Behaviour Research and Therapy 42 (2004) 551–567

www.elsevier.com/locate/brat

Validity of the Eating Disorder Examination Questionnaire

(EDE-Q) in screening for eating disorders
in community samples
J.M. Mond a,
, P.J. Hay b, B. Rodgers c, C. Owen a, P.J.V. Beumont d
a
Department of Psychological Medicine, The Canberra Hospital, Canberra ACT, 2606 Australia
b
Department of Psychiatry, University of Adelaide, Adelaide SA, 5001 Australia
c
Centre for Mental Health Research, The Australian National University, Canberra ACT, 0200 Australia
d
Department of Psychological Medicine, University of Sydney, Sydney NSW, 2006 Australia
Received 9 March 2003; received in revised form 27 May 2003; accepted 4 June 2003

Abstract

In order to examine the concurrent and criterion validity of the questionnaire version of the Eating
Disorders Examination (EDE-Q), self-report and interview formats were administered to a community
sample of women aged 18–45 (n ¼ 208). Correlations between EDE-Q and EDE subscales ranged from
0.68 for Eating Concern to 0.78 for Shape Concern. Scores on the EDE-Q were significantly higher
than those of the EDE for all subscales, with the mean difference ranging from 0.25 for Restraint to
0.85 for Shape Concern. Frequency of both objective bulimic episodes (OBEs) and subjective bulimic epi-
sodes (SBEs) was significantly correlated between measures. Chance-corrected agreement between
EDE-Q and EDE ratings of the presence of OBEs was fair, while that for SBEs was poor. Receiver oper-
ating characteristic (ROC) analysis, based on a sample of 13 cases, indicated that a score of 2.3 on the
global scale of the EDE-Q in conjunction with the occurrence of any OBEs and/or use of exercise as a
means of weight control, yielded optimal validity coefficients (sensitivity ¼ 0:83, specificity ¼ 0:96,
positive predictive value ¼ 0:56). A stepwise discriminant function analysis yielded eight EDE-Q items
which best distinguished cases from non-cases, including frequency of OBEs, use of exercise as a means
of weight control, use of self-induced vomiting, use of laxatives and guilt about eating. The EDE-Q has
good concurrent validity and acceptable criterion validity. The measure appears well-suited to use in pro-
spective epidemiological studies.
# 2003 Elsevier Ltd. All rights reserved.

Keywords: Eating disorders; Assessment; Concurrent and criterion validity; Two-phase design

Corresponding author. Tel.: +61-2-6244-3876; fax: +61-2-6244-3502.
E-mail address: jonathan.mond@act.gov.au (J.M. Mond).

0005-7967/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0005-7967(03)00161-X
552 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567

1. Introduction

It is generally accepted that assessment of the specific psychopathology of eating-disordered

behaviour is best achieved through the administration of a structured or semi-structured inter-
view by clinicians or by trained lay interviewers (Garner, 2002). Frequently, however, con-
straints on time and resources encourage the use of self-report measures. For example in
epidemiological studies of low-prevalence psychiatric disorders, it is usually not possible to con-
duct interview assessment with the total sample. For this reason the use of a two-phase design,
in which interview assessment is conducted only with probable cases identified on the basis of a
self-report measure, is often employed in such studies (Dunn, Pickles, Tansella, & Vazquez-
Barquero, 1999).
Among self-report measures of eating-disordered behaviour, the Eating Attitudes Test (EAT;
Garner & Garfinkel, 1979) has been widely employed as an outcome measure in clinical and
research settings. A 26-item version of the original 40-item scale (EAT-26; Garner, Olmsted,
Bohr, & Garfinkel, 1982) has also frequently been used to detect probable cases of eating dis-
orders in general population surveys. However, the measure was originally developed to assess
the specific behaviours and attitudes of anorexia nervosa (AN) patients, and its validity as a
‘case-finding’ instrument has frequently not been supported (Patton & Szmukler, 1995). An
additional shortcoming of the EAT is that an omnibus score is derived at the expense of dimen-
sional information concerning particular symptoms (Anderson & Williamson, 2002). The Eating
Disorders Inventory (EDI; Garner, Olmsted, & Polivy, 1983), arguably the most comprehensive
self-report measure of eating disorder psychopathology, has also been widely used, but it is too
long for use as a screening instrument and it has not been validated for this purpose (Garner,
1991).
In general, the use as case-finding instruments of measures developed for use in clinical sam-
ples is problematic, since the characteristics of individuals identified as cases in general popu-
lation surveys may differ from those of individuals presenting to services. For example, items
addressing the occurrence of extreme methods of weight control, such as self-induced vomiting
and laxative misuse, may be of limited use in community samples, because the prevalence of
such behaviours is much lower (Garfinkel et al., 1995). Similarly, the extreme dietary restriction
and very low body weights characteristic of AN patients are rarely encountered in general
population surveys (Walters & Kendler, 1995). Instruments such as the EAT may therefore not
be expected to perform well in detecting the relatively more common eating disorders, such as
BED and partial-syndrome cases of AN and BN (Hay, Marley, & Lemar, 1998).
A promising alternative to the EAT is the self-report version of the Eating Disorders Examin-
ation (EDE-Q; Fairburn & Beglin, 1994), a 36-item questionnaire derived from and scored in
the same way as the interview schedule (EDE; Fairburn & Cooper, 1993). The EDE is widely
regarded as the instrument of choice for the assessment and diagnosis of DSM-IV eating dis-
orders (Garner, 2002). The EDE-Q provides a similarly comprehensive assessment of the spe-
cific psychopathology of eating-disordered behaviour in a relatively brief self-report format.
Studies of the validity of the EDE-Q have demonstrated a high level of agreement between the
EDE-Q and EDE in assessing the core attitudinal features of eating disorder psychopathology
in the general population (Fairburn & Beglin, 1994), among female substance abusers (Black &
Wilson, 1996), and in clinical samples of both bulimia nervosa (BN) and binge eating disorder
 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567 553

