part of the tips such as happy, or
even being stressed or tired. These findings were in line with results from the quantitative
analysis in Study 4, Chapter 7. Given that this single tip was based on ‘cognitive
restructuring’ techniques when it was developed (Study 4, Section 7.3.3, Chapter 7), it may
require the inclusion of additional techniques to influence challenging thought processes
when refining the intervention and moving forward in future implementation of the
intervention (Dalle Grave et al., 2013; Wardle et al., 2013; Wight et al., 2015). Stress
management was the behaviour change technique which was associated with the greatest
change in behaviour in the previously mentioned systematic review of internet-based
interventions (Webb et al., 2010). Given only a few studies used it, its use should be
treated with caution, but it was seen to be more effective than emotional control training.
The possible use of stress management strategies to help with ‘emotional over-eating’
should be considered in future work with the ATTI.
204
8.7.2 Consequences of the intervention
Engaged participants showed an increased self-awareness of their behaviour and their

traits, which led to an increase in their self-efficacy for making changes. Self-efficacy
relates to participants’ beliefs about being able to make changes to their behaviour and is a
known predictor of weight loss success (Dalle Grave et al., 2013; Stubbs et al., 2011).
Engaged participants in this study described how the ATTI helped them to feel motivated
about making changes to their behaviours. It provides useful information that allowed
participants to feel more confident in their capacity to change their behaviours. Even when
some of the tips were difficult to follow due to persistent ingrained behaviour such as
avoiding clearing their plate (‘satiety responsiveness’ tip), behaviour change was reported
to be achieved by some. Literature shows that for changes to be made, participants need
to feel confident to initiate new behaviours (Schwarzer, 2008). This study showed that
changes in behaviour had an effect on weight, as well as other positive physical
consequences.
In line with Study 4, when asked about the goal setting sheet in the ATTI, the majority of
participants did not set any goals for themselves. They had either previously set goals prior
to the intervention, or had set themselves an initial goal to follow the tips throughout the
eight weeks. Self-management strategies, such as goal setting and self-monitoring, have
been seen to be among the most commonly recommended strategies in self-help studies
(Hartmann-Boyce et al., 2015), however, they do not appear to have any effect on
behaviour change as part of internet-based health interventions (Webb et al., 2010).
Testing to see if goal-setting is effective for use in the ATTI might need the use of a control
group with and without advice to set goals, to assess differences.
8.7.3 Barriers and facilitators to adherence
Results from this study revealed that keeping to routines was an important facilitator of
adherence to the intervention. These results are in line with previous qualitative studies
which reveal that when changes to diets and physical activity are incorporated into
participants’ daily routines, they are more likely to be maintained (Lally, Wardle, & Gardner,
2011; Stead et al., 2015). When the tips did not fit into their daily lives, engaged
participants made adaptations to incorporate them. Suggestions for on-line shopping or
different ways to carry out ‘resistance training’ instead of using cling film to wrap a problem
foods were made, which could in future be developed. The development of a ‘bank’ of
different tips could be developed for future testing of the ATTI, this could provide different
205
tip options for all the traits. For example, the development of different ‘food
responsiveness’ tip options on the use of cling film as ‘resistance training’, or different
activity options when going out with friends to avoid over-eating, which were mentioned as
tips that engaged participants did not relate to. Other tip suggestions could include
different ways to reduce portion sizes or eating more slowly, which participants found
particularly difficult to follow.
Unplanned events and disrupted routines were barriers to following the tips in engaged
participants and in non-engaged participants’ holidays and weekends away were particular
problems. Documented experiences of ten participants who followed the Top Ten Tips in
an weight loss intervention for an eight week period, revealed that preparatory planning
was needed to avoid past behaviours after a cognitive awareness of previous ingrained
habits (Lally et al., 2008; Lally & Gardner, 2011). Participants in this study suggested that
they would be interested in making the tips into habitual behaviours. Support for planning
to form positive habits could then be included in this intervention in future, given that
planning is an important part of initiating habit formation (Lally et al., 2011). Participants
believed that changing their habits was difficult, in light of patterns that were very
established. They talked about ingrained habits and saw these as barriers to change,
although participants mentioned that the tips helped them to change their self-awareness.
On the other hand, finding ways to integrate the tips into existing routines was identified as
a useful way to support the use of the tips. This again fits with habit-formation theory
where repetition in a consistent context could help behaviours to become automatic
(Beeken et al., 2012).
All participants reported social networks as both facilitators and barriers to adherence to
the tips, in line with previous research (Dalle Grave et al., 2013). Significant evidence exists
for social support in the form of friends, family members or co-workers, as an essential
element of weight loss and weight management programmes (Alm et al., 2008; Metzgar et
al., 2014; Stead et al., 2015; Thomas et al., 2014). Whilst encouragement and support from
family was a facilitator of behaviour change for some, family was also a barrier to
adherence. Results from the interviews support the findings from Study 4, where
participants mentioned lack of support from significant others as a common barrier to
following the tips (Section 7.7.4, Chapter 7). The ATTI could provide suggestions of
different ways that they could enlist support from family and friends, as some participants
suggested that this should be included in the tips. Both engaged and non-engaged
206
participants suggested including access to a newsletter, and support through on-line

forums in order to create a weight management community and provide a space to discuss
personal issues. This has been shown to moderately support behaviour change (Webb et
al., 2010). Again, these suggestions should be considered in any future development of the
ATTI.
8.7.4 Limitations
This study has a number of limitations. Firstly, attempts to contact participants who did not
finish the ATTI or who failed to provide their final weight were largely unsuccessful, so all of
those interviewed had some level of engagement with the ATTI. Better insight might have
been gained into barriers to adherence from those who did not complete the intervention.
Significant differences were seen between those interviewed and the number of replies to
WFQ. Those interviewed were significantly more likely to have replied to ‘5 to 8 WFQ’
14/21 (66.7%), than those not interviewed (‘5 to 8 WFQ’ 2/21 [33.3%]). These results
possibly show differences in engagement with the intervention. Also, perhaps some people
didn’t feel the need to fill in the WFQ every week, but still engaged with the intervention.
Information was obtained from five participants who were considered not to have engaged
with the ATTI. Four interviewed participants reported following the tips for approximately
two weeks overall. Two of those four participants were grouped into those non-engaged,
however they still liked the ATTI and related to the tips, and lost weight, probably due to
following other forms of weight management. Another two participants did not
particularly like the intervention, one of them maintained their weight and the other lost
weight (but followed the tips for a slightly longer five-week period). Only one participant
stopped following the tips because she felt she was not ready to follow them and she
gained weight. Results from more participants who did not follow the tips, weren’t
engaged or did not finish the intervention, could have enriched the results obtained from
the interviews.
A further limitation surrounds information on weight and weight management, which can
be extremely sensitive in nature, potentially leading to social desirability bias occurring in
the interviews (Cameron & Evers, 1990; Johnson et al., 2012). However, all participants
invited to take part in the interview were told that even if they hadn’t followed the
intervention, or the tips, their thoughts would be welcome, even if the input was negative.
Thirdly, although no personal support was given to the participants throughout the ATTI, a
207
few participants mentioned that they felt supported by an intervention. This could have
increased the liking towards the ATTI.
The majority of participants interviewed were women (90.5%) which is also a limitation of
the study. Previous studies have suggested that men often find it easier to lose weight than
women (Jackson, Beeken, & Wardle, 2014; Stubbs et al., 2011), and it is possible that
participants’ experience of the intervention would differ by gender. However, as previously
stated in Study 4, women are more willing to participate in weight loss interventions (Jebb
et al., 2011), and have a higher prevalence of dieting than men (Hartmann-Boyce et al.,
2015; Provencher et al., 2004). Additionally, participants were mostly white, well-educated
and relatively wealthy. A systematic review of self-help programmes suggested that
interventions carried out without the support of a professional may be more effective in
socio-economically advantaged groups (Hartmann-Boyce et al., 2015).
A further limitation includes the use of ‘Big Panel’ members who register to a charity with a
view to participating in weight management studies is a potential cause for selection bias.
As mentioned in Study 4, Section 7.8.4, little is known about whether e-mail contact with
participants still works or if their contact details are correct and many panel members have
been invited to lots of studies before, which could increase or decrease their motivation to
participate in the study. Even though participants appeared to be enthusiastic about the
tips and liked the information in general, difficulties expressed by some participants may
have been exacerbated by previous weight loss failures, and hence, discouraged some
participants from following the tips. The findings of this study therefore cannot be
generalized to other populations and future research should explore the efficacy and
acceptability of personalized appetitive trait based weight management interventions in
more diverse samples.
As per the COREQ 32 item checklist, the only item not covered in this study was number 23
(Transcripts returned) (Appendix 8.3), as the transcripts were not returned to participants
for comments and/or corrections after the interviews, not giving the chance to participants
to clarify any doubts that could have arisen from transcription. Also, the fact that I ran the
eight-week intervention, communicating directly with participants and also personally
conducted all 21 interviews myself could have resulted in researcher bias. For example,
eliciting narrowly defined answers to questions during interviews (e.g. ‘yes’), to questions
such as “did you find the delivery of the tips via e-mail to be adequate?”. However, this
was partially mitigated by the involvement of other researchers, with agreement on inter-
208
rater coding and discussion of the underlying themes. Finally, these interviews took place
shortly after the intervention so provide participants’ views only immediately post-
intervention.
8.7.5 Conclusions
Results from the present study appear to lend support to the utility of a brief, tailored
weight loss intervention of appetitive traits. Overall, interviewees were positive about their
experiences of the ATTI. Interviews were obtained from 16 participants who engaged with
the intervention, although five interviews were carried out in participants who engaged
only briefly with the intervention. Three themes emerged from the interviews which
related to participants’ experiences of the intervention, consequences of the intervention,
and barriers and facilitators to adherence. In general, participants found that they could
engage with the tips because they were simple, clear and their delivery accessible. They
found the tailored aspect of the intervention to be novel and motivational, comparing
favourably with other weight loss problems as it focused on drivers of behaviour. However,
mostly non-engaged participants found the intervention to be too low intensity, and
suggestions were made for the inclusion of more tips, particularly ‘emotional over-eating’
tips, as well as a variety of different tips to suit different behaviours. Participants suggested
that some form of personal contact at the beginning of the intervention could facilitate
engagement with the tips and accountability, and improve specific aspects of the tips to
improve tailoring. Overall, the ATTI appeared to increase self-awareness, lead to changes in
behaviour and have other positive physical consequences such as feeling more energetic
and having impact on physical training. Finally, lack of routines and ‘bad’ habits, were seen
as barriers to adherence to the tips. Engaged participants found support from social
networks facilitated adherence but living with family members who consumed unhealthy
foods in the house also presented barriers.
These results are promising for the initial testing of this novel intervention, which along
with the results from Study 4, Chapter 7, covers the first five Steps of the 6SQuID. It
suggests that tailoring an intervention to an individual’s appetitive traits is acceptable and
could help support weight loss attempts. However, it was difficult to obtain qualitative
data from ‘non-completers’ so more needs to be done to identify reasons for non-
completion given the high level of drop out in Study 4. Future studies should also seek to
refine the intervention based on these results and to explore its effectiveness for weight
loss.
209
Chapter 9. General Discussion
9.1 Introduction
Given the rising prevalence of obesity on a worldwide scale, there is a need for novel and
effective weight management interventions. The significant inter-participant variability in
many behavioural intervention responses has led to a growing interest in the idea of
personalised interventions tailored to individuals’ needs. A potential target for such
interventions could be to use a person’s unique appetitive trait profile to tailor weight
management advice for overweight or obese individuals. However, first there is a need to
further understand the relationships between appetitive traits and weight, and to observe
if the associations seen in childhood still hold into adulthood. The Child Eating Behaviour
Questionnaire is a robust measure of appetitive traits that have consistently shown an
association with weight across infancy and childhood (Ashcroft et al., 2008; van Jaarsveld et
al., 2011). However, studies exploring these appetitive traits across the life course have
been limited by the lack of a comparable self-report measure of these traits for adults.
Measurement of these appetitive traits in adulthood would contribute to our
understanding of how appetite influences weight gain at older ages, as well as providing
potential targets for interventions. Psychometric measures of appetitive traits have not
previously been used to tailor weight management interventions. Therefore, the aim of the
research presented in this thesis was to address gaps in the existing literature by
investigating the relationship between appetitive traits and BMI in adulthood, and to
explore the potential for a weight management intervention to be tailored to an
individual’s appetitive profile.
9.2 Summary of findings and contribution to the literature
The work presented in this thesis was based on four research questions, addressed in five
empirical studies. Study 1 was a systematic review of the literature on psychometric
measures of appetite and appetitive traits. Studies 2 and 3 collected data from an on-line
research panel in order to develop a new measure of appetite, validate it, and study its
relationship with weight. The fourth study developed and tested a brief intervention based
on tips tailored to an individual’s appetitive profile, and aimed to assess facilitators and
210
barriers to their use. The fifth and final study explored participants’ experiences of using
the tips through qualitative research. This chapter discusses the main findings of the thesis
based on the research questions proposed and the contributions they make to the
literature. It reflects on limitations, as well as directions for future research and
implications for practice and policy.
9.2.1 What psychometric measures of appetitive traits currently exist?
Study 1 systematically reviewed existing psychometric measures of appetitive traits, to

assess the need for a novel measure of appetite in adulthood. To the best of my
knowledge, this is the first systematic review to explore psychometric measures of
appetite. Results from this systematic review revealed 32 psychometric measures of
appetite for children, adolescents and adults.
After scoring the reviewed questionnaires to determine their validity and reliability, and
whether associations with weight were reported, 17 measures obtained the highest score
for psychometric robustness. The CEBQ was the only one of these robust questionnaires to
measure an array of aspects of appetite to use behavioural validation studies (Carnell &
Wardle, 2007). Following a further citation search using Google Scholar, three out of the 17
measures emerged as the most widely used: the TFEQ (Stunkard & Messick, 1985), and the
DEBQ (van Strein et al., 1986), mostly for use in adults, and the CEBQ for use in children.
The most common traits measured in adult questionnaires that have been associated with
weight, relate to aspects of ‘restraint’ and disinhibited eating. Many studies using these
measurements were undertaken in populations with weight management issues and eating
disorders. From studies using the CEBQ, most of the evidence points to relationships
between ‘satiety responsiveness’ and ‘food responsiveness’ and weight. No equivalent
measure of ‘satiety responsiveness’ was found in adults. Although ‘food responsiveness’
measured using the CEBQ is similar to ‘external eating’ measured by the DEBQ, the former
is unrelated to dieting and ‘restraint’, and has not been measured in adults. Several
measures assess ‘emotional eating’, however the CEBQ also measures ‘emotional under-
eating’ which has also not previously been measured in adults. The CEBQ also measures
‘enjoyment of food’, ‘slowness in eating’, ‘desire to drink’ and ‘food fussiness’ not currently
measured in adults. Study 1 demonstrated that no currently available questionnaire could
measure similar appetitive traits across the life course, from infancy to adulthood.
However there was potential; for this to be achieved using the BEBQ in infants, the CEBQ in
children and a self-report version of the CEBQ adapted for adults.
211
9.2.2 Can the parent report Ćhild Eating Behaviour Questionnaire´ (CEBQ)
be adapted into a valid and reliable measure of appetitive traits in
adults?
Study 2, Chapter 5, described the development of the Ádult Eating Behaviour

Questionnaire´ (AEBQ). The self-report AEBQ was adapted from the parent-report CEBQ
with the addition of a ‘hunger’ scale, and the removal of the ‘desire to drink’ scale. Based
on piloting, the response options were changed from: ‘never’, ‘rarely’, ‘sometimes’, ‘often’
and ‘always’, to ‘strongly disagree’, ‘disagree’, ‘neither agree nor disagree’, ‘agree’ and
‘strongly agree’. Exploratory factor analysis in a large sample of adults, revealed a final 35
item questionnaire measuring three ‘food approach’ traits (‘hunger and food
responsiveness’ which loaded onto the same factor, ‘emotional over-eating’ and
‘enjoyment of food’) and four ‘food avoidant’ traits (‘satiety responsiveness’, ‘emotional
under-eating’, ‘food fussiness’ and ‘slowness in eating’). Cronbach alpha values were
greater than 0.7 for the seven traits, providing evidence for the internal reliability of the
AEBQ.
Study 3, Chapter 6, confirmed the factor structure of the AEBQ in a second sample of
adults. Confirmatory factor analysis revealed the same structure as Study 2, except a better
model fit was found when ‘hunger’ and ‘satiety responsiveness’ were separated into two
factors, revealing eight final sub-scales. Reliability measurements showed that the AEBQ
was internally reliable and results from a test-retest two weeks apart revealed the AEBQ to
be reliable over time.
The findings from Studies 2 and 3 (described in Chapters 5 and 6 respectively) show that
the AEBQ is a reliable and valid questionnaire that measures eight distinct appetitive traits
in adults.
9.2.3 How do appetitive traits relate to BMI in adults?
After confirming the factor structure of the AEBQ, Study 3 also explored the relationship
between appetitive traits captured by this questionnaire and BMI in adulthood. Results
revealed similar associations between appetitive traits and BMI in adults to those
previously reported in children. Positive associations were seen between BMI and the
‘food approach’ traits; ‘food responsiveness’, ‘emotional over-eating’ and ‘enjoyment of
food’. On the other hand, negative associations were seen between BMI and ‘food
avoidance’ traits; ‘satiety responsiveness’, ‘emotional under-eating’ and ‘slowness in
eating’. Results were consistent with studies in infants (van Jaarsveld et al., 2011) and
212
children (Santos et al., 2011; Sleddens et al., 2008; Spence et al., 2011; Viana et al., 2008;
Webber et al., 2009), but associations were more modest in our sample of adults. This may
be indicative of appetitive traits exerting a differential influence on weight across the life
course. Furthermore, adults may actively restrict their energy intake in an attempt to
control their weight, which could supress the impact of certain traits on BMI, whereas
children typically do not exert such control over their eating.
No associations were found between BMI and the newly added ‘hunger’ scale. In
retrospect, this may be because the items to measure ‘hunger’ within the AEBQ, represent
more of a ‘state’ rather than a ‘trait’ (Blundell et al., 2009; Harrold et al., 2012), and
therefore may be more affected by temporal factors such as the time of the last meal
(Blundell et al., 2009). ‘Food fussiness’ was also unrelated to BMI in this adult sample,
whereas CEBQ measured ‘food fussiness’ has been linked with lower weight in some child
studies (ref). It is possible that ‘food fussiness’ in adults is directed towards a much smaller
number of foods, while greater variation exists in relation to children’s ‘food fussiness’
(Croker et al., 2011; Spence et al., 2011; Webber et al., 2009). Picky eating in adults is also
associated with forms of unhealthy eating (Kauer, Pelchat, Rozin, & Zickgraf, 2015).
However, relationships between ‘food fussiness’ and BMI in children have not always been
consistent (Santos et al., 2011; Svensson et al., 2011). Future research using the AEBQ will
determine if the ‘hunger’ and ‘food fussiness’ scales should be retained, and their
importance in relation to weight as appetitive traits in adults.
9.2.4 Can a weight management intervention tailored to an individual´s

appetitive traits be developed that is acceptable and potentially useful?
Study 4, Chapter 7, involved the development and initial testing of an intervention tailored
to an individual’s appetitive profile, the ‘Appetitive Trait Tailored Intervention’ (ATTI). The
first five steps of Wight et al.’s 6SQuID were followed (Wight et al., 2015), and the study fell
within the ‘development’ phase of the Medical Research Council (MRC) framework for
developing and evaluating complex interventions (Craig et al., 2008).
Tips were developed for ‘food responsiveness’, ‘satiety responsiveness’, ‘emotional over-
eating’ and ‘slowness in eating’ traits, as these provided the widest scope for change. No
tips were developed for ‘hunger’ or ‘food fussiness’ as no relationships were observed
between these traits and BMI in Study 3. Tips were also not developed targeting
‘enjoyment of food’ as the majority of participants scored highly on this trait and it is
problematic to tell someone not to enjoy food. Similarly no tips were developed for
213
‘emotional under-eating’ which does not appear to lead to weight gain (Wardle, Guthrie, et
al., 2001). The tips were developed to help bring about change based on CBT techniques
such as ‘stimulus control’ and ‘response substitution’, using recommendations taken from
‘Shape-Up’, a behavioural healthy lifestyle program for weight loss (Wardle et al., 2013).
Other behaviour change techniques such as goal setting were incorporated into the tips to
support change (Michie, Atkins, et al., 2014).
In order to establish if participants would like to receive such an intervention and how they
would prefer it to be delivered, feasibility questions were included in Study 3. Participants
were asked if they would be interested in participating in an on-line intervention involving
tailored feedback on their appetitive traits based on their AEBQ results. More than half of
the participants (58.4%) replied that they would be interested in receiving this feedback
and tips on how to manage them accordingly. The majority of participants also reported
that they would prefer to receive this information via e-mail (63.4%), rather than in-person
(2.1%), and would also like to be reminded on a weekly basis to continue following the
recommendations (63.4%).
Both the theoretical background and the feasibility study results informed the development
and testing of the intervention on a small scale, within Study 4. Participants completed the
AEBQ and received a set of personalised tips based on their adverse appetitive trait profile
along with weekly reminder e-mails. Out of 53 participants at baseline, a total of 32
participants completed the study and provided a final weight. A mean weight loss of 1.2 kg
was reported over the eight-week intervention. Participants reported that they liked and
used the majority of the tips except for the ‘emotional over-eating’ tip and two of the tips
for managing ‘food responsiveness’. They also reported a number of barriers similar to
those described in previous weight management interventions, such as: ‘force of habit’, or
‘forgetting to carry them out’, as well as external situations making it difficult to follow
specific tips.
In order to gain further insight into the acceptability of the ATTI, semi-structured qualitative
interviews were conducted with 21 intervention participants, five of whom had not
engaged with the study. These findings are reported in Study 5, Chapter 8. Thematic
analysis of semi-structured qualitative interviews revealed three themes with their own
sub-themes. First, the experience of the intervention theme with five sub-themes:
Engaging with the tips and materials; the importance of tailoring; focus on drivers of
behaviour change; too low intensity: a desire for more information; and, the role of
214
personal contact. The second theme, consequences of the intervention with three sub-
themes: Increased self-awareness; behaviour change; and physical consequences. The
third and last theme, barriers and facilitators to adherence with two sub-themes: Routines
and habits; and, social networks. Overall, the results indicated that participants found that
they could engage with the tips because they were simple, clear and their delivery was
accessible. Tailoring of the tips based on appetitive traits was seen as novel and
motivational. The non-engaged participants found the intervention to be too low intensity,
and in particular the ‘emotional over-eating’ tips were seen as not including enough
information. Suggestions were made to include a greater variety of different tips for each
trait. Participants also suggested including some form of personal contact particularly at
the beginning of the intervention, preferably with a health professional. Some of the
consequences of following the ATTI, were that participants reported changes in their
behaviour and physical consequences of following the tips (such as finding other things to
do instead of eating, no longer eating all the food on their plate, feeling more energetic,
and improved physical training), and this was due partly to an increase in self-awareness.
The main barriers to adherence with the tips were a lack of routine and ‘bad’ habits. Social
support was both a facilitator and barrier to adherence, where for example families could
be a help (by providing encouragement) or a hindrance (such as having family members in
the house who were ‘not on a diet’).
Together, Studies 4 and 5 provide evidence supporting the use of AEBQ measured
appetitive traits to inform a personalised weight management intervention targeting the
aspects of an individual’s appetitive profile that put them at greatest risk of weight gain.
However, some aspects of the intervention were less effective and/or engaging and there is
a need to refine the intervention based on the results obtained from both of these studies.
Given the high level of drop out in Study 4, more work is also required to identify reasons
for non-completion. Future work should seek to refine the tips, further explore reasons for
drop out, and ultimately test the effectiveness of the ATTI within a powered study.
9.3 Limitations
Limitations corresponding to each study are outlined in the relevant chapters. However,
some limitations are common to several of the studies and are further discussed below.
215
9.3.1 Self-reported data
Self-reported measures may result in socially desirable responses. Social desirability is a