(BED) patients (Carter, Aime, & Mills, 2001; Wilfley, Schwartz, Spurrell, & Fairburn, 1997).
Acceptable internal consistency and test–retest reliability have also been demonstrated (Luce &
Crowther, 1999). For these reasons, the EDE-Q has increasingly been employed as an outcome
measure and as an adjunct to the use of the EDE in descriptive studies (Anderson &
Williamson, 2002; Pike, Dohm, Striegel-Moore, Wilfley, & Fairburn, 2001).
The validity of the EDE-Q in assessing eating disorder behaviours is less clear. In particular,
significant discrepancies between the EDE-Q and the EDE with respect to assessment of binge
eating behaviour have been reported in both general population and clinical samples (Black &
Wilson, 1996; Carter, Aime, & Mills, 2001; Fairburn & Beglin, 1994; Wilfley, Schwartz, Spurrell
& Fairburn, 1997). These findings are more likely to reflect the inherent difficulty of assessing
binge eating behaviours by self-report rather than a particular failing of the EDE-Q (Meadows,
Palmer, Newball, & Kendrick, 1986), but would nevertheless be expected to detract from the
validity of the measure as a case-finding instrument. Assessment of the frequency of self-induced
vomiting and/or laxative abuse by means of the EDE-Q appears to correspond more closely
with frequency established through interview assessment, though in clinical samples the mean
number of episodes reported may be higher when assessed with the EDE (Carter, Aime, &
Mills, 2001; Fairburn & Beglin, 1994).
To date, only Fairburn and Beglin (1994) have examined the validity of the EDE-Q in a gen-
eral population sample and only concurrent validity, that is, agreement between EDE-Q and
EDE scores, was considered in this study. Beglin and Fairburn (1992) demonstrated the predic-
tive validity of a short-form of the EDE-Q, comprising those (nine) items which best dis-
criminated cases of clinically significant eating disorders from non-cases in a community sample
of women aged 18–35. The items were: presence or absence and number of days of self-induced
vomiting and laxative misuse; frequency of OBEs and of diuretic misuse; preoccupation with
food and calories; guilt about eating; and pursuit of thinness. The resulting scale (EDE-S) was
found to perform better than the EAT in identifying cases and it has subsequently been
employed in some general population surveys (Steinhausen, Winkler, & Meier, 1997; Hay,
Marley & Lemar, 1998). However, the criterion validity of the EDE-Q has not been established.
In preparation for a two-phase epidemiological study of disability and health service uti-
lization associated with the more commonly occurring eating disorders, we examined the val-
idity of the EDE-Q in the population of interest, namely, women aged 18–45 years. Specific
aims of the study were to provide further evidence for the concurrent validity of the EDE-Q, to
demonstrate the criterion validity of the EDE-Q, to establish appropriate cut-off points for use
in screening general population samples and to replicate the findings of Beglin and Fairburn
(1992).

2. Method

2.1. Design and participants

The study was carried out in the Australian Capital Territory (ACT) region of Australia
(population 314,000), a highly urbanized region which includes the city of Canberra. Parti-
cipants were recruited in two phases. At the first phase of the study self-report questionnaires
554 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567

were delivered to a sample of 802 female ACT residents aged 18–45, selected at random from
the electoral roll and stratified by age in 5-year bands. The questionnaire included the EDE-Q,
socio-demographic information and self-reported weight and height, as well as measures of gen-
eral psychological distress (K-10; Kessler & Mroczek, 1994) and impairment in role functioning
(SF-12; Ware, Kosinski, & Keller, 1996). Findings concerning the latter measures, which relate
to other aims of the project, will be reported separately.
Completed questionnaires were received, following reminder letters, from 495 respondents,
which represented a response rate of 69.5%, taking account of individuals who were no longer
resident at the listed address (n ¼ 90; 11.2%). Only information concerning age bracket was
available for non-respondents at the first phase. The age distribution of respondents did not dif-
fer significantly from that of non-respondents (v2 ¼ 7:32, p > 0:05). All participants who
returned a completed questionnaire and who indicated a willingness to be contacted by tele-
phone at a future date (n ¼ 308), were approached to participate in the second phase of the
study, involving administration of the EDE. The conditions of ethics committee approval stipu-
lated that participants who did not provide a contact phone number (n ¼ 187) were not eligible
for interview.
Interviews were completed with 208 individuals, representing a response rate of 76.8% at
the second phase, taking account of participants who could not be contacted (n ¼ 22) or
who were away from or no longer resident in the ACT (n ¼ 15). Demographic characteristics of
the interviewed samples are given in Table 1. Individuals interviewed were older
(mean age ¼ 35:3 vs 32:1 years; t ¼ 4:09; p < 0:05) and more likely to be married (62.0% vs
48.8%; v2 ¼ 16:44, p < 0:05), and to have one or more children (65.9% vs 51.1%, v2 ¼ 13:03,
p < 0:05), than individuals not interviewed (n ¼ 286). The two groups did not differ with respect
to any of the other demographic variables assessed, nor with respect to eating-disordered behav-
iour (as measured by the EDE-Q) or body mass index (BMI, kg/m2; Garrow, 1988).

2.2. Measures

2.2.1. Eating Disorders Examination (12th edition)

The Eating Disorders Examination (Fairburn & Cooper, 1993) is a semistructured, investi-
gator-based interview designed to assess the core attitudinal and behavioural features of patients
with eating disorders. Where the instrument is used for diagnostic purposes, relevant items are
modified to assess a 3-month period in order to generate operational diagnoses according to
DSM-IV criteria (American Psychiatric Association, 1994). Otherwise, items refer to the past 28
days. Behavioural features, such as binge eating and use of self-induced vomiting, are rated in
terms of both the number of days on which these occurred and the number of individual epi-
sodes. For the assessment of binge eating, responses to probe questions concerning the amount
of food consumed and the circumstances of eating are used to categorize episodes of overeating,
based on the presence or absence of two features: loss of control over eating at the time of the
episode (required for subjective and objective bulimic episodes, SBEs and OBEs) and consump-
tion of an ‘objectively large’ amount of food (required for OBEs and episodes of objective over-
eating). For the assessment of attitudinal features, responses to each of 22 items are rated by the
interviewer on a 7-point scale of severity or frequency, with higher scores indicating greater
levels of disturbance. Responses to these items may be divided into four subscales, namely,
 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567 555

Table 1
Demographic characteristics of the interviewed sample (n ¼ 208)
Mean (SD)
Age (years) 35.3 (8.5)
BMI (kg/m2) 25.2 (5.4)

Percentage (%)
Country of birth
Australia 81.7
Other 17.3
First language
English 90.4
Other 9.1
Marital status
Married 62.0
Living as married 7.7
Single 19.7
Separated/divorced 6.7
Children (one or more) 65.9
Education (level completed)
High school (12 years) 84.6
Trade/Tech cert/Dip 23.0
Bachelor’s degree 22.6
Postgraduate qualification 12.0
Employment status
Employed full-time 43.8
Employed part-time/casually 41.8
Home duties 13.5
Studying full-time 11.5
Unemployed 1.9

Restraint, Eating Concern, Weight Concern and Shape Concern. Reliability and validity of the
interview are well established (Fairburn & Cooper, 1993).

2.2.2. Eating Disorders Examination—self report version (EDE-Q)

The EDE-Q (Fairburn & Beglin, 1994) is a 36-item measure derived from the EDE interview
and focusing on the past 28 days. Each item is taken directly from a corresponding EDE item,
with modifications to wording as required. Items addressing eating disorder attitudes are scored
using the same 7-point, forced-choice, rating scheme. Subscale and global scores are derived in
the same way as for the EDE, with scores of four or higher on key items considered to lie in the
clinical range. Frequencies of key behaviours are assessed in terms of the number of episodes of
each behaviour occurring during the past 4 weeks. As in the EDE, an attempt is made to differ-
entiate between OBEs and SBEs, with separate questions addressing these different concepts.
The item ‘pursuit of thinness’, included in Beglin and Fairburn’s (1992) study, was not included
in the revised version of the EDE-Q employed in the present study.
556 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567

2.3. Criteria for cases

Cases of clinically significant eating disorders were identified on the basis of EDE assessment.
Operational criteria for the diagnosis of AN and BN according to DSM-IV are provided in the
EDE and criteria for BED are suggested (Fairburn & Cooper, 1993). While the occurrence of
two or more binge eating episodes per week is required for the diagnosis of BN and BED in
DSM-IV, the Oxford criterion of 12 or more episodes over the past 3 months was followed
because the validity of the twice per week criterion has not been established (Garfinkel et al.,
1995). Operational criteria for cases of EDNOS other than BED are not provided in the EDE.
In the absence of these criteria, individuals were classified as cases of partial syndrome AN or
BN if all but one of the operational criteria for the respective DSM-IV diagnoses, according to
the EDE, were met. Individuals reporting regular (i.e. at least weekly) SBEs and extreme weight
or shape concerns, but not engaging in regular extreme weight control behaviours, were also eli-
gible for the diagnosis of EDNOS, because evidence from general population samples suggests
that individuals meeting these criteria have a symptom severity and course similar to those
meeting criteria for BED (Hay & Fairburn, 1998; Hay, Fairburn, & Doll, 1996).
While the EDE manual specifies a criterion for the use of exercise as a compensatory behav-
iour of ‘intense exercise predominantly intended to use calories or change shape, weight or body
composition occurring on average at least 5 days per week over the past 3 months’, these cri-
teria have not been validated in a general population sample. In the present study, a question was
added to the EDE assessment of exercise to assess the primary motivation for exercise on a scale
of ‘1’ (exercise solely for weight or shape reasons) to ‘5’ (exercise not at all for weight or shape
reasons). Subjects were considered to meet the exercise criterion if they engaged in intense or
moderate to intense exercise solely or primarily for weight or shape reasons (namely, a score of ‘1’
or ‘2’) three or more times per week, (Mond, Hay, Rodgers, Owen & Beumont, submitted).

2.4. Procedure

Because a considerable period of time elapsed between completion of the first and second
phases (mean ¼ 303:2 days, SD ¼ 57:4, range ¼ 141 444), the EDE-Q was re-administered to
interview participants immediately prior to completion of the EDE. A questionnaire which
included the EDE-Q and weight and height information was posted to participants with a letter
confirming the time and place of the interview (mostly in the subject’s home) and with the
request that subjects complete the questionnaire on the day or evening prior to the date of
interview. This step was taken in order to ensure that the EDE-Q and EDE examined the same
28-day period. Subjects were not informed that questions appearing in the questionnaire would
be repeated at the interview.
Interviewers received weekly training in the administration of the EDE from the first author
over a period of 10 weeks. During these sessions, tapes of EDE interviews from previous
research (Hay, Marley & Lemar, 1998) were reviewed and discussed and practice interviews
conducted. Each interviewer also conducted pilot interviews and received detailed feedback con-
cerning these interviews. The process of individual and group feedback continued during the
assessment phase of the study until there was complete agreement between each interviewer and
the first author with respect to the ratings of each item. All interviews were audio-taped, except
 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567 557

where the subject requested that taping not be conducted, and all tapes were reviewed by the
first author. Overall conduct of training and assessment was supervised by the second author
(PH), who had trained within the Oxford group. Formal assessment of interrater reliability was
not conducted because agreement between interviewers has been shown to be excellent given
training and ongoing supervision of this kind (Fairburn & Cooper, 1993). Where necessary,
tapes were forwarded to the second author and a consensus reached on ratings and diagnoses
between the first and second authors. All interviewers had completed a minimum of four years
tertiary education and all had relevant clinical experience. Only female interviewers were
employed.

2.5. Statistical analysis

Correlations between item and subscale scores on the (second administration of the) EDE-Q
and EDE were calculated using Pearson’s r, while t-tests for paired samples were used to test for
statistical significance in the differences between mean scores on items and subscales of the
respective measures. For non-normally distributed data (i.e. binge eating episodes), non-para-
metric tests (Kendall’s tau-b and Wilcoxon’s matched pairs signed rank sum test) were used to
examine the correspondence between EDE-Q and EDE assessment. An additional measure of
agreement with respect to the assessment of binge eating behaviours was provided by the kappa
statistic, which represents the chance-corrected level of agreement between two categorical vari-
ables, in this case the presence or absence of SBEs and OBEs.
For the analysis of criterion validity, the significance of differences between the mean scores
of cases and non-cases on the individual items and subscales of the EDE-Q was tested by means
of Mann–Whitney U-tests. Discriminant function analysis was employed in order to identify
those EDE-Q items which best discriminated between groups, while receiver operating charac-
teristic (ROC) analysis was employed to determine the score on the global EDE-Q scale which
provided the optimal trade-off between sensitivity (proportion of true cases screening positive;
Se) and specificity (proportion of true non-cases screening negative; Sp). The positive predictive
value (proportion of screen positives who are true cases; PPV) of the EDE-Q at the optimal cut-
off point was also calculated. All analyses were conducted using SPSS version 10.0.

3. Results

3.1. Concurrent validity

3.1.1. Comparison of item and subscale scores

There were 195 subjects for whom complete information was available on both EDE-Q and
EDE. Table 2 shows mean scores on each subscale of the EDE-Q and EDE, as well as correla-
tions between subscale scores on the respective measures. Data from Fairburn and Beglin’s
(1994) study are given for comparison.
Scores on individual items of the EDE-Q were higher than those of the EDE for all but five
items, these being: ‘restraint overeating’, ‘food avoidance’, ‘eating in secret’, ‘importance of
558 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567

Table 2
Mean scores on and correlations between EDE-Q and EDE subscales for the present study (n ¼ 195) and for Fair-
burn and Beglin (1994) (n ¼ 243)
Subscale EDE-Q EDE t r
Mean (SD) Mean (SD)
Restraint 1.29 (1.27) 1.04 (1.33) 3.51


0.71



Fairburn and Beglin (1994) 1.25 (1.32) 0.94 (1.09) 6.26


0.81



Eating concern 0.59 (0.84) 0.22 (0.52) 8.26


0.68



a a
Fairburn and Beglin (1994) 0.62 (0.86) 0.27 (0.59)