well-documented issue in questionnaire studies (Allison & Baskin, 2009; Carnell & Wardle,
2008a). These issues might have been particularly relevant to Studies 2, 3, 4 and 5, given
the nature of the information collected, regarding appetite and sensitive issues such as
weight and weight management (Cameron & Evers, 1990). Given the societal preference
for thinness and the prevalence of weight stigma, individuals who are heavier might be
particularly influenced by social desirability bias (Polivy & Herman, 2004; Wee et al., 2013).
These issues could have resulted in under-estimated levels of certain appetitive traits such
as ‘food responsiveness’ or ‘emotional over-eating’ or over-estimates of ‘satiety
responsiveness’. However suggestions have also been made that web-based data
collection can reduce social desirability pressures (Marlow & Wardle, 2014), by reducing
contact with health professionals.
Other issues common to self-report questionnaires are the under-reporting of weight and
over-reporting of height measurements, which in turn leads to BMI under-estimates
(under-reporting bias) (Gorber et al., 2007). This introduces systematic error into self-
reported measures (Rowland, 1990). This could have been an issue for each study included
in the systematic review in Study 1, where convenience samples and mostly self-reported
weights and heights are obtained from participants. This is also an issue in Studies 2, 3, and
4 where all participants self-reported their weights and height. Therefore, it is very likely
that under-estimation of BMI was a common issue throughout this thesis. Self-reported
measures of height and weight are also known to be less accurate in older age groups
(Kuczmarski et al., 2001). Therefore, self-reported height and weight measurements in
those participants above 60 years of age in Studies 2, 3, and 4, might be further under-
estimated, with nearly a quarter of participants who completed the appetitive trait
intervention (Study 4) over the age of 60. Possible implications of such under-reporting
could lead to lack of associations between appetitive traits and weight in older age groups,
compared to younger age groups.
9.3.2 Cross-sectional data
The nature of cross-sectional data in general, does not allow for inferences on causation.
For example, in the case of Study 3, associations were seen between appetitive traits and
BMI in the sample. Results from studies in children have suggested that appetitive traits
are associated with BMI (Carnell & Wardle, 2008b; Sleddens et al., 2008; Viana et al., 2008),
216
and in infancy appetitive traits have been identified as an early marker for future weight
gain (van Jaarsveld, Boniface, Llewellyn, & Wardle, 2014). My results show that appetitive
traits are associated with BMI in adults, however no causal inference can be made from the
results in Study 3.
Prospective longitudinal research is required to study the nature of directionality of the

above mentioned associations. Results from Study 3 however, provided the basis for these
questions to be addressed in future studies.
9.3.3 Loss to follow-up
Study 4 was limited by the very low response rate when the ‘Big Panel’ members were first
contacted, although response rates from this panel may be reduced by faulty e-mails or
excessive contact from different study interventions. Once eligible participants were
identified and 53 participants consented to take part in the ATTI, the study suffered in
particular from high drop-out rates and loss to follow-up, even though many attempts were
made to contact and retain participants through weekly reminders and personal e-mails.
High drop-out rates could have been a consequence of the lack of personal, face-to-face
contact, inherent with internet-based delivery (Arem & Irwin, 2011). Other possible
reasons for the loss to follow-up may have been associated with participants not relating to
the profiling from their AEBQ answers, and personal contact could have allowed for such
issues to be discussed. Also a ‘bank’ of tips for each trait to cover differences in
participants’ behaviours, including sending participants’ different tips to follow per week,
and further reminders to continue using them, might have improved participants’
involvement.
In Study 5, difficulties arose trying to obtain interviews from participants who did not finish
the study or who failed to provide their final weight. Five participants who were
considered not to have engaged with the ATTI were interviewed, out of a total of 21
interview participants. Results from more participants who did not follow the tips, weren’t
engaged or did not finish the study is likely to have enriched the findings. It is possible that
loss to follow-up could have been reduced by providing participants with incentives to
participate in Study 4 and then further incentives for participating in Study 5 (Treweek et
al., 2013).
217
9.4 Implications for future research, practice and policy
The findings of this thesis contribute to our understanding of appetite and weight in adults
and have implications for researchers, health practitioners and policymakers. The results
and limitations of this research opens up several areas for future inquiry in the field of
appetitive traits in adults.
First, there is a need to determine if the AEBQ is a valid measure of appetitive traits in adult
populations in different countries and in different socio-economic and ethnic groups (Cao
et al., 2012; Mallan et al., 2013; Sparks & Radnitz, 2012). Given their particular
characteristics, different relationships between appetitive traits and weight might be seen
in older adults and this requires further investigation (Jackson et al., 2014). Also, given the
self-report nature of the AEBQ, it could potentially be administered to adolescents who are
known to be one of the most vulnerable age groups in relation to weight issues (Lancet,
2012; Moreno et al., 2008). Evidence suggests that parents and adolescents may be
discordant in reporting dietary intake (Northstone, Smith, Cribb, & Emmett, 2013), with less
extreme responses given by self-reporting adolescents than those obtained via parent-
report questionnaires (Waters, Stewart-Brown, & Fitzpatrick, 2003). There is particular
interest from professionals who would like to obtain information about the relationship
between appetitive traits and weight in adolescents (Carnell et al., 2013). Validating the
AEBQ for use in adolescent samples could also enable tailored weight management
interventions focused on modifying appetitive traits directed at this particularly vulnerable
age-group (Neumark-Sztainer, Story, Perry, & Casey, 1999).
The findings of this thesis suggest the appetitive traits that have been most strongly
associated with obesity in children remain important into adulthood. While the evidence is
strongest for associations between BMI and both ‘food responsiveness’ and ‘satiety
responsiveness’ in childhood (Llewellyn & Wardle, 2015; Syrad et al., 2016), the strongest
relationships were seen between BMI and both ‘emotional over-eating’ and ‘emotional
under-eating’ in adults. Furthermore, overall, the associations between BMI and appetitive
traits in adults were found to be of lower magnitude than in children (Carnell & Wardle,
2008a). The high prevalence of weight loss attempts in overweight and obese adults
(Nicklas et al., 2012; Provencher et al., 2004; Wardle & Johnson, 2002), suggests that these
behaviours may play a role in the relationship between appetitive traits and BMI. Adults
may be managing their weight and eating behaviours (Larson et al., 2009; Nicklas et al.,
2012). Future research could involve asking participants if they are currently managing
218
their weight, and also to look at their ability to self-regulate their eating behaviour, given
that those trying to lose weight may not be doing so successfully (Johnson et al., 2012;
Johnson & Wardle, 2014). Studying the moderating effects of current weight loss attempts
and self-regulation of eating behaviour on the relationship between appetitive traits and
BMI could shed further light on adults’ appetitive behaviours and support appropriate
weight loss methods among overweight and obese adults.
Tracking of appetitive traits has been explored across childhood, where ‘food
responsiveness’, ‘emotional over-eating’ and ‘enjoyment of food’ were found to increase
from age four to age 10, in twins from the Twins Early Development Study (TEDS). ‘Satiety
responsiveness’, ‘slowness in eating’, ‘emotional under-eating’ and ‘food fussiness’ were
found to decrease with age (Ashcroft et al., 2008). Currently the TEDS cohort are between
20 and 22 years old (born between 1994 and 1996), and continue to participate in research
(Haworth, Davis, & Plomin, 2013). Administering the AEBQ to the TEDS twins could provide
evidence of the longitudinal continuity and stability of these traits from childhood into
adulthood.
Research is still needed to determine the heritability of these traits in adults. Appetitive
traits have been found to have a heritable component both in children (Carnell et al., 2008),
and infants (Llewellyn et al., 2012). As data becomes available from adult twin populations,
the twin method could be used to quantify genetic and environmental contributions to
AEBQ measured appetitive traits in adulthood (Llewellyn, van Jaarsveld, Johnson, Carnell, &
Wardle, 2010; Llewellyn & Wardle, 2015).
Experimental research using laboratory measures of appetite could be used to further

validate the AEBQ, as was carried out for the CEBQ (Carnell & Wardle, 2007). Laboratory
measured ‘eating in the absence of hunger’ (EAH), would help validate ‘food
responsiveness’ and ‘enjoyment of food’ (Birch et al., 2003; Fisher & Birch, 1999). Caloric
compensation studies could assess the validity of ‘satiety responsiveness’ and ‘slowness in
eating’ under laboratory conditions (Johnson & Birch, 1994; Mattes, Pierce, & Friedman,
1988). Furthermore, neuroimaging studies could also be used to track neurological
appetite pathways triggered by food cues to link the ‘homeostatic’ and ‘hedonic’
neurological appetitive pathways with obesity and assess them against psychometric
measures using the AEBQ (Carnell et al., 2013, 2012). Also validation of the AEBQ against
other measures of appetite such as the DEBQ could be conducted to explore differences
between food responsiveness and ‘external eating’. Validation against measures such as
219
the ‘Intuitive Eating Scale’ (IES and IES-2), might help distinguish differences between
measured ‘satiety responsiveness’ using the AEBQ and the intuitive aspect of satiation
measured by the IES or IES-2 (Tylka & Kroon Van Diest, 2013; Tylka, 2006). More recently,
the ‘Power of Food Scale’ (PFS) has been used to assess the psychological impact of today’s
food environment via three sub-scales (‘food available’, ‘food present’ and ‘food tasted’),
and additional validation could explore the relationship between the PFS and the AEBQ
scales such as ‘food responsiveness’, ‘enjoyment of food’ and ‘satiety responsiveness’, to
assess potential convergent validity between the scales.
Future research should also include obtaining objective measures of weight and height,
which could help reduce BMI under-reporting issues, and result in more accurate estimates
of associations between appetitive traits and BMI (Gorber et al., 2007). It would also be of
interest to obtain additional objective anthropometric measures, including waist
circumference which has previously been found to associate with appetitive traits in
children (Carnell & Wardle, 2008a). Measurement of appetitive traits could also be
correlated to dietary patterns (Emmett et al., 2015) and food preferences (Fildes et al.,
2015), to assess how different food choices relate to different appetitive traits.
This thesis has identified associations between individual appetitive traits and BMI in
adults, but more work is needed to confirm these findings and provide further insight into
the relationships between appetite and weight across adulthood. Ultimately this could help
to identify more effective ways to support individuals' weight loss attempts, and influence
health practitioners’ delivery of weight management advice. Given the rise in obesity
prevalence, there has been an upsurge of weight loss methods provided by a fast growing
industry (“Marketdata Enterprises, Inc.,” 2014). Tailoring advice to individual traits
capitalises on people’s desire to receive personalised advice and could enhance the effects
of other health-promoting messages (Kreuter et al., 1999). There is also a need for brief
interventions which provide simple advice (Clark et al., 2004; Mata et al., 2010), and there
is increasing interest in internet-based delivery of weight management advice as an
affordable option that enables greater coverage than in-person advice (Arem & Irwin, 2011;
Hartmann-Boyce et al., 2015).
The ATTI developed and tested in Studies 4 and 5 could be an acceptable and helpful
approach to weight management that addresses some of these issues. It was relatively
simple and involved instructing participants to follow the tips provided for a period of eight
weeks. Tips were tailored according to participants’ individual AEBQ scores. The
220
information in the tips appeared to help individuals to be able to better direct their
attention to their own ‘risky’ traits and to learn specific ways to help curb them.
Development of an app, and a website with appetitive trait information, as well as an on-
line forum, could improve the support needed by participants to improve their adherence
to the tips (Leske, Strodl, & Hou, 2012; Stubbs et al., 2011).
However, further work is required in order to make definitive recommendations about the
use of appetitive traits in weight management, including refinement of ATTI delivery and of
the tips themselves. Intervention participants may benefit from initial and final contact
with a health professional. Personal contact with a health professional is known to increase
weight loss success (Jebb et al., 2011; Metzgar et al., 2014). Personal contact would also
allow objectively measured height and weight to be taken, as well as a more personalised
AEBQ profiling with sufficient explanation of adverse appetitive traits and their
corresponding tips. However this might impact on the simplicity of the delivery, increasing
the resources required for implementation (Clark et al., 2004; Lally, Chipperfield, & Wardle,
2008; Mata et al., 2010), and would change the intervention’s status as a self-help weight
loss treatment (Hartmann-Boyce et al., 2015).
It may also be worth exploring whether the ATTI could be used alongside other weight loss
programmes to boost effectiveness and the feasibility of its use within different health care
settings. For example, there is the potential to use the ATTI in the primary care setting.
This would address individuals’ desire for input from a healthcare professional, and would
be a relatively simple, brief way of providing tailored weight management information to
better meet individual needs. In this context, it could either be stand alone, or it could be
incorporated into existing weight management or health promotion programmes.
Refining of some of the tips, increasing the number of tips, and varying suggestions of
behaviours to fit the needs of different participants could also improve future adherence to
the ATTI. For example the single ‘emotional over-eating’ tip could be improved by including
more tips which introduce the concept of changing emotions and negative thoughts around
food (Wardle et al., 2013). Instead of suggesting to participants with high ‘food
responsiveness’ to follow the “suggest doing things with friends that don’t involve food, like
going for a walk in the park” tip, maybe providing a different option such as: “if going out
with friends involves eating, try to make healthy food choices, and don’t get carried away
by what your friends are eating”, could improve adherence. These refinements could allow
the rigorous implementation of the ATTI as the sixth step for Quality Intervention
221
Development (6SQuID) defined by Wight et al., (2015), and intervention implementation in

relation to the MRC steps of complex intervention development (Craig et al., 2008). Results
from Studies 4 and 5 also indicate the need to explore why people dropped out of the
study. Further qualitative work with the target population may help to un-pick some of the
issues. Future studies and implementation of the ATTI could also further explore avenues
for reducing drop-out rates, through attempts at personal contact, different reminder
systems, and the use of incentives.
Lastly, this research has implications for policymakers because a better understanding of
the causes of obesity with respect to appetitive traits, highlights how structural changes
may be effective. For example, when considering the importance of ‘food responsiveness’
on adiposity risk, the way that foods are displayed at supermarkets could be manipulated
to discourage unplanned purchases of unhealthy foods. Implementing ‘nudging’27
techniques to encourage healthy food choices both in laboratory experiments and in
naturalistic settings, such as supermarkets, has shown positive results (van Kleef, Otten, &
van Trijp, 2012). Increased availability, visibility and wider assortments of healthy snacks
has been shown to facilitate healthy snack choice (Petrescu et al., 2016; van Kleef et al.,
2012). Reducing portion sizes was one of the ‘satiety responsiveness’ tips used in the ATTI.
Reducing the portion sizes of pre-prepared foods and beverages has been suggested at a
policy level and is supported by the Childhood Obesity Strategy for England, to reduce sugar
consumption (Reed, 2016). Overall, different strategies could be used to provide advice for
the presence of some of these traits on a broader population level.
9.5 Concluding remarks
This thesis aimed to address existing gaps in the literature relating to the relationship
between appetitive traits, weight, and weight management in adulthood. Overall, the
findings make an important contribution to this literature. The AEBQ is a novel measure of
27
Nudging techniques help to modify the environment, to help people make changes to their
behaviours without the conscious awareness of the participants (Petrescu et al., 2016; van Kleef et
al., 2012).
222
appetitive traits that will enable tracking of these traits, and their associations with BMI,
from infancy, into childhood, and now into adulthood. The findings from this thesis suggest
that traits associated with BMI in childhood are also associated with BMI in adulthood,
although associations are smaller. These associations suggest appetitive traits could be a
potential target for weight management interventions. The development and preliminary
testing of the ATTI found individuals’ AEBQ measured appetitive profiles can be used to
inform a tailored intervention to help people with overweight or obesity better manage
their weight. These findings highlight the potential importance of appetitive traits for
weight and weight management in adults, and pave the way for future research to explore
these relationships further.
223
References
References
Allison, D. B., & Baskin, M. L. (Eds.). (2009). Introduction. In Assessment methods for eating
behaviours and weight-related problems. Measures, theory and research. (2nd ed., pp.
vii–xiv). Thousand Oaks, CA: SAGE Publications, Inc.
Alm, M., Soroudi, N., Wylie-Rosett, J., Isasi, C. R., Suchday, S., Rieder, J., & Khan, U. (2008).
A qualitative assessment of barriers and facilitators to achieving behavior goals among
obese inner-city adolescents in a weight management program. The Diabetes
Educator, 34(2), 277–284. doi:10.1177/0145721708314182
Almirall, D., Nahum-Shani, I., Sherwood, N. E., & Murphy, S. A. (2014). Introduction to
SMART designs for the development of adaptive interventions: With application to
weight loss research. Translational Behavioral Medicine, 4(3), 260–274.
doi:10.1007/s13142-014-0265-0
American Educational Research Association, American Psychological Association, &

National Council on Measurement in Education. (1999). Standards for educational and
psychological testing. (American Psychological Association, Ed.). Washington, DC.
Anglé, S., Engblom, J., Eriksson, T., Kautiainen, S., Saha, M.-T., Lindfors, P., … Rimpelä, A.
(2009). Three Factor Eating Questionnaire-R18 as a measure of cognitive restraint,
uncontrolled eating and emotional eating in a sample of young Finnish females. The
International Journal of Behavioral Nutrition and Physical Activity, 6(1), 1–7.
doi:10.1186/1479-5868-6-41
Arem, H., & Irwin, M. (2011). A review of web-based weight loss interventions in adults.
Obesity Reviews, 12(5), e236–e243. doi:10.1111/j.1467-789X.2010.00787.x
Arnow, B., Kenardy, J., & Agras, W. S. (1995). The Emotional Eating Scale: The development
of a measure to assess coping with negative affect by eating. The International Journal
of Eating Disorders, 18(1), 79–90. Retrieved from
http://www.ncbi.nlm.nih.gov/pubmed/7670446
Ashcroft, J., Semmler, C., Carnell, S., van Jaarsveld, C. H. M., & Wardle, J. (2008). Continuity
and stability of eating behaviour traits in children. European Journal of Clinical
Nutrition, 62(8), 985–990. doi:10.1038/sj.ejcn.1602855
Avalos, L. C., & Tylka, T. L. (2006). Exploring a model of intuitive eating with college women.
Journal of Counseling Psychology, 53(4), 486–497. doi:10.1037/0022-0167.53.4.486
Baecke, J. A., Burema, J., Frijters, J. E., Hautvast, J. G., & van der Wiel-Wetzels, W. A. (1983).
Obesity in young Dutch adults: II, daily life-style and body mass index. International
Journal of Obesity, 7(1), 13—24. Retrieved from
http://europepmc.org/abstract/MED/6840963
Banasiak, S. J., Wertheim, E. H., Koerner, J., & Voudouris, N. J. (2001). Test-retest reliability
and internal consistency of a variety of measures of dietary restraint and body
concerns in a sample of adolescent girls. The International Journal of Eating Disorders,
29(1), 85–89. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11135339
224
References
Banna, J. C., Vera Becerra, L. E., Kaiser, L. L., & Townsend, M. S. (2010). Using qualitative
methods to improve questionnaires for spanish speakers: Assessing face validity of a
food behavior checklist. Journal of the American Dietetic Association, 110, 80–90.
doi:10.1016/j.jada.2009.10.002
Baños, R. M., Cebolla, A., Etchemendy, E., Felipe, S., Rasal, P., & Botella, C. (2011).
Validation of the dutch eating behavior questionnaire for children (DEBQ-C) for use
with Spanish children. Nutricion Hospitalaria, 26(4), 890–898.
doi:10.3305/nh.2011.26.4.5238
Barquera, S., Campos, I., & Rivera, J. A. (2013). Mexico attempts to tackle obesity: The
process, results, push backs and future challenges. Obesity Reviews, 14(Suppl 2), 69–
78. doi:10.1111/obr.12096
Bartle, N. C., Hill, C., Webber, L., van Jaarsveld, C. H. M., & Wardle, J. (2013). Emergence
and persistence of overweight and obesity in 7- to 11-year-old children. Obesity Facts,
6(5), 415–23. doi:10.1159/000355677
Beeken, R. J., Croker, H., Morris, S., Leurent, B., Omar, R., Nazareth, I., & Wardle, J. (2012).
Study protocol for the 10 Top Tips (10TT) Trial: Randomised controlled trial of habit-
based advice for weight control in general practice. BMC Public Health, 12(1), 1.
doi:10.1186/1471-2458-12-667
Beeken, R. J., Spanos, D., Fovargue, S., Hunter, R., Omar, R., Hassiotis, A., … Croker, H.
(2013). Piloting a manualised weight management programme (Shape Up-LD) for
overweight and obese persons with mild-moderate learning disabilities: Study
protocol for a pilot randomised controlled trial. Trials, 14(1), 71. doi:10.1186/1745-
6215-14-71
Birch, L. L. (1998). Psychological Influences on the Childhood Diet. The Journal of Nutrition,
128(2 Suppl), 407S–410S. Retrieved from
http://jn.nutrition.org/content/128/2/407S.full.pdf+html
Birch, L. L., & Fisher, J. O. (1998). Development of eating behaviors among children and
adolescents. Pediatrics, 101(Supplement 2), 539–249. Retrieved from
http://pediatrics.aappublications.org/content/101/Supplement_2/539.long
Birch, L. L., Fisher, J. O., & Davison, K. K. (2003). Learning to overeat: Maternal use of
restrictive feeding practices promotes girls’ eating in the absence of hunger. The
American Journal of Clinical Nutrition, 78(2), 215–220. Retrieved from
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2530927&tool=pmcentre
z&rendertype=abstract
Birch, L. L., Fisher, J. O., Grimm-Thomas, K., Markey, C. N., Sawyer, R., & Johnson, S. L.
(2001). Confirmatory factor analysis of the Child Feeding Questionnaire: A measure of
parental attitudes, beliefs and practices about child feeding and obesity proneness.
Appetite, 36(3), 201–210. doi:10.1006/appe.2001.0398
Blake, C. E., Bell, B. A., Freedman, D. A., Colabianchi, N., & Liese, A. D. (2013). The Eating
Identity Type Inventory (EITI). Development and associations with diet. Appetite, 69,
15–22. doi:10.1016/j.appet.2013.05.008
Blundell, J. E., & Cooling, J. (1999). High-fat and low-fat (behavioural) phenotypes: Biology
225
References
or environment? Proceedings of the Nutrition Society, 58, 773–777.
Blundell, J. E., de Graaf, K., Finlayson, G., Halford, J. C. G., Hetherington, M., King, N. A., &
Stubbs, J. (2009). Measuring food intake, hunger, satiety and satiation in the
laboratory. In D. B. Allison & M. L. Baskin (Eds.), Handbook of assessment methods for
eating behaviours and weight-related problems (2nd ed., pp. 283–325). Thousand
Oaks, CA: SAGE Publications, Inc.
Blundell, J. E., & Finlayson, G. (2004). Is susceptibility to weight gain characterized by