Weight concern 1.64 (1.31) 1.12 (1.06) 8.53 0.77



Fairburn and Beglin (1994) 1.59 (1.37) 1.18 (0.93) 7.40


0.79



Shape concern 2.16 (1.44) 1.31 (1.17) 12.07


0.78



Fairburn and Beglin (1994) 2.15 (1.60) 1.34 (1.09) 12.88


0.80



Global 1.42 (1.04) 0.92 (0.89) 10.97


0.84



a a
Fairburn and Beglin (1994) 1.55 (1.21) 0.93 (0.81)




p<0.001.
a
t-Values and correlations not available for eating concern and global scales.

weight’ and ‘importance of shape’ (all p > 0:05). The largest discrepancies were for the items
‘flat stomach’ (mean difference ¼ 1:73), ‘avoidance of exposure’ (1.19), ‘dissatisfaction with
weight’ (1.10) and ‘dissatisfaction with shape’ (1.03) (all p < 0:0001). Correlations between item
pairs ranged from 0.22 (p < 0:01) for the item ‘eating in secret’ to 0.69 (p < :001) for the items
‘dissatisfaction with shape’ and ‘discomfort seeing body’. A total discrepancy score was calcu-
lated for each subject as the average difference between scores on items of the EDE-Q and those
on corresponding EDE items (Black & Wilson, 1996). Total discrepancy scores ranged from
1.13 to 2.74, with a mean of 0.53 (SD ¼ 0:60). A positive correlation was observed between
scores on the global EDE-Q scale and total discrepancy scores (r ¼ 0:49, p < 0:01).

3.1.2. Assessment of binge eating behaviours

The chance corrected agreement between EDE-Q and EDE ratings of the presence of one or
more OBEs was fair (kappa ¼ 0:47; t ¼ 7:31, p < 0:01), while the chance corrected agreement
between EDE-Q and EDE ratings of the presence of one or more SBEs was poor
(kappa ¼ 0:26; t ¼ 3:92, p < 0:01). Among the six subjects who reported one or more OBEs on
both the EDE and EDE-Q, the mean number of binge eating episodes were 13.33 (SD ¼ 12:50),
and 8.17 (SD ¼ 7:57), respectively (Wilcoxon Signed Ranks z ¼ 1:63, p > 0:05; Kendall’s
tau-b ¼ 0:93, p < 0:05), while among the seven subjects who reported one or more SBEs on
both the EDE and EDE-Q, the mean number of episodes were 10.57 (SD ¼ 12:52), and 7.29
(SD ¼ 9:57), respectively (Wilcoxon Signed Ranks z ¼ 0:41, p > 0:05; Kendall’s tau-b ¼ 0:78,
p < 0:05).

3.2. Assessment of weight control behaviours

One subject reported current use of self-induced vomiting as a means of weight control in the
EDE-Q, while one subject reported the use of laxatives. In each case frequencies were below
 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567 559

threshold levels for the DSM-IV diagnosis of BN. The occurrence of each of these behaviours
was confirmed at interview and there were no subjects who reported the use of purging at inter-
view, but not on the EDE-Q. A third subject reported the use of slimming tablets at interview.
On the EDE-Q, 35 subjects (17.9%) reported exercising hard in order to control weight or shape
on one or more occasions during the preceding 4 weeks, while 22 subjects (11.3%) reported the
use of exercise as a compensatory behaviour on four or more occasions. Direct comparison of
exercise behaviour as assessed by the EDE-Q and by the EDE was not possible on account of
the modified assessment of exercise employed for the EDE.

3.3. Criterion validity

A total of 13 subjects (6.3%) met the study criteria for a clinically significant eating disorder.
There were no subjects meeting criteria for DSM-IV AN or BN purging type. There was one
subject who met DSM-IV criteria for non-purging BN and one subject with EDNOS of a BN
purging type, for whom frequencies of OBEs and self-induced vomiting were below threshold. A
further five subjects reported regular use of inappropriate weight control behaviours, with
extreme weight and/or shape concern, but did not describe regular (i.e. 1/week) episodes of
overeating in which a loss of control was experienced. Two subjects met criteria for BED, while
the remaining four subjects met criteria for BED, except that their overeating episodes were not
objectively large.
The mean age of the case group was 30.2 years (SD ¼ 10:2, n ¼ 13), while that of the non-
cases was 35.7 years (SD ¼ 8:3, n ¼ 195) (z ¼ 1:65, p > 0:05). Mean BMI was 26.4 (SD ¼ 4:3)
for the case group and 26.2 (SD ¼ 5:9) for the non-cases (z ¼ 0:66, p > 0:05). Mean scores for
cases and non-cases on the EDE-Q and EDE subscales are given in Table 3, as well as the per-
centage of subjects in each group who reported particular eating disorder behaviours (i.e. binge
eating episodes and use of exercise as a compensatory behaviour) on the EDE-Q. Mann–Whit-
ney U-tests confirmed that individuals diagnosed as cases had significantly higher scores than
non-cases on each subscale of both the EDE-Q and EDE (all p < 0:001).
Additional analysis indicated that the differences between groups were significant at the 0.05
level for all EDE-Q items except ‘flat stomach’ (z ¼ 1:69, p ¼ 0:09). Among the individual
items of the EDE, only ‘avoidance of eating’ did not differ between groups (z ¼ 0:43,
p ¼ 0:67). The mean discrepancy between scores on items of the EDE-Q and on corresponding
items of the EDE was significantly lower among individuals identified as cases than among non-
cases (0.07 vs 0.56; Mann–Whitney U-test z ¼ 2:21, p < 0:05). However, the correlation
between scores on the EDE-Q and EDE global scales did not reach significance among the case
group (Kendall’s tau-b ¼ 0:39, p ¼ 0:07), while this correlation was highly significant among the
non-cases (r ¼ 0:82, p < 0:0001). For the total sample, there was a significant correlation
between BMI and global scores on both the EDE-Q (r ¼ 0:26, p < 0:01) and EDE (r ¼ 0:17,
p < 0:05). The mean BMI of those subjects who scored at or above the 90th percentile of scores
on the global EDE-Q (2.75), but who were non-cases, was 38.1 kg/m2 (SD ¼ 4:3, n ¼ 12).
ROC analysis indicated that for the 22 items comprising the EDE-Q subscales, the optimal
compromise between sensitivity and specificity was achieved at a score of 2.3 on the global
EDE-Q scale (Se ¼ 0:92, Sp ¼ 0:86), yielding a positive predictive value (PPV) of 0.30. Using
560 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567