homeostatic or hedonic risk factors for overconsumption? Physiology & Behavior,
82(1), 21–25. doi:10.1016/j.physbeh.2004.04.021
Blundell, J. E., Stubbs, R. J., Golding, C., Croden, F., Alam, R., Whybrow, S., … Lawton, C. L.
(2005). Resistance and susceptibility to weight gain: Individual variability in response
to a high-fat diet. Physiology & Behavior, 86(5), 614–622.
doi:10.1016/j.physbeh.2005.08.052
Boggiano, M. M., Wenger, L. E., Mrug, S., Burgess, E. E., & Morgan, P. R. (2015). The Kids-
Palatable Eating Motives Scale: Relation to BMI and binge eating traits. Eating
Behaviors, 17, 69–73. doi:10.1016/j.eatbeh.2014.12.014
Boggiano, M. M., Wenger, L. E., Turan, B., Tatum, M. M., Morgan, P. R., & Sylvester, M. D.
(2015). Eating tasty food to cope. Longitudinal association with BMI. Appetite, 87,
365–370. doi:10.1016/j.appet.2015.01.008
Boutelle, K. N., Peterson, C. B., Rydell, S. A., Zucker, N. L., Cafri, G., & Harnack, L. (2011).
Two novel treatments to reduce overeating in overweight children: A randomised
controlled trial. Journal of Consulting and Clinical Psychology, 72(2), 181–204.
doi:10.1038/nature13314.A
Braet, C., & van Strein, T. (1997). Assessment of emotional, externally induced and
restrained behaviour in nine to twelve-year-old obese and non-obese children.
Behaviour Research and Theory, 35(9), 863–873.
Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitatitve Research
in Psychology, 3, 77–101. doi:http://dx.doi.org/10.1191/1478088706qp063oa PLEASE
Brown, R. E., & Kuk, J. L. (2015). Consequences of obesity and weight loss: A devil’s
advocate position. Obesity Reviews, 16, 77–87. doi:10.1111/obr.12232
Brownell, K. D., & Roberto, C. A. (2015). Strategic science with policy impact. The Lancet,
385(9986), 2445–2446. doi:10.1016/S0140-6736(14)62397-7
Bryant, E. J., Caudwell, P., Hopkins, M. E., King, N. A., & Blundell, J. E. (2012). Psycho-
markers of weight loss. The roles of TFEQ Disinhibition and Restraint in exercise-
induced weight management. Appetite, 58(1), 234–241.
doi:10.1016/j.appet.2011.09.006
Burgess, E. E., Turan, B., Lokken, K. L., Morse, A., & Boggiano, M. M. (2014). Profiling
motives behind hedonic eating. Preliminary validation of the Palatable Eating Motives
Scale. Appetite, 72, 66–72. doi:10.1016/j.appet.2013.09.016
Burke, L. E., Conroy, M. B., Sereika, S. M., Okan, U. E., Mindi, A. S., Acharya, S. D., … Glanz,
226
References
K. (2011). The effect of electronic self-monitoring on weight loss and dietary intake: A
randomized behavioral weight loss trial. Obesity, 19(2), 338–344.
doi:10.1038/oby.2010.208
Burke, L. E., Wang, J., & Sevick, M. A. (2011). Self-monitoring in weight loss: A systematic
review of the literature. Journal of the American Dietetic Association, 111(1), 92–102.
doi:10.1016/j.jada.2010.10.008
Caccialanza, R., Nicholls, D., Cena, H., Maccarini, L., Rezzani, C., Antonioli, L., … Roggi, C.
(2004). Validation of the Dutch Eating Behaviour Questionnaire parent version (DEBQ-
P) in the Italian population: A screening tool to detect differences in eating behaviour
among obese, overweight and normal-weight preadolescents. European Journal of
Clinical Nutrition, 58(9), 1217–1222. doi:10.1038/sj.ejcn.1601949
Cameron, R., & Evers, S. E. (1990). Self-report issues in obesity and weight management:
State of the art and future directions. Behavioral Assessment, 12(1), 91–106.
Retrieved from http://psycnet.apa.org/psycinfo/1990-18712-001
Campbell, F., Johnson, M., Messina, J., Guillaume, L., & Goyder, E. (2011). Behavioural
interventions for weight management in pregnancy: A systematic review of
quantitative and qualitative data. BMC Public Health, 11(1), 1–13. doi:10.1186/1471-
2458-11-491
Cao, Y. T., Svensson, V., Marcus, C., Zhang, J. D., & Sobko, T. (2012). Eating behaviour
patterns in Chinese children aged 12-18 months and association with relative weight -
factorial validation of the Children’s Eating Behaviour Questionnaire. The International
Journal of Behavioral Nutrition and Physical Activity, 9(5), 10–1186. doi:10.1186/1479-
5868-9-5
Cappelleri, J. C., Bushmakin, A. G., Gerber, R. A., Leidy, N. K., Sexton, C. C., Karlsson, J., &
Lowe, M. R. (2009). Evaluating the Power of Food Scale in obese subjects and a
general sample of individuals: development and measurement properties.
International Journal of Obesity, 33(8), 913–922. doi:10.1038/ijo.2009.107
Cappelleri, J. C., Bushmakin, A. G., Gerber, R. A., Leidy, N. K., Sexton, C. C., Lowe, M. R., &
Karlsson, J. (2009). Psychometric analysis of the Three-Factor Eating Questionnaire-
R21: Results from a large diverse sample of obese and non-obese participants.
Carnell, S., Benson, L., Pryor, K., & Driggin, E. (2013). Appetitive traits from infancy to
adolescence: Using behavioral and neural measures to investigate obesity risk.
Physiology & Behavior, 121, 79–88. doi:10.1016/j.physbeh.2013.02.015
Carnell, S., Gibson, C., Benson, L., Ochner, C. N., & Geliebter, A. (2012). Neuroimaging and
obesity: Current knowledge and future directions. Obesity Reviews, 13(1), 43–56.
doi:10.1111/j.1467-789X.2011.00927.x.Neuroimaging
Carnell, S., Haworth, C. M. A., Plomin, R., & Wardle, J. (2008). Genetic influence on appetite
in children. International Journal of Obesity, 32(10), 1468–73.
doi:10.1038/ijo.2008.127
Carnell, S., & Wardle, J. (2007). Measuring behavioural susceptibility to obesity: Validation
of the Child Eating Behaviour Questionnaire. Appetite, 48(1), 104–113.
227
References
doi:10.1016/j.appet.2006.07.075
Carnell, S., & Wardle, J. (2008a). Appetite and adiposity in children: Evidence for a
behavioral susceptibility theory of obesity. The American Journal of Clinical Nutrition,
88, 22–29. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18614720
Carnell, S., & Wardle, J. (2008b). Appetitive traits and child obesity: Measurement, origins
and implications for intervention. The Proceedings of the Nutrition Society, 67(4), 343–
355. doi:10.1017/S0029665108008641
Carruth, B. R., & Skinner, J. D. (2000). Revisiting the picky eater phenomenon: Neophobic
behaviors of young children. Journal of the American College of Nutrition, 19(6), 771–
780. doi:10.1080/07315724.2000.10718077
Carter, M. C., Burley, V. J., Nykjaer, C., & Cade, J. E. (2013). Adherence to a smartphone
application for weight loss compared to website and paper diary: Pilot randomized
controlled trial. Journal of Medical Internet Research, 15(4), 1–17.
doi:10.2196/jmir.2283
Carver, C. S., & Scheier, M. F. (1982). Control theory: A useful conceptual framework for
personality-social, clinical, and health psychology. Psychological Bulletin, 92(1), 111–
135. doi:10.1037/0033-2909.92.1.111
Cebolla, A., Barrada, J. R., Van Strein, T., Oliver, E., & Baños, R. M. (2014). Validation of the
Dutch Eating Behavior Questionnaire (DEBQ) in a sample of Spanish women. Appetite,
73, 58–64. doi:10.1016/j.appet.2013.10.014
Celis-Morales, C., Lara, J., & Mathers, J. C. (2015). Personalising nutritional guidance for
more effective behaviour change. Proceedings of the Nutrition Society, 44(June 2014),
1–9. doi:10.1017/S0029665114001633
Cepeda-Benito, A., Gleaves, D. H., Williams, T. L., & Erath, S. A. (2000). The development
and validation of the state and trait food-cravings questionnaires. Behavior Therapy,
31, 151–173. Retrieved from
http://www.sciencedirect.com/science/article/pii/S000578940080009X
Chang, S.-H., Stoll, C. R. T., Song, J., Varela, J. E., Eagon, C. J., & Colditz, G. A. (2014). The
effectiveness and risks of bariatric surgery. JAMA Surgery, 149(3), 275–287.
doi:10.1001/jamasurg.2013.3654
Christian, H., Knuiman, M., Bull, F., Timperio, A., Foster, S., Divitini, M., … Giles-Corti, B.
(2013). A new urban planning code’s impact on walking: The residential environments
project. American Journal of Public Health, 103(7), 1219–1228.
doi:10.2105/AJPH.2013.301230
Clark, M., Hampson, S. E., Avery, L., & Simpson, R. (2004). Effects of a tailored lifestyle self-
management intervention in patients with type 2 diabetes. British Journal of Health
Psychology, 9(3), 365–379. doi:10.1348/1359107041557066
Clark, P. J. (1956). The heritability of certain anthropometric characters as ascertained from

measurements of twins. American Journal of Human Genetics, 8(1), 49–54. Retrieved
from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1716621/
228
References
Cohen, D. A., Marsh, T., Williamson, S., Golinelli, D., & McKenzie, T. L. (2012). Impact and
cost-effectiveness of family fitness zones: A natural experiment in urban public parks.
Health and Place, 18(1), 39–45. doi:10.1016/j.healthplace.2011.09.008
Cohen, J. (1988). Statistical power analysis for the behavioural sciences (2nd ed.). Hillsdale,
NJ.: Lawrence Eribaum Associates, Inc.
Colchero, M. A., Popkin, B. M., Rivera, J. A., & Ng, S. W. (2016). Beverage purchases from
stores in Mexico under the excise tax on sugar sweetened beverages: Observational
study. The British Medical Journal, 352, h6704. doi:10.1136/bmj.h6704
Cole, D. A. (1987). Utility of confirmatory factor analysis in test validation research. Journal
of Consulting and Clinical Psychology, 55(4), 584–594. doi:10.1037/0022-
006X.55.4.584
Collins, C. E., Morgan, P. J., Hutchesson, M. J., & Callister, R. (2013). Efficacy of standard
versus enhanced features in a web-based commercial weight-loss program for obese
adults, part 2: Randomized controlled trial. Journal of Medical Internet Research,
15(7), 1–22. doi:10.2196/jmir.2626
Comrey, A. L., & Lee, H. B. (1992). A first course in factor analysis (Second edi.). Hillsdale,
NJ: Lawrence Erlbaum.
Coons, M. J., Demott, A., Buscemi, J., Duncan, J. M., Pellegrini, C. A., Steglitz, J., … Spring, B.
(2012). Technology Interventions to Curb Obesity: A Systematic Review of the Current
Literature. Current Cardiovascular Risk Reports, 6(2), 120–134. doi:10.1007/s12170-
012-0222-8
Craig, P., Dieppe, P., Macintyre, S., Michie, S., Nazareth, I., & Petticrew, M. (2008).
Developing and evaluating complex interventions: The new Medical Research Council
guidance. British Medical Journal, 337, a1655. doi:10.1136/bmj.a1655
Croker, H., Cooke, L., & Wardle, J. (2011). Appetitive behaviours of children attending
obesity treatment. Appetite, 57(2), 525–529. doi:10.1016/j.appet.2011.05.320
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika,
16(3), 297–334. Retrieved from http://link.springer.com/article/10.1007/BF02310555
Cruwys, T., Leverington, C. T., & Sheldon, A. M. (2015). An experimental investigation of the
consequences and social functions of fat talk in friendship groups. International
Journal of Eating Disorders, 49(1), 84–91. doi:10.1002/eat.22446
Cutler, G. J., Flood, A., Hannan, P. J., & Neumark-Sztainer, D. (2011). Multiple
sociodemographic and socioenvironmental characteristics are correlated with major
patterns of dietary intake in adolescents. Journal of the American Dietetic Association,
111(2), 230–240. doi:10.1016/j.jada.2010.10.052
Dalle Grave, R., Centis, E., Marzocchi, R., El Ghoch, M., & Marchesini, G. (2013). Major
factors for facilitating change in behavioral strategies to reduce obesity. Psychology
Research and Behavior Management, 6, 101–110. doi:10.2147/PRBM.S40460
Dalton, M., Finlayson, G., Hill, A., Blundell, J., Caudwell, P., Gibbons, C., & Näslund, E.
(2014). Preliminary validation and principal components analysis of the Control of
229
References
Eating Questionnaire (CoEQ) for the experience of food craving. European Journal of
Clinical Nutrition, 42(3), 92–101. doi:10.1249/JES.0000000000000019
Daniels, L. A., Mallan, K. M., Nicholson, J. M., Thorpe, K., Nambiar, S., Mauch, C. E., &
Magarey, A. (2015). An early feeding practices intervention for obesity prevention.
Pediatrics, 136(1), e40–e49. doi:10.1542/peds.2014-4108
Dansinger, M. L., Gleason, J. A., Griffith, J. L., Selker, H. P., & Schaefer, E. J. (2005).
Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets for weight loss
and heart disease risk reduction: A randomized trial. JAMA : The Journal of the
American Medical Association, 293(1), 43–53. doi:10.1001/jama.293.1.43
de Lauzon, B., Romon, M., Deschamps, V., Lafay, L., Borys, J.-M. M., Karlsson, J., … Charles,
M. A. (2004). The Three-Factor Eating Questionnaire-R18 is able to distinguish among
different eating patterns in a general population. The Journal of Nutrition, 134(9),
2372–2380. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15333731
de Lauzon-Guillain, B., Basdevant, A., Romon, M., Karlsson, J., Borys, J. M., & Charles, M. A.
(2006). Is restrained eating a risk factor for weight gain in a general population?
American Journal of Clinical Nutrition, 83(3), 132–138. Retrieved from
http://ajcn.nutrition.org/content/83/1/132.long
Deglaire, A., Méjean, C., Castetbon, K., Kesse-Guyot, E., Urbano, C., Hercberg, S., & Schlich,
P. (2012). Development of a questionnaire to assay recalled liking for salt, sweet and
fat. Food Quality and Preference, 23(2), 110–124. doi:10.1016/j.foodqual.2011.08.006
Diehl, J. M. (1999). Einstellungen zu Essen und Gewicht bei 11- bis 16jõhrigen Adoleszenten.
/ [Attitude to eating and body weight by 11- to 16-year-old adolescents].
Schweizerische Medizinische Wochenschrift, 129(5), 162–175. Retrieved from
http://bases.bireme.br/cgi-
bin/wxislind.exe/iah/online/?IsisScript=iah/iah.xis&src=google&base=ADOLEC&lang=p
&nextAction=lnk&exprSearch=10081074&indexSearch=ID
Domecq, J. P., Prutsky, G., Leppin, A., Sonbol, M. B., Altayar, O., Undavalli, C., … Murad, M.
H. (2015). Drugs commonly associated with weight change: A systematic review and
meta-analysis. The Journal of Clinical Endocrinology & Metabolism, 100(2), 363–370.
doi:10.1210/jc.2014-3421
Drapeau, V., Provencher, V., Lemieux, S., Després, J.-P., Bouchard, C., & Tremblay, A.
(2003). Do 6-y changes in eating behaviors predict changes in body weight? Results
from the Québec Family Study. International Journal of Obesity, 27, 808–814.
doi:10.1038/sj.ijo.0802303
Dugard, P., Todman, J., & Staines, H. (Eds.). (2010). Factor analysis. In Approaching
multivariate analysis (2nd ed., pp. 177–205). London and New York: Routledge.
Elfhag, K., & Rössner, S. (2005). Who succeeds in maintaining weight loss? A conceptual
review of factors associated with weight loss maintenance and weight regain. Obesity
Reviews, 6(1), 67–85. doi:10.1111/j.1467-789X.2005.00170.x
Emmett, P. M., Jones, L. R., & Northstone, K. (2015). Dietary patterns in the Avon
Longitudinal Study of Parents and Children. Nutrition Reviews, 73, 207–230.
doi:10.1093/nutrit/nuv055
230
References
Fabricatore, A. N., Wadden, T. A., Moore, R. H., Butryn, M. L., Heymsfield, S. B., & Nguyen,
A. M. (2009). Predictors of attrition and weight loss success: Results from a
randomized controlled trial. Behaviour Research and Therapy, 47(8), 685–691.
doi:10.1016/j.brat.2009.05.004
Faith, M. S., Carnell, S., & Kral, T. V. E. (2013). Genetics of food intake self-regulation in
childhood: Literature review and research opportunities. Human Heredity, 75(2-4),
80–89. doi:10.1159/000353879
Field, A. (2013). Discovering statistics using SPSS. (4th ed.). London: SAGE Publications Ltd.
Fildes, A., Mallan, K. M., Cooke, L., van Jaarsveld, C. H. M., Llewellyn, C. H., Fisher, A., &
Daniels, L. (2015). The relationship between appetite and food preferences in British
and Australian children. International Journal of Behavioral Nutrition and Physical
Activity, 12(1), 116. doi:10.1186/s12966-015-0275-4
Fildes, A., van Jaarsveld, C. H. M., Cooke, L., Wardle, J., & Llewellyn, C. H. (2016). Common
genetic architecture underlying young childrens food fussiness and liking for
vegetables and fruit. American Journal of Clinical Nutrition, 103(4), 1099–1104.
doi:10.3945/ajcn.115.122945
Finer, N. (2015). Predicting therapeutic weight loss. American Journal of Clinical Nutrition,
101(3), 419–420. doi:10.3945/ajcn.114.106195
Finlayson, G., Cecil, J., Higgs, S., Hill, A., & Hetherington, M. (2012). Susceptibility to weight
gain. Eating behaviour traits and physical activity as predictors of weight gain during
the first year of university. Appetite, 58(3), 1091–1098.
doi:10.1016/j.appet.2012.03.003
Finlayson, G., King, N. A., & Blundell, J. E. (2007). Liking vs. wanting food: Importance for
human appetite control and weight regulation. Neuroscience and Biobehavioral
Reviews, 31(7), 987–1002. doi:10.1016/j.neubiorev.2007.03.004
Fisher, J. O., & Birch, L. L. (1999). Restricting access to foods and children’s eating. Appetite,
32(1996), 405–19. Retrieved from
http://www.sciencedirect.com/science/article/pii/S0195666399902313
Fisher, J. O., & Birch, L. L. (2002). Eating in the absence of hunger and overweight in girls
from 5 to 7 y of age. The American Journal of Clinical Nutrition, 76(1), 226–231.
Retrieved from
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2604807&tool=pmcentre
z&rendertype=abstract
Fox, M. C., Ericsson, K. A., & Best, R. (2011). Do procedures for verbal reporting of thinking
have to be reactive? A meta-analysis and recommendations for best reporting
methods. Psychological Bulletin, 137(2), 316–344. doi:10.1037/a0021663
Franz, M. J., VanWormer, J. J., Crain, A. L., Boucher, J. L., Histon, T., Caplan, W., … Pronk, N.
P. (2007). Weight-loss outcomes: A systematic review and meta-analysis of weight-
loss clinical trials with a minimum 1-year follow-up. Journal of the American Dietetic
Association, 107(10), 1755–1767. doi:10.1016/j.jada.2007.07.017
French, S. A., Epstein, L. H., Jeffery, R. W., Blundell, J. E., & Wardle, J. (2012). Eating
231
References
behavior dimensions. Associations with energy intake and body weight. A review.
Appetite, 59(2), 541–549. doi:10.1016/j.appet.2012.07.001
Frosch, D. L., Mello, P., & Lerman, C. (2005). Behavioral consequences of testing for obesity
risk. Cancer Epidemiology, Biomarkers & Prevention, 14, 1485–1489. Retrieved from
http://cebp.aacrjournals.org/content/14/6/1485.full.pdf+html
Frühbeck, G., Toplak, H., Woodward, E., Yumuk, V., Maislos, M., & Oppert, J. M. (2013).
Obesity: The gateway to ill health - An EASO position statement on a rising public
health, clinical and scientific challenge in Europe. Obesity Facts, 6, 117–120.
doi:10.1159/000350627
Fuemmeler, B. F., Lovelady, C. A., Zucker, N. L., & Ostbye, T. (2013). Parental obesity
moderates the relationship between chilhood appetitive traits and weight. Obesity,
21(4), 815–823. doi:10.1002/oby.20144.Parental
Gallant, A. R., Tremblay, A., Pérusse, L., Bouchard, C., Després, J.-P., & Drapeau, V. (2010).
The Three-Factor Eating Questionnaire and BMI in adolescents: Results from the
Québec family study. The British Journal of Nutrition, 104(7), 1074–1079.
doi:10.1017/S0007114510001662
Ganley, R. M. (1988). Emotional eating and how it relates to dietary restraint, disinhibition,
and perceived hunger. International Journal of Eating Disorders, 7(5), 635–647.
doi:10.1002/1098-108X(198809)7:5<635::AID-EAT2260070507>3.0.CO;2-K
Gardner, C. D. (2012). Tailoring dietary approaches for weight loss. International Journal of
Obesity Supplements, 2(S1), S11–S15. doi:10.1038/ijosup.2012.4
Gardner, C. D., Kiazand, A., Kim, S., Stafford, R. S., Balise, R. R., Kraemer, H. C., & King, A. C.
(2007). Comparison of the Atkins, Zone, Ornish and LEARN diets for change in weight
and related risk factors among overweight premenopausal women. Journal of the
American Medical Association, 297(9), 969–978. Retrieved from
http://jama.jamanetwork.com/article.aspx?articleid=205916
Garner, D. M., & Garfinkel, P. E. (1979). The Eating Attitudes Test: An index of the
symptoms of anorexia nervosa. Psychological Medicine, 9(2), 273–9. Retrieved from
Gaskin, J. (2016). Stat Wiki. Confirmatory factor analysis. Retrieved June 10, 2016, from
http://statwiki.kolobkreations.com/index.php?title=Confirmatory_Factor_Analysis
Geliebter, A., & Aversa, A. (2003). Emotional eating in overweight, normal weight, and
underweight individuals. Eating Behaviors, 3(4), 341–347. doi:10.1016/S1471-
0153(02)00100-9
Gill, P., Stewart, K., Treasure, E., & Chadwick, B. (2008). Methods of data collection in
qualitative research: Interviews and focus groups. British Dental Journal, 204(6), 291–
295. doi:10.1038/bdj.2008.192
Göhner, W., Schlatterer, M., Frey, I., Berg, A., & Fuchs, R. (2012). Two-year follow-up of an
interdisciplinary cognitive-behavioral intervention program for obese adults. The
Journal of Psychology: Interdisciplinary and Applied, 146(4), 371–391.
doi:10.1080/00223980.2011.642023
232
References
Gorber, C. S., Tremblay, M., Moher, D., & Gorber, B. (2007). A comparison of direct vs. self-
report measures for assessing height, weight and body mass index: A systematic
review. Obesity Reviews, 8(4), 307–326. doi:10.1111/j.1467-789X.2007.00347.x
Gormally, J., Black, S., Daston, S., & Rardin, D. (1982). The assessment of binge eating
severity among obese persons. Addictive Behaviors, 7(1), 47–55. doi:10.1016/0306-
4603(82)90024-7
Grilo, C. M., & Pogue-Geile, M. F. (1991). The nature of environmental influences on weight
and obesity: A behavior genetic analysis. Psychological Bulletin, 110(3), 520–537.
doi:10.1037/0033-2909.110.3.520
Gudzune, K. A., Doshi, R. S., Mehta, A. K., Chaudhry, Z. W., Jacobs, D. K., Vakil, R. M., …
Clark, J. M. (2015). Efficacy of commercial weight-loss programs. Annals of Internal
Medicine, 162(7), 501–512. doi:10.7326/M14-2238
Gupta, H. (2014). Barriers to and facilitators of long term weight loss maintenance in adult
UK people: A thematic analysis. International Journal of Preventive Medicine, 5(12),
1512–1520. Retrieved from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4336981/
Guss, J. L., & Kissileff, H. R. (2000). Microstructural analyses of human ingestive patterns:
From description to mechanistic hypotheses. Neuroscience and Biobehavioral
Reviews, 24(99), 261–268. doi:10.1016/S0149-7634(99)00079-2
Halford, J. C. G., Boyland, E. J., Cooper, G. D., Dovey, T. M., Smith, C. J., Williams, N., …
Blundell, J. E. (2008). Children’s food preferences: Effects of weight status, food type,
branding and television food advertisements (commercials). International Journal of
Pediatric Obesity, 3(1), 31–38. doi:10.1080/17477160701645152
Harrold, J. A., Dovey, T. M., Blundell, J. E., & Halford, J. C. G. (2012). CNS regulation of
appetite. Neuropharmacology, 63(1), 3–17. doi:10.1016/j.neuropharm.2012.01.007
Hartmann-Boyce, J., Aveyard, P., Koshiaris, C., & Jebb, S. A. (2016). Development of tools to
study personal weight control strategies: OxFAB taxonomy. Obesity, 24(2), 314–320.
doi:10.1002/oby.21341
Hartmann-Boyce, J., Jebb, S. A., Fletcher, B. R., & Aveyard, P. (2015). Self-help for weight
loss in overweight and obese adults: Systematic review and meta-analysis. American
Journal of Public Health, 105(3), e43–e57. doi:10.2105/AJPH.2014.302389
Hartmann-Boyce, J., Johns, D. J., Jebb, S. A., Summerbell, C., & Aveyard, P. (2014).
Behavioural weight management programmes for adults assessed by trials conducted
in everyday contexts: Systematic review and meta-analysis. Obesity Reviews, 15(11),
920–932. doi:10.1111/obr.12220
Hawks, S. R., Merrill, C. G., Gast, J. A., & Hawks, J. F. (2004). Validation of the Motivation for
Eating Scale. Ecology of Food and Nutrition, 43(4), 307–326.
doi:10.1080/03670240490454714
Hawks, S. R., Merrill, R. M., & Madanat, H. N. (2004). The Intuitive Eating Scale:
Development and preliminary validation. American Journal of Health Education, 35(2),
90–99. doi:10.1080/19325037.2004.10603615
233
References
Haworth, C. M. A., Davis, O. S. P., & Plomin, R. (2013). Twins Early Development Study
(TEDS): A genetically sensitive investigation of cognitive and behavioral development
from childhood to young adulthood. Twin Research and Human Genetics, 16(1), 117–
125. doi:10.1017/thg.2012.91
Health and Social Care Information Centre. (2014). Health Survey for England 2013 - Trend
tables. Retrieved from http://www.hscic.gov.uk/catalogue/PUB16077/HSE2013-
Trend-commentary.pdf
Herman, C. P., & Mack, D. (1975). Restrained and unrestrained eating. Journal of
Personality, 43(4), 647–660. Retrieved from
Herman, C. P., & Polivy, J. (1975). Anxiety, restraint and eating behaviour. Journal of
Abnormal Psychology, 84(6), 666–672. Retrieved from
Herman, C. P., Polivy, J., Pliner, P., Threlkeld, J., & Munic, D. (1978). Distractibility in dieters
and nondieters: An alternative view of “externality”. Journal of Personality and Social
Psychology, 36(5), 536–548. Retrieved from
Higgs, S., & Thomas, J. (2016). Social influences on eating. Current Opinion in Behavioral
Sciences, 9, 1–6. doi:10.1016/j.cobeha.2015.10.005
Hill, C., Saxton, J., Webber, L., Blundell, J., & Wardle, J. (2009). The relative reinforcing value
of food predicts weight gain in a longitudinal study of 7-10-y-old children. American
Journal of Clinical Nutrition, 90(2), 276–281. doi:10.3945/ajcn.2009.27479
Hooper, L., Abdelhamid, A., Bunn, D., Brown, T., Summerbell, C. D., & Skeaff, C. M. (2015).
Effects of total fat intake on body weight. Cochrane Database of Systematic Reviews,
2015(8). doi:10.1002/14651858.CD011834
Howard, A. J., Mallan, K. M., Byrne, R., Magarey, A., & Daniels, L. A. (2012). Toddlers’ food
preferences. The impact of novel food exposure, maternal preferences and food
neophobia. Appetite, 59(3), 818–825. doi:10.1016/j.appet.2012.08.022
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–
55. doi:10.1080/10705519909540118
Hubbard, R. R., Palmberg, A., Lydecker, J., Green, B., Kelly, N. R., Trapp, S., & Bean, M. K.
(2015). Culturally-based communication about health, eating, and food: Development
and validation of the CHEF scale. Appetite, 96, 399–407.
doi:10.1016/j.appet.2015.09.024
Hulbert-Williams, L., Nicholls, W., Joy, J., & Hulbert-Williams, N. (2014). Initial validation of
the Mindful Eating Scale. Mindfulness, 5(6), 719–729. doi:10.1007/s12671-013-0227-5
Hutchenson, G., & Sofroniou, N. (1999). The multivariate social scientist: Introductory
statistics using generalized linear models. London, UK: Sage Publications Ltd.
IBM. (2013a). IBM SPSS AMOS, Version 22.0. Armonk, NY, USA: IBM Corp.
234
References
IBM. (2013b). IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY, USA: IBM Corp.
Jackson, S. E., Beeken, R. J., & Wardle, J. (2014). Predictors of weight loss in obese older
adults: Findings from the USA and the UK. Obesity Facts, 7(2), 102–110.
doi:10.1159/000362196
Jansen, A., Theunissen, N., Slechten, K., Nederkoorn, C., Boon, B., Mulkens, S., & Roefs, A.
(2003). Overweight children overeat after exposure to food cues. Eating Behaviors, 4,
197–209. doi:10.1016/S1471-0153(03)00011-4
Jebb, S. A., Ahern, A. L., Olson, A. D., Aston, L. M., Holzapfel, C., Stoll, J., … Caterson, I. D.
(2011). Primary care referral to a commercial provider for weight loss treatment
versus standard care: A randomised controlled trial. Lancet, 378(9801), 1485–1492.
doi:10.1016/S0140-6736(11)61344-5
Jiang, X., Yang, X., Zhang, Y., Wang, B., Sun, L., & Shang, L. (2014). Development and
preliminary validation of Chinese preschoolers’ eating behavior questionnaire. PloS
One, 9(2), e88255. doi:10.1371/journal.pone.0088255
Johns, D. J., Hartmann-Boyce, J., Jebb, S. A., & Aveyard, P. (2014). Diet or exercise
interventions vs combined behavioral weight management programs: A systematic
seview and meta-analysis of direct comparisons. Journal of the Academy of Nutrition
and Dietetics, 114(10), 1557–1568. doi:10.1016/j.jand.2014.07.005
Johnson, F., Beeken, R. J., Croker, H., & Wardle, J. (2014). Do weight perceptions among
obese adults in Great Britain match clinical definitions? Analysis of cross-sectional
surveys from 2007 and 2012. BMJ Open, 4(11), e005561. doi:10.1136/bmjopen-2014-
005561
Johnson, F., Pratt, M., & Wardle, J. (2012). Dietary restraint and self-regulation in eating
behavior. International Journal of Obesity, 36(5), 665–674. doi:10.1038/ijo.2011.156
Johnson, F., & Wardle, J. (2005). Dietary restraint, body dissatisfaction, and psychological
distress: A prospective analysis. Journal of Abnormal Psychology, 114(1), 119–125.
doi:10.1037/0021-843X.114.1.119
Johnson, F., & Wardle, J. (2014). Variety, palatability, and obesity. American Society for
Nutrition, 5, 851–859. doi:10.3945/an.114.007120
Johnson, L., Mander, A. P., Jones, L. R., Emmett, P. M., & Jebb, S. A. (2007). Is sugar-
sweetened beverage consumption associated with increased fatness in children?
Nutrition, 23(7-8), 557–563. doi:10.1016/j.nut.2007.05.005
Johnson, S. L., & Birch, L. L. (1994). Parents’ and children's adiposity and eating style.
Pediatrics, 94, 653–662. Retrieved from
http://pediatrics.aappublications.org/content/94/5/653.long
Johnston, B. C., Kanters, S., Bandayrel, K., Wu, P., Naji, F., Siemieniuk, R. A., … Mills, E. J.
(2014). Comparison of weight loss among named diet programs in overweight and
obese adults. Journal of the American Medical Association, 312(9), 923–933.
doi:10.1001/jama.2014.10397
Jolliffe, I. T. (2002). Principal Component Analysis, Second Edition (2nd Editio.). New York
235
References
Berlin Heidelberg: Springer-Verlag. Retrieved from