Table 3
Mean EDE-Q and EDE subscale scores and frequency of self-reported eating disorder behaviours for individuals
identified as cases (n ¼ 13) and non-cases (n ¼ 195) on the basis of EDE assessment
Cases Non-cases
Mean (SD) Mean (SD)
EDEQ restraint 2.65 (1.48) 1.19 (1.21)
EDEQ eating concern 2.02 (0.95) 0.49 (0.74)
EDEQ weight concern 3.68 (1.08) 1.49 (1.20)
EDEQ shape concern 4.01 (0.98) 2.03 (1.38)
EDEQ global scale 3.09 (0.83) 1.30 (0.96)
EDE restraint 3.17 (1.43) 0.89 (1.19)
EDE eating concern 1.60 (1.03) 0.15 (0.35)
EDE weight concern 3.43 (0.71) 0.98 (0.93)
EDE shape concern 3.67 (0.83) 1.18 (1.03)
EDE global scale 2.97 (0.65) 0.80 (0.73)

Eating disorder behaviours % %

Any OBEsa 58.3 5.5
4 or more OBEs/28 days 25.0 2.2
Any SBEsb 41.7 13.2
4 or more SBEs/28 days 25.0 6.0
Any exercisec 66.7 12.1
4 or more Sessions exercise/28 days 58.3 8.2
a
Objective bulimic episodes.
b
Subjective bulimic episodes.
c
‘Exercising hard as a means of controlling shape or weight’.

the sum of these items as the independent variable (maximum score ¼ 132) rather the mean
score on the global scale, a cut-off score of 56 was obtained. Additional analysis indicated that
only two items distinguished between individuals in the case group who scored above the cut-off
(n ¼ 11) and non-cases who scored above this point (n ¼ 26). These items were the occurrence
of OBEs (Mann–Whitney U-test z ¼ 2:46, p < 0:05) and the use of exercise behaviour as a
compensatory behaviour (z ¼ 2:42, p < 0:05). When the occurrence of any OBEs and/or exer-
cising for weight or shape reasons at least once per week were specified as criteria in addition to
a score of 2.3 on the global scale of the EDE-Q (or total score of 56), PPV increased to 0.56
and specificity to 0.96, with only a small decrease in sensitivity (0.83).
A stepwise discriminant function analysis was conducted in order to identify those EDE-Q
items which best discriminated cases from non-cases. Using a liberal criterion for entry of 0.15
(Tabachnick & Fidell, 1996), eight items were identified, these being: frequency of OBEs, use of
exercise as a means of weight control, use of self-induced vomiting, use of laxatives, ‘guilt about
eating’, ‘social eating’, ‘discomfort seeing body’ and ‘avoidance of exposure’ (v ¼ 105:90,
p < 0:01). While the occurrence of any OBEs was the first variable to enter the model, this was
replaced by frequency of OBEs at a subsequent step.
 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567 561

4. Discussion

4.1. Correspondence between EDE-Q and EDE subscales

The findings of the present study provide further support for self-report assessment of the
attitudinal aspects of eating disorder psychopathology. Consistent with the findings of Fairburn
and Beglin (1994), scores on EDE-Q and EDE were highly correlated and to a similar degree
across the four subscales. Also consistent with previous research, scores on items of the EDE-Q
tended to be higher than those of the corresponding EDE items, reflecting the use of more strin-
gent wording in the interview schedule than in the self-report format for certain items (Wilfley,
Schwartz, Spurrell & Fairburn, 1997). Discrepancies between EDE-Q and EDE scores were
greatest for the Shape Concern subscale, while relatively smaller discrepancies were apparent on
the Restraint subscale. This finding has been taken to reflect the fact that items comprising the
Restraint subscale address relatively unambiguous attitudes and behaviours, while those of the
Shape Concern subscale tap more complex features (Fairburn & Beglin, 1994). However, differ-
ences on the items ‘importance of weight’ and ‘importance of shape’ were minimal in the
present study. This finding is noteworthy given that the undue influence of weight or shape con-
cerns on self-evaluation is a criterion for both AN and BN in DSM-IV.

4.2. Assessment of binge eating and other eating disorder behaviours

Also consistent with findings from previous research, marked discrepancies were apparent
between the two measures with respect to the assessment of binge eating behaviours. In contrast
to the findings of Fairburn and Beglin (1994) and of Black and Wilson (1996), however, fre-
quency of OBEs was higher for the EDE than for the EDE-Q in the present study. Of 17 sub-
jects who reported the occurrence of one or more OBEs on the EDE-Q, only six subjects were
judged to have experienced one or more OBEs according to the EDE. A similar finding, though
with an even greater discrepancy, was apparent in the assessment of SBEs, while the number of
subjects who reported OBEs or SBEs on the EDE but not on the EDE-Q was much smaller.
Assuming that interview assessment of binge eating is the more accurate of the two methods
(Wilson, 1993), then in this study the presence of bulimic episodes was overestimated by self-
report, while the frequency of these episodes was underestimated. Thus, the prevalence of BN,
BED and sub-clinical variants of these disorders may in fact be underestimated by the EDE-Q
in general population studies. The prevalence of recurrent binge eating (i.e. OBEs  1=week) as
assessed by the EDE in the present study (n ¼ 6, 2.9%) is consistent with findings from large-
scale general population surveys employing interview assessment (e.g. Garfinkel et al., 1995).
Data concerning the number of subjects reporting one or more OBEs on the respective meas-
ures, and the number of subjects reporting one or more OBEs on both measures, were not
reported by Fairburn and Beglin (1994) or by Black and Wilson (1996). It is possible that
inclusion in the analysis of subjects reporting OBEs on the EDE-Q but not on the EDE had the
effect of elevating mean binge eating frequency on the EDE-Q relative to that for the EDE in
these studies. Such an effect would be of less consequence in a sample of BED or BN patients,
where most or all subjects would be found to report OBEs on both measures. Consistent with
this interpretation, a higher frequency of OBEs when assessed by the EDE has been reported in
562 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567