http://hbanaszak.mjr.uw.edu.pl/MarketingoweZastosowania/PCA/Jolliffe_2002_Princi
palComponentAnalysis.pdf
Kaiser, H. F. (1958). The Varimax criterion for analytical rotation in factor analysis.
Psychometrika, 23(3), 187–200. Retrieved from
http://link.springer.com/article/10.1007/BF02289233
Kaplan, D. L. (1980). Eating style of obese and nonobese males. Psychosomatic Medicine,
42, 529–538. Retrieved from
http://journals.lww.com/psychosomaticmedicine/Abstract/1980/11000/Eating_Style_
of_Obese_and_Nonobese_Males.1.aspx
Kaplan, H. I., & Kaplan, H. S. (1957). The psychosomatic concept of obesity. Journal of
Nervous and Mental Disease, 125, 181–201. Retrieved from
http://psycnet.apa.org/psycinfo/1959-04434-001
Karlsson, J., Persson, L. O., Sjöström, L., & Sullivan, M. (2000). Psychometric properties and
factor structure of the Three-Factor Eating Questionnaire (TFEQ) in obese men and
women. Results from the Swedish Obese Subjects (SOS) study. International Journal of
Obesity, 24(12), 1715–1725. Retrieved from
Kauer, J., Pelchat, M. L., Rozin, P., & Zickgraf, H. F. (2015). Adult picky eating.
Phenomenology, taste sensitivity, and psychological correlates. Appetite, 90, 219–228.
doi:10.1016/j.appet.2015.03.001
Kim, J.-Y., Oh, S., Steinhubl, S., Kim, S., Bae, W. K., Han, J. S., … Kim, M. J. (2015).
Effectiveness of 6 months of tailored text message reminders for obese male
participants in a worksite weight loss program: Randomized controlled trial. JMIR
mHealth and uHealth, 3(1), e14. doi:10.2196/mhealth.3949
Kirk, S. F. L., Penney, T. L., McHugh, T.-L. F., & Sharma, A. M. (2012). Effective weight
management practice: A review of the lifestyle intervention evidence. International
Journal of Obesity, 36(2), 178–185. doi:10.1038/ijo.2011.80
Kirk, S. F. L., Tytus, R., Tsuyuki, R. T., & Sharma, A. M. (2012). Weight management
experiences of overweight and obese Canadian adults: Findings from a national
survey. Chronic Diseases and Injuries in Canada, 32(2), 63–69. Retrieved from
Kleinert, S., & Horton, R. (2015). Rethinking and reframing obesity. The Lancet, 6736(14),
15–16. doi:10.1016/S0140-6736(15)60163-5
Koenders, P. G., & Van Strein, T. (2011). Emotional eating, rather than lifestyle behavior,
drives weight gain in a prospective study in 1562 employees. Journal of Occupational
and Environmental Medicine, 53(11), 1287–1293.
doi:10.1097/JOM.0b013e31823078a2
Krall, J. S., & Lohse, B. (2011). Validation of a measure of the Satter eating competence
model with low-income females. The International Journal of Behavioral Nutrition and
Physical Activity, 8(1), 26. doi:10.1186/1479-5868-8-26
236
References
Kreuter, M. W., Strecher, V. J., & Glassman, B. (1999). One size does not fit all: The case for
tailoring print materials. Annals of Behavioral Medicine, 21(4), 276–283.
doi:10.1007/BF02895958
Kroon Van Diest, A. M., & Tylka, T. L. (2010). The Caregiver Eating Messages Scale:
Development and psychometric investigation. Body Image, 7(4), 317–326.
doi:10.1016/j.bodyim.2010.06.002
Kuczmarski, M. F., Kuczmarski, R. J., & Najjar, M. (2001). Effects of age on validity of self-
height, weight, and body mass index: Findings from the third National Health and
Nutrition Examination Survey, 1988-1994. Journal of the American Dietetic
Association, 101, 28–34. doi:10.1016/S0002-8223(01)00008-6
Kushner, R. F. (2014). Weight loss strategies for treatment of obesity. Progress in

Cardiovascular Diseases, 56(4), 465–472. doi:10.1016/j.pcad.2013.09.005
Laessle, R. G., Tuschl, R. J., Kotthaus, B. C., & Pirke, K. M. (1989). Behavioral and biological
correlates of dietary restraint in normal life. Appetite, 12(2), 83–94. doi:10.1016/0195-
6663(89)90098-6
Lally, P., Chipperfield, A., & Wardle, J. (2008). Healthy habits: Efficacy of simple advice on
weight control based on a habit-formation model. International Journal of Obesity,
32(4), 700–777. doi:10.1038/sj.ijo.0803771
Lally, P., & Gardner, B. (2011). Promoting habit formation. Health Psychology Review, 7(Supl
1), 1–22. doi:10.1080/17437199.2011.603640
Lally, P., Wardle, J., & Gardner, B. (2011). Experiences of habit formation: A qualitative
study. Psychology, Health & Medicine, 16(4), 484–489.
doi:10.1080/13548506.2011.555774
Lancaster, G., Dodd, S., & Williamson, P. R. (2004). Design and analysis of pilot studies:
Recommendations for good practice. Journal of Evaluation in Clinical Practice, 10(2),
307–312. Retrieved from http://dx.doi.org/ 10.1111/j..2002.384.doc.x
Lancet, T. (2012). Putting adolescents at the centre of health and development. Lancet,
379(9826), 1561. doi:10.1016/S0140-6736(12)60536-4
Lanigan, J. A., Wells, J. C., Lawson, M. S., & Lucas, A. (2001). Validation of food diary method
for assessment of dietary energy and macronutrient intake in infants and children
aged 6-24 months. European Journal of Clinical Nutrition, 55(2), 124–9. Retrieved from
Lappalainen, R., & Epstein, L. H. (1990). A behavioral economics analysis of food choice in
humans. Appetite, 14, 81–93. doi:0195-6663(90)90002-P [pii]
Larson, N. I., Neumark-Sztainer, D., & Story, M. (2009). Weight control behaviors and
dietary intake among adolescents and young adults: longitudinal findings from Project
EAT. Journal of the American Dietetic Association, 109(11), 1869–1877.
doi:10.1016/j.jada.2009.08.016
Laureati, M., Bergamaschi, V., & Pagliarini, E. (2015). Assessing childhood food neophobia:
Validation of a scale in Italian primary school children. Food Quality and Preference,
237
References
40, 8–15. doi:10.1016/j.foodqual.2014.08.003
Leske, S., Strodl, E., & Hou, X.-Y. (2012). A qualitative study of the determinants of dieting
and non-dieting approaches in overweight/obese Australian adults. BMC Public
Health, 12, 1086. doi:10.1186/1471-2458-12-1086
Lindroos, A. K., Lissner, L., Mathiassen, M. E., Karlsson, J., Sullivan, M., Bengtsson, C., &
Sjostrom, L. (1997). Dietary intake in relation to restrained eating, disinhibition, and
hunger in obese and nonobese Swedish women. Obesity Research, 5(3), 175–182.
doi:10.1002/j.1550-8528.1997.tb00290.x
Llewellyn, C. H., Carnell, S., & Wardle, J. (2011). Eating behaviour and weight in children. In
L. A. A. Moreno, I. Pigeot, & W. Ahrens (Eds.), Epidemiology of obesity in children and
asolescents. Prevalence and etiology (Kindle.). New York, NY, USA: Springer Science +
Business Media, LLC.
Llewellyn, C. H., Trzaskowski, M., Plomin, R., & Wardle, J. (2013). Finding the missing
heritability in pediatric obesity: The contribution of genome-wide complex trait
analysis. International Journal of Obesity, 37(11), 1506–1509. doi:10.1038/ijo.2013.30
Llewellyn, C. H., Trzaskowski, M., van Jaarsveld, C. H. M., Plomin, R., & Wardle, J. (2014).
Satiety mechanisms in genetic risk of obesity. JAMA Pediatrics, 168(4), 338–344.
doi:10.1001/jamapediatrics.2013.4944
Llewellyn, C. H., van Jaarsveld, C. H. M., Johnson, L., Carnell, S., & Wardle, J. (2010). Nature
and nurture in infant appetite: Analysis of the Gemini twin birth cohort. American
Journal of Clinical Nutrition, 91(12), 1172–1179.
doi:10.3945/ajcn.2009.28868.INTRODUCTION
Llewellyn, C. H., van Jaarsveld, C. H. M., Johnson, L., Carnell, S., & Wardle, J. (2011).
Development and factor structure of the Baby Eating Behaviour Questionnaire in the
Gemini birth cohort. Appetite, 57(2), 388–396. doi:10.1016/j.appet.2011.05.324
Llewellyn, C. H., van Jaarsveld, C. H. M., Plomin, R., Fisher, A., & Wardle, J. (2012). Inherited
behavioral susceptibility to adiposity in infancy: A multivariate genetic analysis of
appetite and weight in the Gemini birth cohort. American Journal of Clinical Nutrition,
95, 633–639. doi:10.3945/ajcn.111.023671.Am
Llewellyn, C. H., & Wardle, J. (2015). Behavioral susceptibility to obesity: Gene–

environment interplay in the development of weight. Physiology & Behavior, 152,
494–501. doi:10.1016/j.physbeh.2015.07.006
Loewen, R., & Pliner, P. (2000). The Food Situations Questionnaire: A measure of children’s
willingness to try novel foods in stimulating and non-stimulating situations. Appetite,
35(3), 239–250. doi:10.1006/appe.2000.0353
Loh, D. A., Moy, F. M., Zaharan, N. L., & Mohamed, Z. (2013). Eating behaviour among
multi-ethnic adolescents in a middle-income country as measured by the self-reported
Children’s Eating Behaviour Questionnaire. PLoS One, 8(12), e82885.
doi:10.1371/journal.pone.0082885
Lohse, B., Satter, E., Horacek, T., Gebreselassie, T., & Oakland, M. J. (2007). Measuring
eating competence: Psychometric properties and validity of the ecSatter Inventory.
238
References
Journal of Nutrition Education and Behavior, 39(5), S154–S166.