samples of both BED and BN patients (Wilfley, Schwartz, Spurrell, & Fairburn, 1997; Carter,
Aime, & Mills, 2001). A likely explanation for these findings is that recall of binge eating behav-
iours is enhanced with EDE assessment through the use of calendar prompts and detailed ques-
tioning (Wilfley, Schwartz, Spurrell, & Fairburn, 1997).
We are aware of only one other study in which agreement between the EDE-Q and EDE with
respect to the occurrence of binge eating episodes has been assessed. Kalarchian, Wilson, Brolin,
and Bradley (2000) found moderate agreement between the EDE-Q and EDE with respect to
the occurrence of regular (two or more/week) OBEs (kappa ¼ 0:45) in a sample of morbidly
obese gastric bypass surgery candidates. However, only 13 of 23 (56.5%) subjects classified as
regular binge eaters on the EDE-Q were confirmed by the EDE, while of 22 subjects who were
classified as regular binge eaters according to the EDE, only 13 (59.1%) had reported regular
episodes of binge eating on the EDE-Q. It may also be noted that frequency of OBEs was
higher when assessed with the EDE than with the EDE-Q in Kalarchian, Wilson, Brolin, and
Bradley’s (2000) study, consistent with the findings of Carter, Aime & Mills, 2001, with those of
Wilfley, Schwartz, Spurrell, & Fairburn (1997) and with those of the present study.
Wilfley, Schwartpurrell, and Fairburn (1997) noted that assessment of binge eating is more
problematic in BED patients than in BN patients, since, in the absence of purging and/or other
extreme methods of weight control, episodes are less discrete and more difficult to identify. This
applies equally, and perhaps more so, to the assessment of binge eating in general population
samples. In a recent study of bulimic-type eating disorders in primary practice, individuals
reporting regular SBEs, with or without the use of extreme weight control behaviours, com-
prised the majority of cases of clinically significant eating disorders (Hay, Marley, & Lemar,
1998). While the clinical significance of SBEs is currently a source of contention (Pratt, Niego,
& Agras, 1998), the findings of the present study suggest that assessment of SBEs by means of
self-report may be no less problematic than that of OBEs. Results reported by Grilo, Masheb,
and Wilson (2001a) in a sample of BED patients support this conclusion. These authors found
that while frequency of OBEs was significantly correlated for EDE-Q and EDE assessment, a
non-significant correlation was observed for SBE frequency. An additional finding of this study
was that the frequency of SBEs was significantly higher when assessed by self-monitoring than
when assessed by the EDE-Q. Both findings were replicated in a different sample of BED
patients Grilo, Masheb & Wilson (2001b), leading the authors to conclude that ‘the EDE-Q
does not appear to have use for assessing forms of overeating other than OBEs’ (p. 420). How-
ever, EDE-Q and EDE scores for SBE frequency were significantly correlated in the study of
BN patients by Carter, Aime, & Mills (2001). The validity of the EDE-Q in assessing the occur-
rence and frequency of SBEs requires further investigation.
Low base rates of self-induced vomiting and laxative misuse precluded comparison between
interview and self-report assessment of these behaviours in the present study. Carter, Aime, &
Mills, 2001 reported significant correlations between the frequency of episodes of both self-
induced vomiting and laxative misuse assessed by the EDE-Q and the EDE in a sample of BN
patients, though the mean number of episodes of self-induced vomiting was significantly higher
for the EDE. Given the unambiguous nature of purging behaviours, problems of definition
would not be expected to be a source of discrepancies were this relationship to be examined in a
larger general population sample. More problematic may be the deliberate underreporting of
such behaviours in either or both forms of assessment (Evans & Wertheim, 2002; Vitousek,
 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567 563

Daly, & Heiser, 1991). With respect to the assessment of exercise behaviour, assessment of exer-
cise frequency is unlikely to be problematic, though discrepancies between interview and self-
report formats in the assessment of motivation for exercise may be worthy of investigation
(Hubbard, Gray, & Parker, 1998).

4.3. Criterion validity

To our knowledge, this is the first study to consider the criterion validity of the EDE-Q. Dif-
ferences in mean scores between cases and non-cases were highly significant for each of the four
subscales and of similar magnitude for each subscale. The fact that the EDE-Q item ‘flat stom-
ach’ did not discriminate significantly between groups is likely to reflect a ceiling effect with
respect to the endorsement of this item, in that more than one third of subjects indicated a defi-
nite desire to have a flat stomach every day. Given the current fashion not only for leanness,
but for body tone (Gordon, 2000), it may be necessary to revise the wording of this question in
future versions of the EDE-Q. The failure of the EDE item ‘avoidance of eating’ to discriminate
between groups reflects the fact that only four individuals received a rating other than ‘0’ on
this item and underscores the point made previously that items addressing the more extreme
forms of weight control behaviour may not be useful in identifying those cases of eating dis-
orders typically encountered in general population surveys.
Results of the discriminant function analysis showed considerable overlap with those of
Beglin and Fairburn’s (1992) study. Of note is that the EDE-Q item addressing frequency of
OBEs provided information useful in discriminating cases of clinically significant eating dis-
orders from non-cases in both studies, notwithstanding those discrepancies between self-report
and interview assessment of binge eating referred to above. In contrast to the findings of Beglin
and Fairburn, however, items addressing the frequency of self-induced vomiting and laxative
abuse did not enter the stepwise model in the present study, while the use of exercise as a means
of weight control was included. These differences may be largely a product of differences in the
prevalence of purging behaviours in the different studies. Of 243 subjects interviewed in Beglin
and Fairburn’s (1992) study, 14 (5.8%) reported current use of self-induced vomiting or lax-
atives on the EDE, while in the present study only two individuals out of a total of 208 subjects
interviewed (1.0%) reported the use of these behaviours. This discrepancy may in turn reflect the
different age profiles of subjects in the respective studies, since BED is relatively more common,
and BN relatively less common, among women aged 35–45 (Johnson, Spitzer, & Williams,
2001). In addition, criteria for a case were expanded to include individuals with recurrent SBEs,
and a more liberal criterion for the use of exercise as a compensatory behaviour was employed.
Black and Wilson (1996) reported that, in a sample of female substance abusers, scores on the
global EDE-Q scale were significantly correlated with the total discrepancy between EDE-Q and
EDE scores, suggesting that the EDE-Q became less reliable as eating disorder symptoms
became more frequent. While this finding was replicated in the present study when the total
sample was considered, the discrepancy between EDE-Q and EDE scores was greater among
non-cases than among cases. In fact, EDE scores were higher than EDE-Q scores for a number
of subjects in the case group. On the other hand, the correlation between scores on the global
scales of the respective measures was lower among cases than among non-cases. One possible
interpretation of these findings is that symptomatic individuals are more likely to underreport
564 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567

eating disordered attitudes and behaviours when assessed by self-report, either through lack of
insight or conscious dissimulation or both (Vitousek, Daly, & Heiser, 1991). The higher mean
frequencies of binge eating and self-induced vomiting reported for interview assessment in clini-
cal samples is consistent with this interpretation (Carter, Aime, & Mills, 2001; Wilfley Schwartz,
Spurrell, & Fairburn, 1997).
Scores on both the EDE-Q and EDE were positively correlated with BMI in the present
study. Further, BMI was markedly elevated among individuals who scored at or above the 90th
percentile of EDE-Q scores, but who were non-cases (n ¼ 10). For this reason, the 22 items of
the EDE-Q from which subscale and global scores are derived, performed poorly in discriminat-
ing cases from non-cases in the present study. Similarly, the low PPV of the EAT in general
population surveys has been attributed to a preponderance of overweight dieters among individ-
uals scoring above the recommended cut-off point (King, 1989; in Patton & Szmukler, 1995). A
recently developed measure was found to have a PPV of 0.24 in a sample of general practice
attenders (Luck et al., 2002). To some extent, these findings reflect the inherent difficulty of
screening for low-prevalence psychiatric disorders in community samples. While the prevalence
of partial-syndrome eating disorders is higher than that of either AN or BN, it has been shown
that for disorders with a prevalence of less than 10%, PPV may be less than 0.5, even when sen-
sitivity and specificity are near perfect (Williams, Hand, & Tarnopolsky, 1982). In the present
study, the PPV of the EDE-Q improved substantially—from 0.30 to 0.56—when items address-
ing the occurrence of OBEs and/or exercise for weight or shape reasons were specified as cri-
teria in addition to the global scale score.
Findings relating to criterion validity must be interpreted with caution on account of the
small number of cases on which this analysis was based and given that validity coefficients may
vary across samples, even among different sub-groups in the same population (Goldberg, 1981;
Szmukler, 1985). However, both the total sample size and the number of cases were comparable
to those in the studies of Beglin and Fairburn (1992) and Fairburn and Beglin (1994). The fact
that data relating to criterion validity were based on a second administration of the EDE-Q
may also limit the generalizability of the present findings. Post-hoc analysis indicated a small
decrease (0.2) on the global scale of the EDE-Q between first and second administrations
among individuals identified as cases. While this difference was not significant, a slightly higher
threshold than that suggested by the ROC analysis may be appropriate. It should also be noted
that the findings of the present study were based on a sub-group of participants who completed
both measures. While individuals interviewed were similar to those not interviewed with respect
to eating disorder psychopathology and BMI, some 220 women (31%) chose not to participate
in the study and only information concerning age was available for these individuals. The extent
to which the study sample was representative of the total phase one sample is therefore
unknown.
Several other self-report instruments have recently been developed (Ghaderi & Scott, 2002;
Luck et al., 2002; Rosenvinge et al., 2001; Stice, Telch, & Rizvi, 2000). However, we are aware
of only one measure with demonstrated validity in a community sample (Stice, Telch & Rizvi,
2000). An additional advantage of the EDE-Q is that normative data are available for both gen-
eral population and clinical samples, while such data are not available for more recently
developed measures. Norms for the EDE-Q among adolescent girls have also been established
 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567 565

(Carter, Stewart, & Fairburn, 2001) and use of EDE-Q in this population is likely to increase
(Passi, Bryson, & Lock, 2002).
In conclusion, the EDE-Q is a useful measure of eating disorder psychopathology, showing a
high level of agreement with the EDE interview in the assessment of attitudinal features. Prob-
lems with the assessment of binge eating and with low PPV remain. However, the EDE-Q items
addressing occurrance and frequency of OBEs appear to discriminate well between cases and
non-cases diagnosed according to the EDE and PPV appears to be superior to that of other
self-report measures. The instrument appears ideally suited for use in two-phase epidemiological
studies in which individuals screening positive are to be followed up at successive time periods.

Acknowledgements

The research was conducted while the first author was in receipt of a Research Training Fel-
lowship from the NSW Institute of Psychiatry.

References
American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders. (4th ed.). Wash-
ington, DC: APA ((DSM-IV)).
Anderson, D. A., & Williamson, D. A. (2002). Outcome measurement in eating disorders. In W. W. IsHak, & T.
et al. (Eds.), Outcome measurement in psychiatry: A critical review (pp. 289–301). Washington, DC: American
Psychiatric Publishing Inc.
Beglin, S. J., & Fairburn, C. G. (1992). Evaluation of a new instrument for the detection of eating disorders in com-
munity samples. Psychiatry Research, 44, 191–201.
Black, C. M. D., & Wilson, G. T. (1996). Assessment of eating disorders: interview versus questionnaire. Inter-
national Journal of Eating Disorders, 20, 43–50.
Carter, J. C., Aime, A. A., & Mills, J. S. (2001). Assessment of bulimia nervosa: a comparison of interview and self-
report questionnaire methods. International Journal of Eating Disorders, 30, 187–192.
Carter, J. C., Stewart, D. A., & Fairburn, C. G. (2001). Eating Disorder Examination Questionnaire: norms for
young adolescent girls. Behaviour Research and Therapy, 39, 625–632.
Dunn, G., Pickles, A., Tansella, M., & Vazquez-Barquero, J. L. (1999). Two-phase epidemiological surveys in psychi-
atric research. British Journal of Psychiatry, 174, 95–100.
Evans, L., & Wertheim, E. H. (2002). An examination of willingness to self-disclose in women with bulimic symp-
toms considering the context of disclosure and negative affect levels. International Journal of Eating Disorders, 31,
344–348.
Fairburn, C. G., & Beglin, S. J. (1994). Assessment of eating disorders: interview or self-report questionnaire?. Inter-
national Journal of Eating Disorders, 16, 363–370.
Fairburn, C. G., & Cooper, Z. (1993). The eating disorder examination. In C. G. Fairburn, & G. T. Wilson (Eds.),
Binge eating: Nature, assessment and treatment (12th ed.) (pp. 317–360). New York: Guilford Press.
Garfinkel, P. E., Lin, E., Goering, P., Spegg, C., Goldbloom, D. S., Kennedy, S., Kaplan, A. S., & Woodside, D. B.
(1995). Bulimia nervosa in a Canadian community sample: prevalence and comparison of subgroups. American
Journal of Psychiatry, 152, 1052–1058.
Garner, D. M. (1991). Eating disorder inventory-2: Professional manual. Odessa, FL: Psychological Assessment
Resources.
Garner, D. (2002). Measurement of eating disorder psychopathology. In C. G. Fairburn, & K. D. Brownell (Eds.),
Eating disorders and obesity: A comprehensive handbook (2nd ed.) (pp. 141–146). Guilford Press: New York.
566 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567