doi:10.1016/j.jneb.2007.04.371
Lowe, M. R., Butryn, M. L., Didie, E. R., Annunziato, R. A., Thomas, J. G., Crerand, C. E., …
Halford, J. C. G. (2009). The Power of Food Scale. A new measure of the psychological
influence of the food environment. Appetite, 53(1), 114–118.
doi:10.1016/j.appet.2009.05.016
Mackenbach, J. D., Tiemeier, H., Ende, J. Van Der, Nijs, I. M. T., Jaddoe, V. W. V, Hofman, A.,
… Jansen, P. W. (2012). Relation of emotional and behavioral problems with body
mass index in preschool children: The generation R study. Journal of Developmental
and Behavioral Pediatrics, 33(8), 641–648. doi:10.1097/DBP.0b013e31826419b8
Madigan, C. D., Daley, A. J., Kabir, E., Aveyard, P., & Brown, W. (2015). Cluster analysis of
behavioural weight management strategies and associations with weight change in
young women: A longitudinal analysis. International Journal of Obesity, 39(11), 1601–
1606. doi:10.1038/ijo.2015.116
Maes, H. H. M., Neale, M. C., & Eaves, L. J. (1997). Genetic and environmental factors in
relative body weight and human adiposity. Behavior Genetics, 27(4), 325–351.
Retrieved from http://link.springer.com/article/10.1023/A:1025635913927
Malik, V. S., & Hu, F. B. (2007). Popular weight-loss diets: From evidence to practice. Nature
Clinical Practice. Cardiovascular Medicine, 4(1), 34–41. doi:10.1038/ncpcardio0726
Malik, V. S., Willett, W. C., & Hu, F. B. (2012). Global obesity: Trends, risk factors and policy
implications. Nature Reviews. Endocrinology, 9(1), 13–27.
doi:10.1038/nrendo.2012.199
Mallan, K. M., Liu, W. H., Mehta, R. J., Daniels, L. A., Magarey, A., & Battistutta, D. (2013).
Maternal report of young children’s eating styles. Validation of the Children's Eating
Behaviour Questionnaire in three ethnically diverse Australian samples. Appetite, 64,
48–55. doi:10.1016/j.appet.2013.01.003
Mallan, K. M., Nambiar, S., Magarey, A. M., & Daniels, L. A. (2014). Satiety responsiveness in
toddlerhood predicts energy intake and weight status at four years of age. Appetite,
74, 79–85. doi:10.1016/j.appet.2013.12.001
Marketdata Enterprises, Inc. (2014). The U.S. Weight loss market: 2014 Status report and
forecast. Report # FS45. Retrieved June 16, 2016, from
http://www.marketdataenterprises.com/studies/#WORKSITEWEIGHTLOSS
Marlow, L. A., & Wardle, J. (2014). Development of a scale to assess cancer stigma in the
non-patient population. BMC Cancer, 14, 285. doi:10.1186/1471-2407-14-285
Mata, J., Todd, P. M., & Lippke, S. (2010). When weight management lasts. Lower perceived
rule complexity increases adherence. Appetite, 54(1), 37–43.
doi:10.1016/j.appet.2009.09.004
Mattes, R. D., Pierce, C. B., & Friedman, M. I. (1988). Daily caloric intake of normal-weight
adults: Response to changes in dietary energy density of a luncheon meal. American
Journal of Clinical Nutrition, 48, 214–219.
239
References
Meisel, S. F., Beeken, R. J., van Jaarsveld, C. H. M., & Wardle, J. (2015). Genetic
susceptibility testing and readiness to control weight: Results from a randomized
controlled trial. Obesity, 23(2), 305–312. doi:10.1002/oby.20958
Meisel, S. F., & Wardle, J. (2014a). “Battling my biology”: Psychological effects of genetic
testing for risk of weight gain. Journal of Genetic Counseling, 23(2), 179–186.
doi:10.1007/s10897-013-9628-9
Meisel, S. F., & Wardle, J. (2014b). Responses to FTO genetic test feedback for obesity in a
sample of overweight adults: A qualitative analysis. Genes & Nutrition, 9(1), 374.
doi:10.1007/s12263-013-0374-2
Metzgar, C. J., Preston, A. G., Miller, D. L., & Nickols-Richardson, S. M. (2014). Facilitators
and barriers to weight loss and weight loss maintenance: A qualitative exploration.
Journal of Human Nutrition and Dietetics, 28(6), 593–603. doi:10.1111/jhn.12273
Meule, A., Hermann, T., & Kübler, A. (2014). A short version of the food cravings
questionnaire-trait: The FCQ-T-reduced. Frontiers in Psychology, 5, 190.
doi:10.3389/fpsyg.2014.00190
Meyer, J. E., & Pudel, V. E. (1972). Experimental studies in obese and normal weight
subjects. Journal of Psychosomatic Research, 16, 305–308. Retrieved from
http://www.sciencedirect.com.libproxy.ucl.ac.uk/science/article/pii/00223999729001
41
Michie, S., Abraham, C., Whittington, C., McAteer, J., & Gupta, S. (2009). Effective
techniques in healthy eating and physical activity interventions: A meta-regression.
Health Psychology, 28(6), 690–701. doi:10.1037/a0016136
Michie, S., Ashford, S., Sniehotta, F. F., Dombrowski, S. U., Bishop, A., & French, D. P.
(2011). A refined taxonomy of behaviour change techniques to help people change
their physical activity and healthy eating behaviours: The CALO-RE taxonomy.
Psychology and Health, 26(11), 1479–1498. doi:10.1080/08870446.2010.540664
Michie, S., Atkins, L., & West, R. (2014). The Behaviour Change Wheel. A guide to designing
interventions. Great Britain: Silverback Publishing.
Michie, S., Richardson, M., Johnston, M., Abraham, C., Francis, J., Hardeman, W., … Wood,
C. E. (2013). The behavior change technique taxonomy (v1) of 93 hierarchically
clustered techniques: Building an international consensus for the reporting of
behavior change interventions. Annals of Behavioral Medicine, 46(1), 81–95.
doi:10.1007/s12160-013-9486-6
Michie, S., West, R., Brown, J., & Gainforth, H. (2014). ABC of Behaviour Change Theories.
An essential resource for researchers policy makers and practicioners. Great Britain:
Silverback Publishing.
Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for
systematic reviews and meta-analyses: The PRISMA statement. Annals of Internal
Medicine, 339(4), 332–336. doi:10.1136/bmj.b2535
Moreno, L. A., González-Gross, M., Kersting, M., Molnár, D., de Henauw, S., Beghin, L., …
Marcos, A. (2008). Assessing, understanding and modifying nutritional status, eating
240
References
habits and physical activity in European adolescents: The HELENA (Healthy Lifestyle in
Europe by Nutrition in Adolescence) Study. Public Health Nutrition, 11(3), 288–99.
doi:10.1017/S1368980007000535
Morse, J. M. (1995). The significance of saturation. Qualitative Health Research, 5(2), 147–
149. Retrieved from http://qhr.sagepub.com.libproxy.ucl.ac.uk/content/5/2/147
National Institute of Health. (1998). Clinical guidelines on the identification, evaluation and
treatment of overweight and obesity in adults. The Evidence Report. National
Institutes of Health. National Heart, Lung and Blood Institute.
Naude, C. E., Schoonees, A., Senekal, M., Young, T., Garner, P., & Volmink, J. (2014). Low
carbohydrate versus isoenergetic balanced diets for reducing weight and
cardiovascular risk: A systematic review and meta-analysis. PLoS One, 9(7), e100652.
doi:10.1371/journal.pone.0100652
Nawaz, H., Chan, W., Abdulrahman, M., Larson, D., & Katz, D. L. (2001). Self-reported
weight and height. American Journal of Preventive Medicine, 20(4), 294–298.
doi:10.1016/S0749-3797(01)00293-8
Neumark-Sztainer, D., Story, M., Perry, C., & Casey, M. A. (1999). Factors affecting food
choices in adolescents: Findings from focus-group discussions with adolescents.
Journal of the American Dietetic Association, 99, 929–37.
Neumark-Sztainer, D., Wall, M., Larson, N. I., Eisenberg, M. E., & Loth, K. (2011). Dieting and
disordered eating behaviors from adolescence to young adulthood: findings from a
10-year longitudinal study. Journal of the American Dietetic Association, 111(7), 1004–
11. doi:10.1016/j.jada.2011.04.012
Neville, S., & Pickard, J. (2016, March). UK tax on sugary drinks is “nannying” and
“impractical.” Financial Times. London, UK. Retrieved from
http://www.ft.com/cms/s/0/5b41e47a-ec53-11e5-bb79-
2303682345c8.html#axzz46sie4NCa
Ng, M., Fleming, T. D., Robinson, M., Thomson, B., Graetz, N., Margono, C., … Gakidou, E.
(2014). Global, regional, and national prevalence of overweight and obesity in children
and adults during 1980-2013: A systematic analysis for the Global Burden of Disease
Study 2013. Lancet, 384(9945), 766–781. doi:10.1016/S0140-6736(14)60460-8
Nguyen, N., Champion, J. K., Ponce, J., Quebbemann, B., Patterson, E., Pham, B., … Favretti,
F. (2012). A review of unmet needs in obesity management. Obesity Surgery, 22(6),
956–966. doi:10.1007/s11695-012-0634-z
NHS Choices. Your health, your choices. (2016). Obesity treatment. Retrieved June 19,
2016, from http://www.nhs.uk/Conditions/Obesity/Pages/Treatment.aspx
NICE Clinical Guideline 189. (2014). Obesity: Identification , assessment and management of
overweight and obesity in children, young people and adults. Retrieved from
http://www.nice.org.uk/guidance/cg189
Nicklas, J. M., Huskey, K. W., Davis, R. B., & Wee, C. C. (2012). Successful weight loss among
obese U.S. adults. American Journal of Preventive Medicine, 42(5), 481–485.
doi:10.1016/j.amepre.2012.01.005
241
References
Nijs, I. M. T., Franken, I. H. A., & Muris, P. (2007). The modified Trait and State Food-
Cravings Questionnaires: Development and validation of a general index of food
craving. Appetite, 49(1), 38–46. doi:10.1016/j.appet.2006.11.001
Nolan, L. J., Halperin, L. B., & Geliebter, A. (2010). Emotional Appetite Questionnaire.
Construct validity and relationship with BMI. Appetite, 54(2), 314–319.
doi:10.1016/j.appet.2009.12.004
Nordmann, A. J., Suter-Zimmermann, K., Bucher, H. C., Shai, I., Tuttle, K. R., Estruch, R., &
Briel, M. (2011). Meta-analysis comparing Mediterranean to low-fat diets for
modification of cardiovascular risk factors. The American Journal of Medicine, 124(9),
841–851. doi:10.1016/j.amjmed.2011.04.024
Northstone, K., Smith, A. D., Cribb, V. L., & Emmett, P. M. (2013). Dietary patterns in UK
adolescents obtained from a dual-source FFQ and their associations with socio-
economic position, nutrient intake and modes of eating. Public Health Nutrition, 17(7),
1–10. doi:10.1017/S1368980013001547
O’Rahilly, S., & Farooqi, I. S. (2008). Human obesity as a heritable disorder of the central
control of energy balance. International Journal of Obesity, 32, S55–S61.
doi:10.1038/ijo.2008.239
Ogden, C. L., Yanovski, S. Z., Carroll, M. D., & Flegal, K. M. (2007). The epidemiology of
obesity. Gastroenterology, 132(6), 2087–102. doi:10.1053/j.gastro.2007.03.052
Ogden, J., Liakopoulou, E., Antilliou, G., & Gough, G. (2012). The Meaning of Food (MOF):
The development of a new measurement tool. European Eating Disorders Review,
20(5), 423–426. doi:10.1002/erv.1141
Oppenheim, A. N. (2003). Questionnaire design, interviewing and attitude measurements.

(3rd ed.). London, UK: Continuum.
Pagoto, S., & Appelhans, B. (2013). A call for an end to the diet debates. JAMA, 01655, 2–3.
doi:10.1186/1475-2891-10-9.6
Parkinson, K. N., Drewett, R. F., Le Couteur, A. S., & Adamson, A. J. (2010). Do maternal
ratings of appetite in infants predict later Child Eating Behaviour Questionnaire scores
and body mass index? Appetite, 54(1), 186–190. doi:10.1016/j.appet.2009.10.007
Petrescu, D. C., Hollands, G. J., Couturier, D.-L., Ng, Y.-L., Marteau, T. M., Hollands, G., …
Zwicker, L. (2016). Public acceptability in the UK and USA of nudging to reduce
obesity: The example of reducing sugar-sweetened beverages consumption. Plos One,
11(6), e0155995. doi:10.1371/journal.pone.0155995
Piernas, C., Barquera, S., & Popkin, B. M. (2014). Current patterns of water and beverage
consumption among Mexican children and adolescents aged 1–18 years: Analysis of
the Mexican National Health and Nutrition Survey 2012. Public Health Nutrition,
17(10), 2166–2175. doi:10.1017/S1368980014000998
Piernas, C., Ng, S. W., & Popkin, B. (2013). Trends in purchases and intake of foods and
beverages containing caloric and low-calorie sweeteners over the last decade in the
United States. Pediatric Obesity, 8(4), 294–306. doi:10.1111/j.2047-
6310.2013.00153.x
242
References
Pliner, P. (1994). Development of measures of food neophobia in children. Appetite, 23,

147–163. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7864609
Pliner, P., & Hobden, K. (1992). Development of a scale to measure the trait of food
neophobia in humans. Appetite, 19, 105–20. Retrieved from
Polivy, J., & Herman, C. P. (1976a). Clinical depression and weight change: A complex
relation. Journal of Abnormal Psychology, 85(3), 338–340. Retrieved from
http://psycnet.apa.org/journals/abn/85/3/338/
Polivy, J., & Herman, C. P. (1976b). Effects of alcohol on eating behavior: Influence of mood
and perceived intoxication. Journal of Abnormal Psychology, 85(6), 601–606.
Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/993457
Polivy, J., & Herman, C. P. (2004). Sociocultural idealization of thin female body shapes: An
introduction to the special issue on body image and eating disorders. Journal of Social
and Clinical Psychology, 23(1), 1–6. doi:10.1521/jscp.23.1.1.26986
Polivy, J., Herman, C. P., Younger, J. C., & Erskine, B. (1979). Effects of a model on eating
behavior: The induction of a restrained eating style. Journal of Personality, 47(1), 100–
17. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/430328
Popkess-Vawter, S., Gerkovich, M. M., & Wendel, S. (2000). Reliability and validity of the
overeating tension scale. Journal of Nursing Measurement, 8(2), 145–60. Retrieved
from
http://www.ingentaconnect.com/content/springer/nmeas/2000/00000008/00000002
/art00005
Powers, S. W., Chamberlin, L. A., van Schaick, K. B., Sherman, S. N., & Whitaker, R. C. (2006).
Maternal feeding strategies, child eating behaviors, and child BMI in low-income
African-American preschoolers. Obesity, 14(11), 2026–2033.
doi:10.1038/oby.2006.237
Provencher, V., Drapeau, V., Tremblay, A., Després, J.-P., Bouchard, C., & Lemieux, S.
(2004). Eating behaviours, dietary profile and body composition according to dieting
history in men and women of the Québec Family Study. The British Journal of
Nutrition, 91(6), 997–1004. doi:10.1079/BJN20041115
Puhl, R., & Heuer, C. (2009). The stigma of obesity: A review and update. Obesity, 17(5),
941–964. doi:10.1038/oby.2008.636
QSR International Pty Ltd. (2012). NVivo qualitative data analysis software (version 10).
Rapoport, L., Clark, M., & Wardle, J. (2000). Evaluation of a modified cognitive-behavioural
programme for weight management. International Journal of Obesity, 24(12), 1726–
37. doi:10.1038/sj.ijo.0801465
Reed, E. (2016). Childhood obesity & food policy in the UK. Department of Health. Retrieved
June 15, 2016, from
http://ec.europa.eu/health/nutrition_physical_activity/docs/ev_20160406_co01_en.p
df
243
References
Roberto, C. A., Swinburn, B. A., Hawkes, C., Huang, T. T.-K., Costa, S. A., Ashe, M., …
Brownell, K. D. (2015). Patchy progress on obesity prevention: Emerging examples,
entrenched barriers, and new thinking. The Lancet, 385(9985), 2400–2409.
doi:10.1016/S0140-6736(14)61744-X
Rodenburg, G., Kremers, S. P. J., Oenema, A., & van de Mheen, D. (2012). Associations of
children’s appetitive traits with weight and dietary behaviours in the context of
general parenting. PLoS One, 7(12), e50642. doi:10.1371/journal.pone.0050642
Rollins, B. Y., Riggs, N. R., Spruijt-Metz, D., McClain, A. D., Chou, C.-P., & Pentz, M. A. (2014).
Psychometrics of the Eating in Emotional Situations Questionnaire (EESQ) among low-
income Latino elementary-school children. Eating Behaviors, 12(2), 156–159.
doi:10.1016/j.eatbeh.2011.01.004.Psychometrics
Rosenbaum, M., Hirsch, J., Gallagher, D. A., & Leibel, R. L. (2008). Long-term persistence of
adaptive thermogenesis in subjects who have maintained a reduced body weight.
American Journal of Clinical Nutrition, 88(4), 906–912. doi:88/4/906 [pii]
Rowland, M. L. (1990). Self-reported weight and height. The American Journal of Clinical
Nutrition, 52(6), 1125–1133. Retrieved from
http://ajcn.nutrition.org/content/52/6/1125.long
Santos, J. L., Ho-Urriola, J. A., González, A., Smalley, S. V, Domínguez-Vásquez, P., Cataldo,
R., … Hodgson, M. I. (2011). Association between eating behavior scores and obesity
in Chilean children. Nutrition Journal, 10(1), 108. doi:10.1186/1475-2891-10-108
Schacht, M., Richter-appelt, H., & Schulte-markwort, M. (2006). Eating Pattern Inventory for
children: A new self-rating questionnaire for preadolescents. Journal of Clinical
Psychology, 62(10), 1259–1273. doi:10.1002/jclp
Schachter, S. (1968). Obesity and eating. Internal and external cues differentially affece the
eating behaviour of obese and normal subjects. Science, 161(3843), 751–756.
Schachter, S., & Gross, L. P. (1968). Manipulated time and eating behavior. Journal of
Personality and Social Psychology, 10(2), 98–106. Retrieved from
http://psycnet.apa.org/journals/psp/10/2/98/
Schembre, S. M., & Geller, K. S. (2011). Psychometric properties and construct validity of
the Weight-Related Eating Questionnaire in a diverse population. Obesity, 19(12),
2336–2344. doi:10.1038/oby.2011.96.Psychometric
Schembre, S. M., Greene, G., & Melanson, K. (2009). Development and validation of a
weight-related eating questionnaire. Eating Behaviors, 10(2), 119–124.
doi:10.1016/j.eatbeh.2009.03.006
Schousboe, K., Willemsen, G., Kyvik, K. O., Mortensen, J., Boomsma, D. I., Cornes, B. K., …
Harris, J. R. (2003). Sex differences in heritability of BMI: A comparative study of
results from twin studies in eight countries. Twin Research, 6(5), 409–421.
doi:10.1375/136905203770326411
Schwarzer, R. (2008). Modeling health behavior change: How to predict and modify the
adoption and maintenance of health behaviors. Applied Psychology, 57(1), 1–29.
244
References
doi:10.1111/j.1464-0597.2007.00325.x
Seger, J., Horn, D. B., Westman, E. C., Scinta, W., Richardson, L. A., Primack, C., … Bays, H. E.
(2013). American Society of Bariatric Physicians obesity algorithm: Adult adiposity
evaluation and treatment. Retrieved June 19, 2016, from
http://www.asbp.org/obesityalgorithm.html
Shamseddeen, H., Zelada Getty, J., Hamdallah, I. N., & Ali, M. R. (2011). Epidemiology and
economic impact of obesity and type 2 diabetes. The Surgical Clinics of North America,
91(6), 1163–72, vii. doi:10.1016/j.suc.2011.08.001
Silventoinen, K., Rokholm, B., Kaprio, J., & Sørensen, T. I. A. (2010). The genetic and
environmental influences on childhood obesity: A systematic review of twin and
adoption studies. International Journal of Obesity, 34(1), 29–40.
doi:10.1038/ijo.2009.177
Sleddens, E. F. C., Kremers, S. P. J., & Thijs, C. (2008). The children’s eating behaviour
questionnaire: Factorial validity and association with Body Mass Index in Dutch
children aged 6-7. The International Journal of Behavioral Nutrition and Physical
Activity, 5(1), 49. doi:10.1186/1479-5868-5-49
Soussignan, R., Schaal, B., Boulanger, V., Gaillet, M., & Jiang, T. (2012). Orofacial reactivity
to the sight and smell of food stimuli. Evidence for anticipatory liking related to food
reward cues in overweight children. Appetite, 58(2), 508–516.
doi:10.1016/j.appet.2011.12.018
Sparks, M. A., & Radnitz, C. L. (2012). Confirmatory factor analysis of the Children’s Eating
Behaviour Questionnaire in a low-income sample. Eating Behaviors, 13(3), 267–70.
doi:10.1016/j.eatbeh.2012.03.002
Spence, J. C., Carson, V., Casey, L., & Boule, N. (2011). Examining behavioural susceptibility
to obesity among Canadian pre-school children: the role of eating behaviours.
International Journal of Pediatric Obesity, 6(2-2), e501–e507.
doi:10.3109/17477166.2010.512087
Stead, M., Craigie, A. M., Macleod, M., McKell, J., Caswell, S., Steele, R. J. C., & Anderson, A.
S. (2015). Why are some people more successful at lifestyle change than others?
Factors associated with successful weight loss in the BeWEL randomised controlled
trial of adults at risk of colorectal cancer. International Journal of Behavioral Nutrition
and Physical Activity, 12, 87. doi:10.1186/s12966-015-0240-2
Stevens, J. P. (2009). Applied multivariate statistics for social sciences. (5th ed.). Hove, Easat
Sussex: Routledge.
Stok, F. M., Verkooijen, K. T., de Ridder, D. T. D., de Wit, J. B. F., & de Vet, E. (2014). How
norms work: Self-identification, attitude, and self-efficacy mediate the relation
between descriptive social norms and vegetable intake. Applied Psychology: Health
and Well-Being, 6(2), 230–250. doi:10.1111/aphw.12026
Stotts, J. L., & Lohse, B. (2007). Reliability of the ecSatter Inventory as a tool to measure
eating competence. Journal of Nutrition Education and Behavior, 39(5), S167–S170.
doi:10.1016/j.jneb.2007.03.091
245
References
Strayhorn, J. M. (2002). Self-control: Theory and research. Journal of the American Academy
of Child and Adolescent Psychiatry, 41(1), 7–16. Retrieved from
Streiner, D. L., & Norman, G. R. (2015). Health measurement scales: A practical guide to
their development and use. (5th ed.). Oxford, UK: Oxford University Press. Retrieved
from https://read.amazon.com/
Stubbs, J., Whybrow, S., Teixeira, P. J., Blundell, J. E., Lawton, C., Westenhoefer, J., … Raats,
M. (2011). Problems in identifying predictors and correlates of weight loss and
maintenance: Implications for weight control therapies based on behaviour change.
Obesity Reviews : An Official Journal of the International Association for the Study of
Obesity, 12(9), 688–708. doi:10.1111/j.1467-789X.2011.00883.x
Stunkard, A. J., & Fox, S. (1971). The relationship of gastric motility and hunger. A summary
of the evidence. Psychosomatic Medicine, 33(2), 123–34. Retrieved from
Stunkard, A. J., Harris, J. R., Pedersen, N. L., & McClearn, G. E. (1990). The Body-Mass Index
of twins who have been reared apart. New England Journal of Medicine, 322(21),
1483–7. Retrieved from
http://www.nejm.org/doi/pdf/10.1056/NEJM199005243222102
Stunkard, A. J., & Messick, S. (1985). The Three-Factor Eating Questionnaire to measure
dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research, 29(1),
71–83. doi:10.1016/0022-3999(85)90010-8
Sumithran, P., & Proietto, J. (2013). The defence of body weight: A physiological basis for
weight regain after weight loss. Clinical Science, 124(4), 231–241.
doi:10.1042/CS20120223
Survey Monkey. (2016). Retrieved May 24, 2016, from https://www.surveymonkey.net/
Svensson, M., Hult, M., van der Mark, M., Grotta, A., Jonasson, J., von Hausswolff-Juhlin, Y.,
… Trolle Lagerros, Y. (2014). The change in eating behaviors in a web-based weight
loss program: A longitudinal analysis of study completers. Journal of Medical Internet
Research, 16(11), e234. doi:10.2196/jmir.3131
Svensson, V., Lundborg, L., Cao, Y. T., Nowicka, P., Marcus, C., & Sobko, T. (2011). Obesity
related eating behaviour patterns in Swedish preschool children and association with
age, gender, relative weight and parental weight - factorial validation of the Children’s
Eating Behaviour Questionnaire. The Iternational Journal of Behavioral Nutrition and
Physical Activity, 8(1), 134. doi:10.1186/1479-5868-8-134
Sweetman, C., Wardle, J., & Cooke, L. (2008). Soft drinks and “desire to drink” in
preschoolers. The International Journal of Behavioral Nutrition and Physical Activity, 5,
60. doi:10.1186/1479-5868-5-60
Syrad, H., Johnson, L., Wardle, J., & Llewellyn, C. H. (2016). Appetitive traits and food intake
patterns in early life. American Journal of Clinical Nutrition, 103(1), 231–235.
doi:10.3945/ajcn.115.117382.
Tanofsky-Kraff, M., Ranzenhofer, L. M., Yanovski, S. Z., Schvey, N. A., Faith, M. S., Gustafson,
246
References
J., & Yanovski, J. A. (2008). Psychometric properties of a new questionnaire to assess

eating in the absence of hunger in children and adolescents. Appetite, 51(1), 148–155.
doi:10.1016/j.appet.2008.01.001
Tanofsky-Kraff, M., Theim, K. R., Yanovski, S. Z., Bassett, A. M., Burns, N. P., Ranzenhofer, L.
M., … Yanovski, J. A. (2007). Validation of the Emotional Eating Scale adapted for use
in children and adolescents (EES-C). Journal of Eating Disorders, 40, 232–240.
doi:10.1002/eat
Tapper, K., & Pothos, E. M. (2010). Development and validation of a Food Preoccupation
Questionnaire. Eating Behaviors, 11(1), 45–53. doi:10.1016/j.eatbeh.2009.09.003
Tavakol, M., & Dennick, R. (2011). Making sense of Cronbach’s alpha. International Journal
of Medical Education, 2, 53–55. doi:10.5116/ijme.4dfb.8dfd
Taylor, C. M., Wernimont, S. M., Northstone, K., & Emmett, P. M. (2015). Picky/fussy eating
in children: Review of definitions, assessment, prevalence and dietary intakes.
Appetite, 95, 349–359. doi:10.1016/j.appet.2015.07.026
Te Morenga, L., Mallard, S., & Mann, J. (2012). Dietary sugars and body weight: Systematic
review and meta-analyses of randomised controlled trials and cohort studies. British
Medical Journal, 346, e7492. doi:10.1136/bmj.e7492
Teixeira, P. J., Going, S. B., Sardinha, L. B., & Lohman, T. G. (2005). A review of psychosocial
pre-treatment predictors of weight control. Obesity Reviews, 6(1), 43–65.
doi:10.1111/j.1467-789X.2005.00166.x
Teixeira, P. J., Palmeira, A. L., Branco, T. L., Martins, S. S., Minderico, C. S., Barata, J. T., …
Sardinha, L. B. (2004). Who will lose weight? A reexamination of predictors of weight
loss in women. The International Journal of Behavioral Nutrition and Physical Activity,
1(1), 12. doi:10.1186/1479-5868-1-12
Teixeira, P. J., Silva, M. N., Coutinho, S. R., Palmeira, A. L., Mata, J., Vieira, P. N., … Sardinha,
L. B. (2010). Mediators of weight loss and weight loss maintenance in middle-aged
women. Obesity, 18(4), 725–735. doi:10.1038/oby.2009.281
Temple, J. L., Legierski, C. M., Giacomelli, A. M., Salvy, S., & Epstein, L. H. (2008).
Overweight children find food more reinforcing and consume more energy than do
nonoverweight children. American Journal of Clinical Nutrition, 87(5), 1121–1127.
The Obesity Society and American College of Cardiology/American Heart Association Task
Force on Practice Guidelines. Based on a systematic review from The Obesity Expert
Panel, 2013. (2014). Executive Summary : Guidelines (2013) for the management of
overweight and obesity in adults. A report of the American College of
Cardiology/American Heart Association Task Force on Practice Guidelines and The
Obesity Society. Obesity, 22(November 2013), 5–39. doi:10.1002/oby.20821
Thomas, J. G., Bond, D. S., Phelan, S., Hill, J. O., & Wing, R. R. (2014). Weight-loss
maintenance for 10 years in the National Weight Control Registry. American Journal of
Preventive Medicine, 46(1), 17–23. doi:10.1016/j.amepre.2013.08.019
Thompson, B. (1951). Exploratory and confirmatory factor analysis: Understanding concepts

and applications (1st c2004 ed.). Washington DC: American Psychological Association.
247
References
Tinsley, H. E., & Tinsley, D. J. (1987). Uses of factor analysis in counseling psychology
research. Journal of Counseling Psychology, 34(4), 414–424. doi:10.1037//0022-
0167.34.4.414
Tong, A., Sainsbury, P., & Craig, J. (2007). Consolidated criteria for reporting qualitative
research (COREQ): A 32-item checklist for interviews and focus groups. International
Journal for Quality in Health Care, 19(6), 349–357. doi:10.1093/intqhc/mzm042
Torres, A., Sarmiento, O. L., Stauber, C., & Zarama, R. (2013). The ciclovia and cicloruta
programs: Promising interventions to promote physical activity and social capital in
Bogota, Colombia. American Journal of Public Health, 103(2), 23–30.
doi:10.2105/AJPH.2012.301142
Treweek, S., Lockhart, P., Pitkethly, M., Cook, J. A., Kjeldstrøm, M., Johansen, M., …
Mitchell, E. D. (2013). Methods to improve recruitment to randomised controlled
trials: Cochrane systematic review and meta-analysis. BMJ Open, 3(2), e002360.
doi:10.1136/bmjopen-2012-002360
Truby, H., Baic, S., DeLooy, A., Fox, K. R., Livingstone, M. B. E., Logan, C. M., … Millward, D. J.
(2006). Randomised controlled trial of four commercial weight loss programmes in the
UK: Initial findings from the BBC “diet trials”. BMJ (Clinical Research Ed.), 332(7553),
1309–1314. doi:10.1136/bmj.38833.411204.80
Truby, H., Hiscutt, R., Herriot, A. M., Stanley, M., Delooy, A., Fox, K. R., … Livingstone, M. B.
E. (2008). Commercial weight loss diets meet nutrient requirements in free living
adults over 8 weeks: A randomised controlled weight loss trial. Nutrition Journal, 7,
25. doi:10.1186/1475-2891-7-25
Tsai, A. G., & Wadden, T. A. (2006). The evolution of very-low-calorie diets: An update and
meta-analysis. Obesity, 14(8), 1283–1293. doi:10.1038/oby.2006.146
Tylka, T. L. (2006). Development and psychometric evaluation of a measure of intuitive

eating. Journal of Counseling Psychology, 53(2), 226–240. doi:10.1037/0022-
0167.53.2.226
Tylka, T. L., & Kroon Van Diest, A. M. (2013). The Intuitive Eating Scale-2: Item refinement
and psychometric evaluation with college women and men. Journal of Counseling
Psychology, 60(1), 137–153. doi:10.1037/a0030893
van Jaarsveld, C. H. M., Boniface, D., Llewellyn, C. H., & Wardle, J. (2014). Appetite and
growth: A longitudinal sibling analysis. JAMA Pediatrics, 168(4), 345–350.
doi:10.1001/jamapediatrics.2013.4951
van Jaarsveld, C. H. M., Llewellyn, C. H., Johnson, L., & Wardle, J. (2011). Prospective
associations between appetitive traits and weight gain in infancy. The American
Journal of Clinical Nutrition, 94(6), 1562–1567.
doi:10.3945/ajcn.111.015818.Individual
van Kleef, E., Otten, K., & van Trijp, H. C. M. (2012). Healthy snacks at the checkout counter:
a lab and field study on the impact of shelf arrangement and assortment structure on
consumer choices. BMC Public Health, 12(1), 1072. doi:10.1186/1471-2458-12-1072
248
References
van Strein, T. (2002). Dutch Eating Behaviour Questionnaire: Manual. Bury St Edmonds, UK:
Thames Valley Test Co.
van Strein, T., Frijters, J. E. R., Bergers, G. P. A., & Defares, P. B. (1986). The Dutch Eating
Behavior Questionnaire (DEBQ) for assessment of restrained, emotional, and external
eating behavior. International Journal of Eating Disorders, 5(2), 295–315.
doi:10.1002/1098-108X(198602)5:2<295::AID-EAT2260050209>3.0.CO;2-T
van Strein, T., Herman, C. P., Engels, R. C. M. E., Larsen, J. K., & van Leeuwe, J. F. J. (2007).
Construct validation of the Restraint Scale in normal-weight and overweight females.
Appetite, 49(1), 109–121. doi:10.1016/j.appet.2007.01.003
van Strein, T., & Oosterveld, P. (2008). The Children’ s DEBQ for assessment of restrained,
emotional, and external eating in 7- to 12 year-old children. International Journal of
Eating Disorders, 41, 72–81. doi:10.1002/eat
van Strein, T., van de Laar, F. A., van Leeuwe, J. F. J., Lucassen, P. L. B. J., van den Hoogen, H.
J. M., Rutten, G. E. H. M., & van Weel, C. (2007). The dieting dilemma in patients with
newly diagnosed type 2 diabetes: Does dietary restraint predict weight gain 4 years
after diagnosis? Health Psychology, 26(1), 105–112. doi:10.1037/0278-6133.26.1.105
Vandenbroeck, P., Goossens, J., & Clemens, M. (2007). Foresight. Tackling obesities: Future
choices - Obesity system atlas. doi:10.1037/e602972011-001
Viana, V., Sinde, S., & Saxton, J. C. (2008). Children’s Eating Behaviour Questionnaire:
associations with BMI in Portuguese children. The British Journal of Nutrition, 100(2),
445–450. doi:10.1017/S0007114508894391
Wadden, T. A., Webb, V. L., Moran, C. H., & Bailer, B. A. (2012). Lifestyle modification for
obesity: New developments in diet, physical activity, and behavior therapy.
Circulation, 125(9), 1157–70. doi:10.1161/CIRCULATIONAHA.111.039453
Walker, K. L., Hadjistavropoulos, T., Gagnon, M. M., & MacNab, Y. C. (2015). Development
and validation of the Hunger Sensitivity Scale (HSS) among university students.
Canadian Journal of Behavioural Science, 47(1), 1–11. doi:10.1037/a0035764
Wardle, J. (1987a). Eating style a validation study of the Dutch Eating Behaviour
Questionnaire in normal subjects and women with eating disorders. Journal of
Psychosomatic Research, 31(2), 161–169. doi:10.1016/0022-3999(87)90072-9
Wardle, J. (1987b). Hunger and satiety: A multidimensional assessment of responses to

caloric loads. Physiology & Behavior, 40(5), 577–582. doi:10.1016/0031-
9384(87)90100-4
Wardle, J., & Beales, S. (1988). Control and loss of control over eating: An experimental
investigation. Journal of Abnormal Psychology, 97(1), 35–40. doi:10.1037/0021-
843X.97.1.35
Wardle, J., & Boniface, D. (2008). Changes in the distributions of body mass index and waist
circumference in English adults, 1993/1994 to 2002/2003. International Journal of
Obesity, 32(3), 527–32. doi:10.1038/sj.ijo.0803740
Wardle, J., & Carnell, S. (2009). Appetite is a heritable phenotype associated with adiposity.
249
References
Annals of Behavioral Medicine, 38, 25–30. doi:10.1007/s12160-009-9116-5
Wardle, J., Carnell, S., Haworth, C. M. A., Farooqi, I. S., O’Rahilly, S., & Plomin, R. (2008).
Obesity associated genetic variation in FTO is associated with diminished satiety. The
Journal of Clinical Endocrinology and Metabolism, 93(9), 3640–3643.
doi:10.1210/jc.2008-0472
Wardle, J., & Gibson, E. L. (2001). Impact of stress on diet: Process and implications. In S.
Stansfield & M. Marmot (Eds.), Stress and heart disease (pp. 35–40). London: British
Medical Journal Books.
Wardle, J., Guthrie, C. A., Sanderson, S., & Rapoport, L. (2001). Development of the
Children’s Eating Behaviour Questionnaire. Journal of Child Psychology and Psychiatry,
42(7), 963–970. doi:10.1111/1469-7610.00792
Wardle, J., & Johnson, F. (2002). Weight and dieting: Examining levels of weight concern in
British adults. International Journal of Obesity, 26(8), 1144–1149.
doi:10.1038/sj.ijo.0802046
Wardle, J., & Johnson, F. (2015). Weight Regulation. In J. D. Wright (Ed.), International
Encyclopedia of the Social & Behavioral Sciences (pp. 468–474). Elsevier Ltd.
doi:10.1016/B978-0-08-097086-8.14139-X
Wardle, J., Liao, L.-M., Rapoport, L., Hillsdon, M., Croker, H., & Edwards, C. (2001). Shape-
Up. A lifestyle programme to manage your weight. London, UK: Weight Concern.
Retrieved from www.weightconcern.com
Wardle, J., Liao, L.-M., Rapoport, L., Hillsdon, M., Croker, H., & Edwards, C. (2013). Shape-
Up. A self-help guide to managing your weight (Reprint). (J. Wardle, P. Chadwick, A.
Chipperfield, H. Croker, N. Gokool, L.-M. Liao, … C. Blanckenberg, Eds.). London, UK:
Weight Concern. Retrieved from www.weightconcern.org.uk
Wardle, J., Marsland, L., Sheikh, Y., Quinn, M., Fedoroff, I., & Ogden, J. (1992). Eating style
and eating behaviour in adolescents. Appetite, 18(3), 167–183.
doi:http://dx.doi.org/10.1016/0195-6663(92)90195-C
Waters, E., Stewart-Brown, S., & Fitzpatrick, R. (2003). Agreement between adolescent self-
report and parent reports of health and well-being: Results of an epidemiological
study. Child: Care, Health and Development, 29(6), 501–509. doi:10.1046/j.1365-
2214.2003.00370.x
Webb, T. L., Joseph, J., Yardley, L., & Michie, S. (2010). Using the internet to promote health
behavior change: a systematic review and meta-analysis of the impact of theoretical
basis, use of behavior change techniques, and mode of delivery on efficacy. Journal of
Medical Internet Research, 12(1), e4. doi:10.2196/jmir.1376
Webber, L., Hill, C., Saxton, J., Van Jaarsveld, C. H. M., & Wardle, J. (2009). Eating behaviour
and weight in children. International Journal of Obesity, 33(1), 1–14.
doi:10.1038/ijo.2008.219.
Wee, C. C., Davis, R. B., Huskey, K. W., Jones, D. B., & Hamel, M. B. (2013). Quality of life
among obese patients seeking weight loss surgery: The importance of obesity-related
social stigma and functional status. Journal of General Internal Medicine, 28, 231–238.
250
References
doi:10.1007/s11606-012-2188-0
Weight Concern. (2016a). Fighting obesity with knowledge... Retrieved June 19, 2016, from
http://www.weightconcern.org.uk/
Weight Concern. (2016b). Shape-Up programme. Retrieved June 19, 2016, from
http://www.weightconcern.org.uk/node/31
Weight Concern. (2016c). The Big Panel. Join the debate. Retrieved June 19, 2016, from
http://weightconcern.org.uk/?q=unsubscribe
Westenhoefer, J. (1991). Dietary restraint and disinhibition: Is restraint a homogeneous

construct? Appetite, 16(1), 45–55. Retrieved from
Westenhoefer, J., Stunkard, a J., & Pudel, V. (1999). Validation of the flexible and rigid
control dimensions of dietary restraint. The International Journal of Eating Disorders,
26(1), 53–64. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10349584
Westerterp-Plantenga, M. S., Lemmens, S. G., & Westerterp, K. R. (2012). Dietary protein –

its role in satiety, energetics, weight loss and health. British Journal of Nutrition,
108(S2), S105–S112. doi:10.1017/S0007114512002589
WHO. (2000). Obesity: preventing and managing the global epidemic. WHO Technical
Report Series 894. Geneva, Switzerland. Retrieved from
http://www.who.int/nutrition/publications/obesity/WHO_TRS_894/en/
WHO. (2014). Global status report on noncommunicable diseases 2014. Geneva,

Switzerland. Retrieved from http://www.who.int/nmh/publications/ncd-status-
report-2014/en/
WHO. (2015). Overweight and obesity fact sheets N°311. Geneva, Switzerland. Retrieved
from http://www.who.int/mediacentre/factsheets/fs311/en/
Wight, D., Wimbush, E., Jepson, R., & Doi, L. (2015). Six steps in quality intervention
development (6SQuID). Journal of Epidemiology and Community Health, 70(5), 520–
525. doi:10.1136/jech-2015-205952
Willett, W. C., Dietz, W. H., & Colditz, G. A. (1999). Guidelines for healthy weight. New
England Journal of Medicine, 341(6), 427–4. doi:10.1056/NEJM199908053410607
Williamson, D. A., Lawson, O. J., Brooks, E. R., Wozniak, P. J., Ryan, D. H., Bray, G. a, &
Duchmann, E. G. (1995). Association of body mass with dietary restraint and
disinhibition. Appetite, 25, 31–41. doi:10.1006/appe.1995.0039
Williamson, D. A., Martin, C. K., York-Crowe, E., Anton, S. D., Redman, L. M., Han, H., &
Ravussin, E. (2007). Measurement of dietary restraint: Validity tests of four
questionnaires. Appetite, 48(2), 183–192. doi:10.1016/j.appet.2006.08.066
Wing, R. R., Crane, M. M., Thomas, J. G., Kumar, R., & Weinberg, B. (2010). Improving
weight loss outcomes of community interventions by incorporating behavioral
strategies. American Journal of Public Health, 100(12), 2513–2519.
doi:10.2105/AJPH.2009.183616
251
References
Wing, R. R., & Phelan, S. (2005). Long-term weight loss maintenance. The American Journal
of Clinical Nutrition, 82(1), 222S–225S. doi:82/1/222S [pii]
Wright, C., & Birks, E. (2000). Risk factors for failure to thrive: A population- based survey.
Child: Care, Health and Development, 26(1), 5–16.
Yaemsiri, S., Slining, M. M., & Agarwal, S. K. (2010). Perceived weight status, overweight
diagnosis, and weight control among US adults: The NHANES 2003 – 2008 study.
252
Appendices
Appendices
Appendix 4.1 Complete electronic search strategy for the systematic review in Chapter 4
1. (eating adj2 behavio$).mp. [mp=title, abstract, original title, name of substance word,
subject heading word, keyword heading word, protocol supplementary concept word, rare
disease supplementary concept word, unique identifier]
2. appetit$.mp.
3. 1 or 2
4. exp Questionnaires/
5. scale$.mp.
6. measure$.mp.
7. instrument$.mp.
8. 4 or 5 or 6 or 7
9. exp Food/
10. eat$.mp.
11. 9 or 10
12. validat$.mp.
13. (factor$ adj2 struct$).mp. [mp=title, abstract, original title, name of substance word,
subject heading word, keyword heading word, protocol supplementary concept word, rare
disease supplementary concept word, unique identifier]
14. reliability.mp.
15. development.mp.
16. adaptation.mp.
17. 12 or 13 or 14 or 15 or 16
18. 3 and 8 and 11 and 17
19. limit 18 to humans
253
Appendices
Appendix 4.2 Standards for educational and psychological testing – does not include all
of the standards relevant to the development of high stake instrument, fairness in testing
and issues of cultural sensitivity (Chapter 4)
Standards
Test development
1. The purpose of the test must be clearly stated, with a definition of the construct
being measured.
2. Specifications about normative or standardization sample must be given.
3. The items and response formats should be reviewed by a panel of experts,
whose qualifications should be specified.
4. Any pilot testing should be described, including characteristics of the sample(s)
tested.
5. The criteria for keeping and rejecting items, whether based on judgement,
classical test theory, or item response theory, must be given.
6. If the items are selected on the basis of empirical relationships (e.g. factor
analysis, item-total correlations) rather than on theoretical grounds, then there
should be at least one cross-validational study to confirm the results. Any
discrepancies between the results of the studies should be documented.
7. Some evidence should be given regarding the content coverage of the scale.
8. If the items are weighted, a rationale (either statistical or theoretical) should be
given for the weights.
9. If scoring involves more than simply adding up the responses, then detailed
instructions should be given, including any training that is required of the raters
or scorers.
10. If a scales is used only for research purposes, this should be clearly stated to the
test taker.
11. If a short form of the test is developed, then two things must be specified: the
procedures or criteria by which items were selected for deletion, and how the
short form’s psychometric properties compare against the original (e.g.
reliability, validity).
12. If, due to research or theory, the definition of the domain has changed
significantly from the time the instrument was originally developed, then the
scale should be modified to reflect this.
Reliability
1. The reliability and standard error measurement (SEM) must be reported for the
total score. If the instrument has sub-scales, then this information must also be
given.
2. When differences scores are interpreted, the reliabilities and SEMs of the
differences cores should be reported.
3. The sample must be described in sufficient detail to allow the readers to
determine if the data apply to their groups.
4. The procedures that were used must be explained (e.g. test-retest interval, any
training given to the raiters, etc.).
5. If the reliability coefficients were adjusted for restriction in range, then both
adjusted and unadjusted values should be reported.
6. If there is reason to believe that the reliability may vary with age, or with
different groups, the reliabilities and SEMs should be reported separately for
these groups, as soon as sufficient data are available.
254
Appendices
Validity
1. Because no test is valid in all people and in all situations, the population for
which the test is appropriate should be clearly stated, including relevant
sociodemographic information.
2. If the scale is to be used in a novel way, validity data must be gathered to
support this new use.
3. A rationale should be given for the domains covered or not covered in the
content validation phase of development.
4. When any phase of development depends on the opinions of experts, raters or
judges, the qualifications of these people should be given, as well as any training
or instructions they may have received.
5. Sufficient details should be reported about any validational studies to allow users
to judge the relevance or the findings to their local conditions.
6. When the validational studies involve relating the new scale to other measures,
the rationale and psychometric properties of the measures must be given.
7. If adjustments have been made for restriction in the range of scores, both
adjusted and unadjusted coefficients should be reported.
Source: (American Educational Research Association et al., 1999; Streiner & Norman, 2015)
255
Appendix 4.3 Validity and reliability measures of questionnaires from the systematic review in Chapter 4 – Extended version
Internal Test-retest
# Reference Convergent/content/criterion validity Discriminant validity Psychometric evaluation
reliability reliability
1 (Arnow et al., α=0.81 (0.72 2-week test Higher levels of all EES sub-scales No measures of psychological adjustment (BDI, 4
1995) to 0.78) retest (r = 0.79, correlated with greater severity of SCL-90-R, RSE) were significantly related to the
p<0.001). binge eating (BES) (p<0.001). EES sub-scales.
No associations between the EES sub-scales and
Significant associations between EES TFEQ-CR.
and Anger/Frustration and Depression
sub-scales and TFEQ-D scale (p<0.05).
Emotional Eating Scale - Adapted for use in Children and Adolescents (EES-C)
2 (Tanofsky-Kraff α (0.83 to 3.4±2.6 months Higher levels of all EES-C sub-scales in The EES-C Anger/Frustration and EES-C 4
et al., 2007) 0.95) ICC (0.59 to 0.74) children with recent LOC eating Unsettled were unrelated to measures of trait
episodes than No LOC (p’s < 0.05). and state anxiety, externalised behaviours.
LOC participants had higher EES-C-Depression
than No LOC participants which reveals sub-
scales may discriminate against measures of
general psychopathology.
3 (Blake et al., α (0.61 to r=0.78 to 0.84 Convergent validity with dietary intake - 3
2013) 0.82) for healthy, measures (p<0.05 to p<0.001)
emotional, and
picky eating
identity types
r=0.66 for meat

eating types

4 (Burgess et al., α (0.73 to (Boggiano et al., Significant associations between the A small but significant association between the 4
2014) 0.91) 2015) PEMS sub-scales and YFAS food PEMS Coping subscale and BIS scores (p<0.01).
r=0.98, p<0.001 dependence and BES scores (p<0.01 for Other sub-scales were not significantly
Appendices
all sub-scales). associated with BIS.
256