Garner, D. M., & Garfinkel, P. E. (1979). The Eating Attitudes Test: an index of the symptoms of anorexia nervosa.
Psychological Medicine, 9, 273–279.
Garner, D. M., Olmsted, M. P., Bohr, Y., & Garfinkel, P. E. (1982). The Eating Attitudes Test: psychometric fea-
tures and clinical correlates. Psychological Medicine, 12, 871–878.
Garner, D. M., Olmsted, M. P., & Polivy, J. (1983). Development and validation of a multidimensional eating dis-
order inventory for anorexia nervosa and bulimia. International Journal of Eating Disorders, 2, 15–34.
Garrow, J. S. (1988). Obesity and related diseases. Edinburgh: Churchill Livingstone.
Ghaderi, A., & Scott, B. (2002). The preliminary reliability and validity of the survey for eating disorders (SEDs): a
self-report questionnaire for diagnosing eating disorders. European Eating Disorders Review, 10, 61–76.
Goldberg, D. (1981). Estimating the prevalence of psychiatric disorder from the results of a screening test.
In J. K. Wing, P. Bebbington, & L. N. Robins (Eds.), What is a case? (pp. 129–136). London: Grant McIntyre.
Gordon, R. A. (2000). Eating disorders: Anatomy of a social epidemic. (2nd ed.). Oxford, UK: Blackwell.
Grilo, C. M., Masheb, R. M., & Wilson, G. T. (2001a). A comparison of different methods for assessing the features
of eating disorders in patients with binge eating disorder. Journal of Consulting and Clinical Psychology, 69, 317–
322.
Grilo, C. M., Masheb, R. M., & Wilson, G. T. (2001b). Different methods for assessing the features of eating dis-
orders in patients with binge eating disorder: a replication. Obesity Research, 9, 418–422.
Hay, P. J., & Fairburn, C. G. (1998). The validity of the DSM-IV scheme for classifying bulimic eating disorders.
International Journal of Eating Disorders, 23, 7–15.
Hay, P. J., Fairburn, C. G., & Doll, H. (1996). The classification of bulimic eating disorders: a community based
cluster-analytic study. Psychological Medicine, 26, 801–812.
Hay, P. J., Marley, J., & Lemar, S. (1998). Covert eating disorders: the prevalence, characteristics and help-seeking of
those with bulimic eating disorders in general practice. Primary Care Psychiatry, 4, 95–99.
Hubbard, S. T., Gray, J. G., & Parker, S. (1998). Differences among women who exercise for ‘food related’ and ‘non-
food related’ reasons. European Eating Disorders Review, 6, 255–265.
Johnson, J. G., Spitzer, R. L., & Williams, J. B. W. (2001). Health problems, impairment and illnesses associated
with bulimia nervosa and binge eating disorder among primary care and obstetric gynaecology patients. Psycho-
logical Medicine, 31, 1455–1466.
Kalarchian, M. A., Wilson, G. T., Brolin, R. E., & Bradley, L. (2000). Assessment of eating disorders in bariatric
surgery candidates: self-report questionnaire versus interview. International Journal of Eating Disorders, 28, 465–
469.
Kessler, R. C., & Mroczek, D. (1994). Final versions of our non-specific psychological distress scale (memo 10/3/94).
Ann Arbor: Survey Research Centre of the Institute for Social Research, University of Michigan.
King, M.B. (1989). Eating disorders in a general practice population. Prevalence, characteristics and follow-up at 12
to 18 months. Psychological Medicine Monograph Supplements, 14, 1–34.
Luck, A. J., Morgan, J. F., Reid, F., O’Brien, A., Brunton, J., Price, C., Perry, L., & Lacey, J. H. (2002). The
SCOFF questionnaire and clinical interview for eating disorders in general practice: comparative study. British
Medical Journal, 325, 755–756.
Luce, K. H., & Crowther, J. H. (1999). The reliability of the eating disorder examination-self-report questionnaire
version (EDE-Q). International Journal of Eating Disorders, 25, 349–351.
Meadows, G. N., Palmer, R. L., Newball, E. U. M., & Kendrick, J. M. T. (1986). Eating attitudes and disorder in
young women: a general practice based survey. Psychological Medicine, 16, 351–357.
Mond, J. M., Hay, P. S., Rodgers, B., Owen, C. A., & Beumont, P. J. V. Criteria for ‘excessive exercise’ in eating dis-
orders research: when is exercise behaviour ‘clinically significant’? (manuscript submitted for publication).
Passi, V. A., Bryson, S. W., & Lock, J. (2002). Assessment of eating disorders in adolescents with anorexia nervosa:
self-report questionnaire versus interview. International Journal of Eating Disorders, 33, 45–54.
Patton, G. C., & Szmukler, G. I. (1995). Epidemiology of eating disorders. In A. Jablensky (Ed.), Epidemiological
psychiatry (pp. 307–328). London: Balliere Tindall.
Pike, K. M., Dohm, F. A., Striegel-Moore, R. H., Wilfley, D. E., & Fairburn, C. G. (2001). A comparison of black
and white women with binge eating disorder. American Journal of Psychiatry, 158, 1455–1460.
Pratt, E. M., Niego, S. H., & Agras, W. S. (1998). Does the size of a binge matter?. International Journal of Eating
Disorders, 24, 307–312.
 J.M. Mond et al. / Behaviour Research and Therapy 42 (2004) 551–567 567

Rosenvinge, J. H., Perry, J. A., Bjorgum, L., Bergensen, T. D., Silvera, D. H., & Holte, A. (2001). A new instrument
measuring disturbed eating patterns in a community populations: development and initial validation of a five-item
scale (EDS-5). European Eating Disorders Review, 9, 123–132.
Steinhausen, H., Winkler, C., & Meier, M. (1997). Eating disorders in adolescence in a Swiss epidemiological study.
International Journal of Eating Disorders, 22, 147–151.
Stice, E., Telch, C. F., & Rizvi, S. L. (2000). Development and validation of the eating disorder diagnostic scale: a
brief self-report measure of anorexia, bulimia and binge-eating disorder. Psychological Assessment, 12, 123–131.
Szmukler, G. I. (1985). The epidemiology of anorexia nervosa and bulimia. Journal of Psychiatric Research, 19, 143–
153.
Tabachnick, B. G., & Fidell, L. S. (1996). Using multivariate statistics. (3rd ed.). New York: Harper Collins.
Vitousek, K. B., Daly, J., & Heiser, C. (1991). Reconstructing the internal world of the eating-disordered individual:
overcoming denial and distortion in self-report. International Journal of Eating Disorders, 10, 647–666.
Walters, E. E., & Kendler, K. S. (1995). Anorexia nervosa and anorexic-like symptoms in a population-based female
twin sample. American Journal of Psychiatry, 152, 64–71.
Ware, J. E., Kosinski, M., & Keller, S. D. (1996). A 12-item short-form health survey: construction of scales and pre-
liminary tests of reliability and validity. Medical Care, 34, 220–233.
Wilfley, D. E., Schwartz, M. B., Spurrell, E. B., & Fairburn, C. G. (1997). Assessing the specific psychopathology of
binge eating disorder patients: interview or self-report?. Behaviour Research and Therapy, 35, 1151–1159.
Williams, P., Hand, D., & Tarnopolsky, A. (1982). The problem of screening for uncommon disorders—a comment
on the Eating Attitudes Test. Psychological Medicine, 12, 431–434.
Wilson, G. T. (1993). Assessment of binge eating. In C. G. Fairburn, & G. T. Wilson (Eds.), Binge eating: Nature,
assessment and treatment (pp. 227–249). New York: Guilford Press.