5 (Boggiano, α (0.64 to - - - 1
Wenger, Mrug, 0.90)
Burgess, &
Morgan, 2015)
6 (Cappelleri, S1: (Lowe et al., TFEQ-R21: TFEQ-R21: 4
Bushmakin, α (0.81 to 2009) UE: 0.64 (0.54–0.70) EE: 0.51 (0.40- CR: -0.16 (-0.27 to -0.05), suggesting the PFS and
Gerber, Leidy, 0.88) S3: (r =0.77, 0.63) suggesting the PFS and TFEQ-R21 TFEQ-R21 CR were measuring different aspects
Sexton, p<0.001) UE and EE sub-scales were measuring of eating.
Karlsson, et al., S2: similar but distinct aspects of eating.
2009) α (0.82 to
0.91)

7 (Cepeda-Benito S1: FCQ-T: α 3-weeks FCQ-T correlated to TFEQ-D and TFEQ- FCQ-T was largely uncorrelated to TFEQ-CR: r 4
et al., 2000) (0.81 to 0.94) S1: FCQ-T: r (0.72 H: r (0.31 to 0.66) (0.04 to 0.46).
to 0.88) FCQ-S was largely uncorrelated to TFEQ sub-
S2: FCQ-S: α scales: r (0.04 to 0.41), where only Lack of
(0.82 to 0.88) S2: FCQ-S: r (0.40 Control and Desire correlated.
to 0.63) not
stable over time
Brief version of the Food-Cravings Questionnaire-Trait (FCQ-T) (FCQ-T-r)

8 (Meule et al., α=0.94. - FCQ-T-r against the RS (0.32 to 0.78) - 2
2014) with 2/ items (p<0.01)

9 (Nijs et al., S1: G-FCQ-T S2: S2: S2: No correlations between G-FCQ-T and DEBQ- 4
2007) α=0.90 3-weeks G-FCQ-T positively correlated DEBQ-EE RS r=0.04
G-FCQ-S ICC 0.79 r=0.71, p<0.01; DEBQ-ExtE r=0.51,
α=0.92 p<0.01
S2: G-FCQ-T G-FCQ-S positively correlated with

α=0.90 DEBQ-EE r=0.19, p<0.01; and DEBQ-ExtE
G-FCQ-S r=0.30, p<0.01.
α=0.92
Appendices
10 (Dalton et al., α (0.88 to - Craving Control was negatively related - 2
2014) 0.66) with TFEQ-D (p<0.001) and TFEQ-H
257
(p<0.001) and binge eating tendency

(p<0.001)
TFEQ-D and binge eating tendency were
negatively related to Positive Mood
(p<0.001).

11 (Geliebter & α=0.78 and r (0.71 to 0.95) (Nolan et al., 2010) (Nolan et al., 2010) 4
Aversa, 2003) 0.75 for Significant positive correlation between Low correlations of EMAQ positive emotions and
EMAQ-NE the Negative Emotions and Situations situations scores with the DEBQ-EE score.
and EMAQ-PE scores of the EMAQ and the DEBQ-EE
(P<0.0001).
α=0.65 and
0.57 for
EMAQ-NS
and EMAQ-PS
12 (Hawks, Merrill, α (0.75 to 4 weeks MFES Emotional eating subscale was - 3
& Madanat, 0.95) r=(0.55 to 0.77) highly correlated with each of the three
2004) EES sub-scales (p<0.001) and with
TFEQ-D (p<0.001) and TFEQ-H
(p<0.001).
Environmental and Social eating sub-
scales showed similar but weaker
correlation with EES (p<0.001 to
p=0.068) the TFEQ-D (p<0.001 to
p=0.079) and TFEQ-H (p<0.001 to
p=0.004).
Physical eating was significantly
correlated only with the TFEQ-D
(p<0.05).

13 (Hawks et al., α (0.42 to 4-weeks CBDS and total scores for each of the CBDS and scores for IES self-care sub-scale r=– 4
2004) 0.93) r=0.56 and 0.87. four factors were r=–0.84 (p<.0001) for 0.023 (p = 0.659)
intrinsic eating, r=–0.42 (p<.0001) for
extrinsic eating, r=–0.484 (p<0.001) for
anti-dieting,
Appendices
14 (Hulbert- α (0.75) - Significant positive inter-correlations - 2
258
Williams et al., 5/6 subscales with several mindfulness and body

2014) acceptance questionnaires (not detailed
here)
15 (Jiang et al., α=0.92 (0.74 2- weeks Construct validity: Dimensions of - 3
2014) to 0.87) 0.72 positive eating (food responsiveness,
exogenous eating, emotional eating,
and initiative eating) tended to be
positively correlated to each other and
negatively correlated to dimensions of
negative eating (food fussiness, eating
habit, satiety responsiveness) (p<0.05
to p<0.01).
16 (Loewen & α=0.80 (0.71 29.9 day mean Self-reported FSQ predicted willingness - 3
Pliner, 2000) and 0.73) r=0.64 (0.64 and to try new foods under laboratory
0.56) conditions, and better than parent
report of their child’s neophobia.

17 (Lohse et al., α (0.65 to (Stotts & Lohse, Eating-competent persons (i.e. ecSI - 3
2007) 0.84) 2007) score ≥32) exhibited lower feelings of
2- to 6-week; TFEQ-CR, TFEQ-D, and TFEQ-H
r=0.68 (0.52 to (p≤0.001).
0.70)

18 (Ogden et al., α (0.6 to 0.9) - - - 1
2012)

19 (Pliner & α (0.88) 2 to 4 weeks: Correlations between FNS and the GNS FNS scores non-significant when correlated 4
Hobden, 1992) r(38)=0.91 and for the two samples were r(128) = 0.54, against measures of composite anxiety
r(31)=0.87, p<0.01 and r(71)=0.62, p<0.01. r(28)=0.26, as well as non-significant for low or
p<0.01. high fear anxiety conditions (fear
manipulations): r(24)= 0.18 and r(25)=0.26.
15-weeks:
r(59)=0.82,
p<0.01
Appendices
20 (Pliner, 1994) - - High correlations between behavioural - 2
259
measures of food neophobia (state) and

paper and pencil measures of FNS-C
(trait)
21 (Laureati et al., α=0.71 t-test ICFNS scores were significantly and - 3
2015) comparisons negatively correlated with willingness
(p<0.05) to taste and liking of unfamiliar food.
22 (Popkess- S1: α (0.74 to - S4: Significant correlation between OTS - 2
Vawter et al., 0.88) and BULIT (r=0.37, p<0.01)
2000) S2: α (0.69 to
0.87)
S3: α (0.74 to
0.93)
S4: α (0.70 to
0.92)
23 (Rollins et al., α=0.86 (0.70 - Only criterion validity of the EESQ sub- - 2
2014) and 0.81) scales with the food frequency and
eating behaviour measures, stratified
by gender (p<0.05, p<0.01, p<0.001).

24 (Schacht et al., α (0.72 to - - - 1
2006) .93)

25 (Stunkard & S3: Dieters: (Ganley, 1988) (Gormally et al., 1982) Discriminant measures were shown between 4
Messick, 1985) α (0.79 to One month: Binge severity, quantified by a scale subgroups of dieters and free eaters in S3
0.84) Free α (0.80 to 0.93) devised for that purpose correlated (p<0.001).
eaters: α with TFEQ-D (r=0.61, p<0.001) but not
(0.84 to 0.92) with TFEQ-CR (r= -0.14, NS). Binge (Gormally et al., 1982)
Combined severity correlated with TFEQ-H (r=0.54, Binge severity, did not correlated with TFEQ-CR
sample: α p < 0.001). (r= -0.14, NS)
(0.85 to 0.93)
Three Factor Eating Questionnaire revised version TFEQ-R18

26 (Karlsson et al., S1: α (0.76 to - 12/21 items passed convergent validity. Item discriminant validity revealed separating 3
2000) 0.85) TFEQ-D and TFEQ-H was a problem
Appendices
S2: α (0.77 to
0.85)
260
Three Factor Eating Questionnaire revised version TFEQ-R21_TFEQ-R18-V2

27 (Cappelleri, S1: α (0.70 to - - - 1
Bushmakin, 0.92)
Gerber, Leidy, S2: α (0.78 to
Sexton, Lowe, 0.94)
et al., 2009)

28 (Tanofsky-Kraff α (0.80 to α: 0.65 to 0.70, Good convergent validity with The EAH-C Boredom/Fatigue scale discriminated 4
et al., 2008) 0.88) p’s <0.01 emotional eating and loss of control from depressive symptoms (r=0.13, p=0.20), and
episodes (p<0.01) (against measures of the EAH-C External scale was not associated with
depression and anxiety). State anxiety (r=0.05, p=0.67) and depressive
symptoms (r=0.12, p=0.25).

29 (Tylka, 2006) S1: α=0.89 S4: 3-week r=0.90 (Avalos & Tylka, 2006) IES total scores negligibly related to optimism 4
(0.72 to 0.89) (0.74 to 0.88) IES total scores were moderately to and unrelated to proactive coping. Impression
S2: α=0.85 strongly related to self-esteem and management was not related to the total IES,
(0.85 to 0.87) satisfaction with life and moderately the Unconditional Permission to Eat subscale, or
related to optimism and proactive the Eating for Physical Rather Than Emotional
coping (p<0.001). Reasons subscale.
Unconditional Permission to Eat
subscale was strongly related in a
negative direction to eating disorder
symptomatology (EAT-26) (p<0.001).
The Eating for Physical Rather Than
Emotional Reasons subscale was
moderately to strongly related to and
satisfaction with life (p<0.001).
Intuitive Eating Scale 2 (IES-2)

30 (Tylka & Kroon S1: F: α=0.89 3 weeks Correlations were r=0.87 for women IES-2 scores were unrelated or negligibly related 4
Van Diest, (0.81 to 0.93); r=0.88 among and r=0.91 for men between the to social desirability for women (p=0.4) and men
2013) M: α=0.87 women and original IES and the IES-2 total scores. (p=0.1).
(0.82 to 0.92) r=0.92 among IES-2 total scores were positively
S2: F: α=0.88 men for the IES-2 related to body appreciation, self-
(0.81 to 0.93); total score. esteem, and satisfaction with life
M: α=0.89 (p<0.01). IES was inversely related to
(0.83 to 0.92) eating disorder symptomatology, poor
Appendices
S3: F: α=0.85 interoceptive awareness, body
(0.77 to 0.92); surveillance, body shame, and
261
M: α=0.88 internalization of media appearance

(0.82 to 0.92) ideals (p<0.01).
The Dutch Eating Behaviour Questionnaire (DEBQ)
31 (van Strein et α (0.80 to α=0.94 (0.65 to Construct validity: Positive correlations Existent data but unavailable as only included in 4
al., 1986) 0.95) 0.84) (Banasiak et between emotional and external eating a publication when DEBQ is purchased for use
al., 2001) and weak relationships between (van Strein, 2002).
restraint and external eating. Similar
results have been replicated in (Cebolla
et al., 2014; J Wardle, 1987a).
The Children’s Dutch Eating Behaviour Questionnaire (DEBQ - C)
32 (van Strein & α (0.73 to (Baños et al., Construct validity: - 3
Oosterveld, 0.82) 2011) (1 month) In both sexes DEBQ-C-EE and DEBQ-C-
2008) DEBQ-C-EE ExtE was significantly interrelated
α=0.39 (0.22- (p<0.01), but DEBQ-C-R was not
0.54), DEBQ-C-R associated with either DEBQ-C-EE and
α=0.71 (0.61- DEBQ-C-ExtE (controlling for BMI and
0.79), DEBQ-C- age).
ExtE α=0.64
(0.52-0.74).
33 (Braet & van (Caccialanza - Significant relationship between DEBQ- - 2
Strein, 1997) et al., 2004) P-EE and DEBQ-P-ExtE and various
α (0.79 to nutritional parameters (p<0.01 to
0.86) p<0.001).
34 (Walker et al., S1: α=0.95 S2: One month, HSS was significantly associated with Absence of significant correlations with general 4
2015) S2: α=0.90 r=0.81 (p<0.001) TFEQ-H (p<0.05), TFEQ-D (p<0.001). anxiety, depression and anxiety sensitivity.
35 (Wardle, α (0.74 to t (0.52 [EOE] & (Carnell & Wardle, 2007) Behavioural (Loh et al., 2013) See details below. 4
Guthrie, et al., 0.91) 0.64 [EUE] to validation:
2001) 0.87) Higher SR was associated with lower
intake in the EWH test, better average
caloric compensation, slower eating
and lower average total energy intake.
Higher scores on FR were associated
with faster eating rate and greater total
energy intake. Higher scores on EF were
associated with greater EWH intake,
Appendices
faster eating rate and greater total
energy intake and a marginally
262
significant association between higher

EF and poorer average caloric
compensation.
36 (Llewellyn, van α (0.73 to - Construct validity: - 2
Jaarsveld, et al., 0.81) ‘Satiety responsiveness’ and ‘slowness
2011) in eating’ were positively correlated and
the size of the correlation was only
slightly smaller than in older children
(0.52–0.67) (Wardle et al., 2001).
Self-report measure of the CEBQ for 13-year-old adolescents (CEBQ-self-report)
37 (Loh et al., 0.48 to 0.76 0.72 to 0.90 - The AVE values were greater than the R-squared 2
2013) values between the constructs between Phase 1
and Phase 2 models, indicating sufficient
discriminant validity.
38 (Westenhoefer, Rigid control (Westenhoefer et (Westenhoefer et al., 1999) Discriminant analysis in a subgroup of 4
1991) α=0.77 al., 1999) Increased ‘rigid control’ is associated moderately highly restrained eaters with either
Flexible Scales measured with increasing disinhibition (p<0.001). low or high disinhibition (n= 1759) revealed
control at different time Increasing ‘flexible control’ is associated different sets of restraint behaviours and
α=0.79 points with decreasing disinhibition (p<0.001). cognitions differentiate between high and low
disinhibition.
S1, S2, S3, etc. = Study 1, Study 2, Study 3

AVE: Average variance extracted; BDI: Beck Depression Inventory; BIS: Behavioural Inhibition Scores; BULIT: Bulimia test; CBDS: Cognitive Behavioural Dieting Scale;
EE: Emotional Eating; ExtE: External Eating; CR: Cognitive Restraint; DI: Dysregulation Inventory; GNS: General Neophobia Scale, LOC: Loss of control; MOE: Meanings
of Eating Questionnaire; NCOG: Non-clinical overweight group; No LOC: No loss of control; NWG: Normal weight group;
RSE: Rosenberg self-esteem scale; SCL-90-R: Symptom checklist; UE: Uncontrolled eating.
BEBQ: Baby Eating Behaviour Questionnaire; CEBQ: Child Eating Behaviour Questionnaire; CEBQ-self-report: Self-report measure of the CEBQ; CPEBQ: Chinese Pre-
schoolers’ Eating Behaviour Questionnaire; CoEQ: Control of Eating Questionnaire; DEBQ: Dutch Eating Behaviour Questionnaire; DEBQ-C: Children’s Dutch Eating
Behaviour Questionnaire; DEBQ-P: Dutch Eating Behaviour Questionnaire parent version; EAH-C: Eating in the Absence of Hunger; ecSI: ecSatter Inventory; EES:
Emotional Eating Scale; EESQ: Eating in Emotional Situations Questionnaire; EES-C: Emotional Eating Scale; EITI: Eating Identity Type Inventory; EMAQ: Emotional
Appetite Questionnaire; EPI-C: Eating Pattern Inventory for Children; FCQ-S and FCQ-T: State and Trait Food-Cravings Questionnaires; FCQ-T-r and FCQ-T-r: Brief
version of the Food Craving Questionnaire-Trait; FSQ: food Situations Questionnaire; FNS: Food Neophobia Scale; FNS-C: Food Neophobia Scale for children; G-FCQ-T
and G-FCQ-S: General index of food craving; HSS: Hunger Sensitivity Scale; ICFNS: Italian Food Neophobia Scale for children; IES: Intuitive Eating Scale; IES-2: Intuitive
Eating Scale-2; IES-H: Intuitive Eating Scale-H; K-PEMS: Palatable Eating Motives Scale for kids; MES: Mindful Eating Scale; MOF: Meaning of Food Questionnaire; OTS:
Appendices
Overeating Tension Scales; PEMS: Palatable Eating Motives Scale; PFS: Power of Food Scale; MFES: Motivation for Eating Scale; TFEQ: Three Factor Eating
Questionnaire; TFEQ-R18: Three Factor Eating Questionnaire revised version; Three Factor Eating Questionnaire revised version TFEQ-R21 _TFEQ-R18-V2.
263
Appendices
Appendix 5.1 Published paper
264
Appendices
265
Appendices
266
Appendices
267
Appendices
268
Appendices
269
Appendices
270
Appendices
Appendix 5.2 Conference presentations
Hunot C., Beeken R.J., Croker H., Wardle J. Development of the ‘adult eating behaviour
questionnaire’ for appetitive trait measurement. Obesity Facts 2015;8(suppl 1):89-90.
Poster at the European Conference on Obesity (ECO 2015), Prague, Czech Republic.
Hunot C., Beeken R.J., Croker H., Klienman N., Wardle J. Associations between appetitive
traits and weight in adults in Britain. Obesity Facts 2015;8(suppl 1):85-86. Poster at the
European Conference on Obesity (ECO 2015), Prague, Czech Republic.
Hunot C., Beeken R.J., Croker H., Wardle J. Development of the ‘adult eating behaviour
questionnaire’ for appetitive trait measurement. Oral presentation at the XIII Symposium
of Mexican Students and Studies. July 24th 2015. University College London, UK.
271
Appendices
Appendix 5.3 The Child Eating Behaviour Questionnaire (CEBQ)
272
Appendices
273
Appendices
http://www.ucl.ac.uk/hbrc/resources/resources_eb
274
Appendices
Appendix 5.4 Weight Concern ‘Shape-Up’ manual ‘hunger’ or ‘craving’ questions
Source: (Wardle, Liao, et al., 2001)
275
Appendices
Appendix 5.5 ‘Adult Eating Behaviour Questionnaire’ used for piloting (49-item) in Study
2, Chapter 5
276
Appendices
277
Appendices
278
Appendices
279
Appendices
280
Appendices
281
Appendices
282
Appendices
283
Appendices
Appendix 5.6 Illustrative example of themes obtained from adult piloting (Study 2,
Chapter 5)
Regarding response option: ‘never’, ‘rarely’, ‘sometimes’, ‘often’ and ‘always’.

“response options do not always fit the question” (n=1)
“phrasing with the scale is ambiguous and I could not provide meaningful answers as a
result” (n=1)
“Some of the questions don't really work with the answers given” (n=1)
“The 'I often feel' questions seem strange given that the response options are frequency
based” (n=1)
“I think the response options would make more sense if they were agree to disagree rather
than never to always” (n=1)
“Maybe one needs to differentiate questions between frequency (appropriate for some
questions), and a True/False scale (appropriate for others)” (n=1)
Regarding item: Given the choice, I would always have food in my mouth
“sounds a bit odd”(n=4)
“over the top” (n=1)
“the questions 'given the choice I would always be eating/have food in my mouth' implies
there is a barrier to eating” (n=1)
Regarding item: I am interested about food
“the question [I am interested about food] sound a bit strange, it is weird” (n=3).
284
Appendices
Appendix 5.7 Adult Eating Behaviour Questionnaire (47-items) (Study 2, Sample 1)
285
Appendices
286
Appendices
287
Appendices
288
Appendices
289
Appendices
290
Appendices
291
Appendices
292
Appendices
293
Appendices
Appendix 5.8 Ethical Approval, Study 2, Chapter 5
294
Appendices
295
Appendices
Appendix 5.9 AEBQ items compared to the original CEBQ items
Item
AEBQ AEBQ item CEBQ item
source
I often feel so hungry that I have
New
to eat something right away
I often notice my stomach
New
rumbling
H New If I miss a meal I get irritable
If my meals are delayed I get light-
New
headed
New I often feel hungry
I often feel hungry when I am with
New
someone who is eating
New I am always thinking about food
When I see or smell food that I
New
like, it makes me want to eat
Given the choice, I would eat most Given the choice, my child would
CEBQ (FR)
of the time eat most of the time
Even if my child is full up s/he
FR
NI - finds room to eat his/her favourite
food
If given the chance, my child
NI - would always have food in his/her
mouth
NI - My child is always asking for food
If allowed to, my child would eat
NI -
too much
CEBQ
I eat more when I'm annoyed My child eats more when annoyed
(EOE)
CEBQ
I eat more when I'm worried My child eats more when worried
(EOE)
New I eat more when I'm upset NI
EOE
CEBQ
I eat more when I´m anxious My child eats more when anxious
(EOE)
New I eat more when I'm angry NI
My child eats more when s/he has
NI -
nothing else to do
CEBQ (EF) I love food My child loves food
My child looks forward to
CEBQ (EF) I look forward to mealtimes
EF mealtimes
CEBQ (EF) I enjoy eating My child enjoys eating
NI - My child is interested in food
I often leave food on my plate at My child leaves food on his/her
CEBQ (SR)
the end of a meal plate at the end of a meal
I often get full before my meal is My child gets full before his/her
CEBQ (SR)
finished meal is finished
SR CEBQ (SR) I get full up easily My child gets full up easily
I cannot eat a meal if I have had a My child cannot eat a meal if s/he
CEBQ (SR)
snack just before has had a snack just before
NI - My child has a big appetite*
296
Appendices
Item
AEBQ AEBQ item CEBQ item
source
New I eat less when I'm worried
CEBQ
I eat less when I'm angry My child eats less when angry
(EUE)
CEBQ
I eat less when I'm upset My child eats less when upset
(EUE)
New I eat less when I'm annoyed
EUE
New I eat less when I'm anxious
My child eats more when she is
NI -
happy
My child eats less when s/he is
NI - tired
CEBQ (FF) I refuse new foods at first My child refuses new foods at first
My child is difficult to please with
NI -
meals
I often decide that I don’t like a My child decides that s/he doesn’t
CEBQ (FF)
food, before tasting it like a food, even without tasting it
FF My child enjoys tasting new
CEBQ (FF) I enjoy tasting new foods*
foods*
I am interested in tasting food I My child is interested in tasting
CEBQ (FF)
haven't tasted before* food s/he hasn’t tasted before*
My child enjoys a wide variety of
CEBQ (FF) I enjoy a wide variety of foods*
foods*
CEBQ (SE) I eat slowly My child eats slowly
My child takes more than 30
CEBQ (SE) I am often last at finishing a meal
minutes to finish a meal
SE I eat more and more slowly during My child eats more and more
CEBQ (SE)
the course of a meal slowly during the course of a meal
My child finishes his/her meal
CEBQ (SE) I often finish my meal (s) quickly*
quickly*
H, ‘hunger’; FR, ‘food responsiveness’; EOE, ‘emotional over-eating’; EF, ‘enjoyment of food’; SR,
‘satiety responsiveness’; EUE, ‘emotional under-eating’; FF, ‘food fussiness’; SE, ‘slowness in eating’.
297
Appendices
Appendix 5.10 Adult Eating Behaviour Questionnaire with scoring system
298
Appendices
299
Appendices
300
Appendices
https://www.ucl.ac.uk/hbrc/resources/resources_eb/AEBQ
301
Appendices
Appendix 6.1 Relevant parts of the Self-Regulation of Eating Behaviour Questionnaire

which contained the Adult Eating Behaviour Questionnaire items (Study 3, Sample 2)
302
Appendices
303
Appendices
304
Appendices
305
Appendices
306
Appendices
307
Appendices
308
Appendices
309
Appendices
310
Appendices
Appendix 6.2 Ethical approval, Study 3, Chapter 6
311
Appendices
312
Appendices
Appendix 7.1 Feasibility questions obtained from an on-line panel who completed the
AEBQ (n=711)a (Sample 2)
Feasibility questions regarding the interest of participants in

Frequency
participating in an intervention involving feedback on their
n (%)
AEBQ responses and receiving tailored appetitive trait feedback
Would you be interested in receiving feedback on your appetitive traits
(i.e. styles of eating that could make you gain or lose weight) and tips n=954b
on how to manage them accordingly?
Yes 558 (58.5%)
No 243 (25.5%)
Maybe 153 (16.0%)
What format would you like to receive this information in?
In person 15 (2.1%)
Via e-mail 611 (85.9%)
Via phone 8 (1.1%)
On-line 77 (10.8%)
Do you think that knowing about your appetitive traits would change
how you eat?
Yes 362 (50.9%)
No 45 (6.3%)
Maybe 304 (42.8%)
Would you be interested in taking part in a study looking at the effect
of giving people feedback on their appetitive traits?
Very likely to take part 254 (35.7%)
Likely to take part 186 (26.2%)
Somewhat likely to take part 185 (26.0%)
Probably would not take part 86 (12.1%)
If the study on appetitive trait feedback took place over eight weeks,
how often would you be interested in receiving input/tips as feedback?
Daily 111 (15.6%)
Weekly 451 (63.4%)
Fortnightly 65 (6.8%)
Monthly 50 (5.2%)
Never 34 (4.8%)
Is there any information you think would be particularly useful?
(Choose as many options as you like)
Tips on becoming aware of how hungry you are 273/681 (28.6%)
Healthy food options 444/628 (46.5%)
Tips on how to like healthy foods more 295/659 (30.9%)
Tips for managing emotional eating 283/671 (29.7%)
Tips on how to control how much you eat when around 362/592 (37.9%)
tempting food
Tips on resisting eating 326/628 (34.2%)
Tips on eating self-awareness (do you know when you are
373/581 (39.1%)
hungry?)
a
Data was analysed only for those participants who replied they ‘yes’ or ‘maybe’ would be
interested in participating in an intervention involving feedback on their AEBQ responses and
receiving tailored appetitive trait feedback (i.e. n=711).
b
Initial total data collected n=954.
313
Appendices
Appendix 7.2 Individualised appetitive trait tip feedback for high 'food responsiveness'
and high ‘emotional over-eating’, low 'satiety responsiveness' and fast eating (low
'slowness in eating')
314
Appendices
315
Appendices
316
Appendices
317
Appendices
318
Appendices
319
Appendices
320
Appendices
Appendix 7.3 Individualised appetitive trait tip feedback for high 'food responsiveness',
low 'satiety responsiveness' and fast eating (low 'slowness in eating')
321
Appendices
322
322
Appendices
323
323
Appendices
324
324
Appendices
325
325
Appendices
326
326
Appendices
Appendix 7.4. Initial Survey Monkey questionnaire sent to members of the ‘Big Panel’ in
Study 4, Chapter 7
327
Appendices
328
Appendices
329
Appendices
330
Appendices
331
Appendices
332
Appendices
333
Appendices
Appendix 7.5 Survey Monkey questionnaire sent to potential ‘Big Panel’ members after
a previous first contact, to assess inclusion criteria for participation in Study 4, Chapter 7
334
Appendices
335
Appendices
336
Appendices
337
Appendices
338
Appendices
Appendix 7.6 Weekly follow-up questionnaires (WFQ) (Study 4, Chapter 7)
339
Appendices
340
Appendices
341
Appendices
342
Appendices
343
Appendices
344
Appendices
Appendix 7.7 Consent form for participation in the Appetitive Trait Tailored Intervention
(ATTI) (Study 4, Chapter 7)
345
Appendices
Appendix 7.8 Ethical approval, Study 4, Chapter 7
346
Appendices
347
Appendices
Appendix 7.9 Number of tips given to each participant for high ‘food responsiveness’ and
‘emotional over-eating’ and low ‘satiety responsiveness’ and ‘slowness in eating’ scores
(n=53) (Study 4, Chapter 7)
No. of tips per

Participant ID FR Mean EOE Mean SR Mean SE Mean
participant (n)
1 3.75 3.2 2.75 1.25 4
2 3.75 2.6 2.5 3.75 2
3 3.5 2.6 3 2.25 2
4 4 3.6 2.25 3 3
5 4 3.8 2.5 2.5 4
6 4.25 3.4 2.25 2.75 4
7 3.5 2.8 3.75 2.75 2
8 3.25 3 2.75 2.5 3
9 3.75 3.4 2.75 1.75 4
10 4.25 3.4 3.25 4 2
11 4 2.6 2.5 3 2
12 4 3.4 2.5 2 4
13 4.25 3.4 2.25 1.5 4
14 3.25 2.6 2 2.25 3
15 4.25 3 2.5 2 3
16 3.75 3.2 2 1.75 4
17 3.25 2.8 3 3.5 2
18 3.25 2.8 2.75 1.75 3
19 3.25 2.4 3 3 1
20 4 3.4 2.75 2 4
21 3.5 3.6 3.5 2.5 3
22 3.5 2.4 3.5 2.5 2
23 4.5 3.2 3 3.5 2
24 5 3.4 3.5 4 2
25 4.75 3.2 3 1.5 3
26 3.75 3 3 2.25 2
27 4.25 2 2.25 2.25 3
28 4 2.6 2 4 2
29 2.5 3.4 3.5 2.25 2
30 3 2.8 2.5 2.75 2
31 4.5 3.8 3.5 3 2
32 4 3.2 2.25 1.5 4
33 4.5 3.2 2.5 1.25 4
34 3.25 3.8 2 2.25 4
35 3.5 3.4 3.75 4 2
36 3.5 3.2 3.25 2.75 3
37 4.5 3.8 2.25 1.75 4
38 3.75 3 2.5 2 3
39 3 2.6 3 2.5 1
40 4.75 3.2 2 1 4
41 3.5 3.4 2.5 2.25 4
42 3.5 2.8 3 2.75 2
348
Appendices
No. of tips per

Participant ID FR Mean EOE Mean SR Mean SE Mean
participant (n)
43 3.75 2.2 3.25 2.75 2
44 4.25 3.6 2 1 4
45 3.5 3.2 3 2.75 3
46 3.5 3.4 2.75 2.25 4
47 4.75 4.2 3 1 3
48 4 2.8 3.25 3.5 1
49 3.75 3.4 3.25 2.5 3
50 3.75 2.8 1.5 2.2 3
51 4 2.2 3.75 2.75 2
52 4.25 3.2 3.5 1.75 3
53 5 3.2 2 1.25 4
Tips given to 50 31 29 42
participants n(%) (94.3%) (58.5%) (54.7%) (79.2%)
High scores
Low scores
FR=Food responsiveness; EOE=Emotional over-eating; SR=Satiety responsiveness; SE=Slowness in
eating.
349
Appendices
Appendix 7.10 Initial BMI, final BMI, initial weight, final weight and change in weight for
each participant in the appetitive trait intervention (n=53) (Study 4, Chapter 7)
Initial Weight
Participant Final BMI Weight
BMI I-Weight F-Weight loss (%)
ID Kg/m2 change (kg)
Kg/m2 (kg) (kg)
1 40.1 - 127 - - -
2 39.9 38.1 92.1 88 -4.1 10.7
3 34.0 33.3 83.9 82 -1.9 5.0
4 26.4 26.7 78.9 79.8 0.9 -2.3
5 32.3 33.0 91.2 93 1.8 -4.7
6 39.0 38.6 110 109 -1 2.6
7 28.5 - 70.3 - - -
8 27.3 27.0 74.4 73.5 -0.9 2.3
9 26.6 26.6 70.8 70.8 0 0.0
10 43.9 43.0 116.6 114.3 -2.3 6.0
11 38.2 37.9 101.6 100.7 -0.9 2.3
12 56.6 - 159.7 - - -
13 27.7 27.7 73.5 73.5 0 0.0
14 39.8 39.8 108.4 108.4 0 0.0
15 28.6 - 85.7 - - -
16 33.8 33.8 112 112 0 0.0
17 25.3 24.7 74.9 73 -1.9 5.0
18 36.6 36.1 90.3 88.9 -1.4 3.7
19 27.1 - 69.4 - - -
20 38.3 - 88.5 - - -
21 29.4 - 83 - - -
22 45.0 - 127 - - -
23 46.2 47.1 113.9 116.1 2.2 -5.7
24 36.1 36.1 84.4 84.4 0 0.0
25 28.6 28.6 82.6 82.6 0 0.0
26 43.7 43.5 116.1 115.7 -0.4 1.0
27 31.6 - 78 - - -
28 27.0 25.6 71.7 68 -3.7 9.7
29 36.3 - 102.5 - - -
30 31.8 31.6 85.5 85 -0.5 1.3
31 46.3 - 126 - - -
32 29.6 - 83.5 - - -
33 36.0 - 95.7 - - -
34 32.3 32.7 88 88.9 0.9 -2.3
35 29.4 29.0 68 67.1 -0.9 2.3
36 51.2 49.9 156.9 152.9 -4 10.4
37 30.2 29.8 74.4 73.5 -0.9 2.3
38 26.5 25.1 71.2 67.6 -3.6 9.4
39 29.7 30.4 85.7 88 2.3 -6.0
40 25.9 - 73 - - -
41 42.2 41.7 109.3 108 -1.3 3.4
42 33.5 - 88.9 - - -
350
Appendices
Initial Weight
Participant Final BMI Weight
BMI I-Weight F-Weight loss (%)
ID Kg/m2 change (kg)
Kg/m2 (kg) (kg)
43 50.2 - 133.4 - - -
44 56.0 - 148.8 - - -
45 30.7 - 79.5 - - -
46 34.2 33.8 96.6 95.3 -1.3 3.4
47 28.5 27.0 77.6 73.5 -4.1 10.7
48 41.7 37.6 113.4 102.5 -10.9 28.4
49 27.1 27.7 70.3 71.7 1.4 -3.7
50 49.6 - 143.3 - - -
51 39.1 - 113 - - -
52 35.8 - 94 - - -
53 38.8 38.1 103 101.1 -1.9 5.0
Lost weight
Gained weight
Stayed the same weight
F: Final; I: Initial
351
Appendices
Appendix 8.1 Semi-structured interview guide for partitipants of the ‘Appetitive Trait
Tailored Intervention’ (ATTI) (Study 5, Chapter 8)
352
Appendices
353
Appendices
354
Appendices
355
Appendices
356
Appendices
Appendix 8.2 Consolidated criteria for reporting qualitative research (COREQ); 32 item
checklist (Study 5, Chapter 8)
No Item Guide questions/description

Domain 1: Research team and reflexivity
Personal characteristics
1 Interviewer/facilitator Which author conducted the interviews?
2 Credentials What were the researcher’s credentials?
3 Occupation What was their occupation at the time of the study?
4 Gender Was the researcher male or female?
5 Experience and training What experience or training did the researcher have?
Relationship with participants
6 Relationship established Was a relationship established prior to study
commencement?
7 Participant knowledge of the What did the participants know about the researcher?
interviewer
8 Interviewer characteristics What characteristics were reported about the
interviewer?
Domain 2: Study design
Theoretical framework
9 Methodological orientation and What methodological orientation was stated to
theory underpin the study?
Participant selection
10 Sampling How were participants selected?
11 Method of approach How were participants approached?
12 Sample size How many participants were in the study?
13 Non-participation How many people refused to participate or dropped
out? Reasons?
Setting
14 Setting of data collection Where was the data collected?
15 Presence of non-participants Was anyone else present besides the participants and
researchers?
16 Description of the sample What are the important characteristics of the sample?
Data collection
17 Interview guide Were questions, prompts, guides provided by the
authors? Was it pilot tested?
18 Repeat interviews Were repeat interviews carried out? If yes, how
many?
19 Audio/visual recording Did the researcher use audio or visual recording to
collect the data?
20 Field notes Were field notes made during and/or after the
interview?
21 Duration What was the duration of the interviews?
22 Data saturation Was data saturation discussed?
23 Transcripts returned Were transcripts returned to participants for
comment and/or corrections?
Domain 3: Analysis and findings
Data analysis
24 Number of data coders How many data coders coded the data?
25 Description of the coding tree Did authors provide a description of the coding tree?
26 Derivation of themes Were themes identified in advance or derived from
the data?
27 Software What software, if applicable, was used to manage the
357
Appendices
data?
28 Participant checking Did participants provide feedback on the findings?
Reporting
29 Quotations presented Were participant quotations presented to illustrate
the themes? Was each quotation identified?
30 Data and findings consistent Was there consistency between the data presented
and the findings?
31 Clarity of major themes Were major themes clearly presented in the findings?
32 Clarity of minor themes Is there a description of diverse cases or discussion of
minor themes?
Source: Tong, Sainsbury, & Craig, 2007
358
Appendix 8.3 Appetitive traits tips given, Initial BMI, final BMI, initial weight, final weight and change in weight for each participant
interviewed after the appetitive trait intervention (n=21) (Study 5, Chapter 8)
Initial BMI Final BMI I-Weight F-Weight

Participant ID FR Mean EOE Mean SR Mean SE Mean Weight change (kg)
Kg/m2 Kg/m2 (kg) (kg)
2 3.75 2.6 2.5 3.75 39.9 38.1 92.1 88 4.1
4 4 3.6 2.25 3 26.4 26.7 78.9 79.8 -0.9
6 4.25 3.4 2.25 2.75 39.0 38.6 110 109 1
8 3.25 3 2.75 2.5 27.3 27.0 74.4 73.5 0.9
9 3.75 3.4 2.75 1.75 26.6 26.6 70.8 70.8 0
10 4.25 3.4 3.25 4 43.9 43.0 116.6 114.3 2.3
13 4.25 3.4 2.25 1.5 27.7 27.7 73.5 73.5 0
16 3.75 3.2 2 1.75 33.8 33.8 112 112 0
17 3.25 2.8 3 3.5 25.3 24.7 74.9 73 1.9
18 3.25 2.8 2.75 1.75 36.6 36.1 90.3 88.9 1.4
23 4.5 3.2 3 3.5 46.2 47.1 113.9 116.1 -2.2
25 4.75 3.2 3 1.5 28.6 28.6 82.6 82.6 0
26 3.75 3 3 2.25 43.7 43.5 116.1 115.7 0.4
28 4 2.6 2 4 27.0 25.6 71.7 68 3.7
30 3 2.8 2.5 2.75 31.8 31.6 85.5 85 0.5
35 3.5 3.4 3.75 4 29.4 29.0 68 67.1 0.9
37 4.5 3.8 2.25 1.75 30.2 29.8 74.4 73.5 0.9
41 3.5 3.4 2.5 2.25 42.2 41.7 109.3 108 1.3
46 3.5 3.4 2.75 2.25 34.2 33.8 96.6 95.3 1.3
47 4.75 4.2 3 1 28.5 27.0 77.6 73.5 4.1
48 4 2.8 3.25 3.5 41.7 37.6 113.4 102.5 10.9
Appendices
FR=Food responsiveness; EOE=Emotional over-eating; SR=Satiety responsiveness; SE=Slowness in eating.
359

CEBQ

Uploaded by

Copyright:

Available Formats

CEBQ

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CEBQ

Uploaded by

Copyright:

Available Formats

APPPETITIVE TRAITS AND THEIR RELATIONSHIPS TO WEIGHT AND

Claudia Madeleine Elizabeth Hunot

A thesis submitted for the degree of Doctor of Philosophy

UNIVERSITY COLLEGE LONDON

Chapter 1. Weight management .......................................................................................... 19

1.2 Approaches to weight management ........................................................................... 20

1.2.1 Structural approaches ......................................................................................... 21

1.2.2 Pharmacological interventions ............................................................................ 22

1.2.3 Surgical interventions .......................................................................................... 23

1.2.4 Lifestyle interventions ......................................................................................... 24

1.4 Tailoring weight management interventions to improve outcomes .......................... 29

1.5 Summary of the findings ............................................................................................. 31

Chapter 2. Appetitive traits and weight............................................................................... 32

2.2 Appetite ....................................................................................................................... 32

2.2.1 The homeostatic control of appetite .................................................................. 33

2.2.2 The hedonic control of appetite .......................................................................... 34

2.2.3 Interplay of homeostatic and hedonic control of appetite ................................. 34

2.3 Appetitive traits ........................................................................................................... 35

2.4 Measures of appetite and appetitive traits................................................................. 36

2.4.1 Experimental measures of appetite and appetitive traits .................................. 37

2.4.2 Psychometric measures of appetite and appetitive traits .................................. 41

2.5 Appetitive traits and weight ........................................................................................ 43

2.5.1 The genetics of obesity........................................................................................ 43

2.5.2 The “Behavioural Susceptibility Theory” of obesity ............................................ 45

2.5.3 Evidence for the “Behavioural Susceptibility Theory” ........................................ 45

2.6 Summary of the findings ............................................................................................. 48

Chapter 3. Aims of the thesis ............................................................................................... 51

Chapter 4. Study 1: Systematic review on questionnaire measures of appetite and

4.2 Aim .............................................................................................................................. 55

4.3 Methods ...................................................................................................................... 55

4.3.1 Information sources and search strategy ............................................................ 55

4.3.2 Eligibility criteria .................................................................................................. 56

4.3.3 Study selection .................................................................................................... 57

4.3.4 Data extraction process – Classifying and coding studies ................................... 58

4.3.5 Assessing the robustness of the questionnaires ................................................. 58

4.3.6 Most commonly measured appetitive traits by age group ................................. 60

4.4 Results ......................................................................................................................... 60

4.4.1 Search results ...................................................................................................... 60

4.4.2 Evaluation of psychometric properties: Reliability and validity of the

4.4.3 Overall robustness of the questionnaires ........................................................... 78

4.4.4 Most commonly measured appetitive traits by age group ................................. 84

4.5 Discussion .................................................................................................................... 85

4.5.1 Measures that emcompass ‘emotional’ aspects of appetite .............................. 87

4.5.2 Measures that emcompass ‘restraint’ aspects of appetite ................................. 89

4.5.3 Measures that emcompass ‘food and eating/externality’ aspects of appetite .. 89

4.5.4 Limitations ........................................................................................................... 90

Chapter 5. Study 2: Development of the ‘Adult Eating Behaviour Questionnaire’ ............ 96

5.2 Aims ............................................................................................................................. 96

5.3 Methods ...................................................................................................................... 98

5.3.1 Translation of the CEBQ into the AEBQ ............................................................... 98

5.3.2 Review of items from other questionnaires on appetite from existing

5.3.3 Piloting in a sample of adults ............................................................................ 101

5.3.4 Assessing the factor structure of the AEBQ ...................................................... 102

5.3.5 Summary statistics............................................................................................. 107

5.3.6 Internal reliability .............................................................................................. 107

5.4 Ethical approval ......................................................................................................... 108

5.5 Results ....................................................................................................................... 108

5.5.1 Assessment of the factor structure ................................................................... 108

5.5.2 Criteria used for eliminating items .................................................................... 109

5.5.3 Summary statistics............................................................................................. 114