See

discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/238274320

Development and validation

of a multidimensional eating
disorder inventory of anorexia
nervosa and...

Article in International Journal of Eating Disorders January 1983

DOI: 10.1002/1098-108X(198321)2:23.0.CO;2-6

CITATIONS READS

2,051 42,070

3 authors, including:

David M. Garner
River Centre Clinic
178 PUBLICATIONS 17,408 CITATIONS

SEE PROFILE

All content following this page was uploaded by David M. Garner on 01 March 2017.

The user has requested enhancement of the downloaded file.

 DEVELOPMENT AND VALIDATION OF A
MULTIDIMENSIONAL EATING DISORDER
INVENTORY FOR ANOREXIA NERVOSA
AND BULIMIA

David M. Garner, Ph.D., Marion P. Olmstead, M.A.,

Janet Polivy, Ph.D.

ABSTRACT

The development and validation of a new measure, the Eating Disorder In-
ventory (EDI) is described. The EDI is a 64 item, self-report, multiscale meas-
ure designed for the assessment of psychological and behavioral traits
common in anorexia nervosa (AN) and bulimia. The EDI consists of eight sub-
scales measuring: 1) Drive for Thinness, 2) Bulimia, 3) Body Dissatisfaction,
4) Ineffectiveness, 5) Perfectionism, 6) Interpersonal Distrust, 7) Interocep-
tive Awareness and 8) Maturity Fears. Reliability (internal consistency) is
established for all subscales and several indices of validity are presented.
First, AN patients (N = 113) are differentiated from female comparison (FQ
subjects (N = 577) using a cross-validation procedure. Secondly, patient self-
report subscale scores agree with clinician ratings ofsubscale traits. Thirdly,
clinically recovered AN patients score similarly to FCs on all subscales.
Finally, convergent and discriminant validity are established for subscales.
The EDI was also administered to groups of normal weight bulimic women,
obese, and normal weight but formerly obese women, as well as a male com-
parison group. Group differences are reported and the potential utility of the
EDI is discussed.
I
INTRODUCTION

During the past decade, anorexia nervosa and more recently bulimia in
normal-weight women have been the focus of growing attention from the
public sector and various health disciplines. There is a consensus that they are
no longer rare disorders, but rather are prevalent with a marked
preponderance in young women. A significant morbidity and mortality have
been associated with anorexia nervosa while the risks of bulimia in normal-
weight women have not yet been well documented. Attempts to objectively

David M. Garner, Ph.D. is Associate Professor, Department of Psychiatry and Psychology, University of
Toronto, and Coordinator of Research, Department of Psychiatry, Toronto General Hospital; Marion P.
Olmstead, M.A. is a Doctoral Candidate in Clinical Psychology, York University, Toronto; Janet Polivy,
Ph.D. is Associate Professor, Department of Psychiatry and Psychology, University of Toronto. Please
address reprint requests and correspondence to: Dr. David M. Gamer. Department of Psychiatry, Toronto
General Hospital, 101 College Street, Toronto, Ontario M5G 1L7.
I

INT'LIOURNAL OF EATING DISORDERS/VOL.2 NO.2/WINTER 1983 15

Multidimensional Eating Disorder Inventory

measure response to treatment have resulted in the development of several

measures of the symptoms of these eating disorders.
Slade (1973) originally proposed an observer-rating scale for assessing three
dimensions of anorexic behavior in an inpatient setting. A self-report measure
of the symptoms of anorexia nervosa, the Eating Attitudes Test (EAT), was
developed by Garner and Garfinkel (1979), and has been recently abbreviated
based on a factor analysis (Garner, Olmsted, Bohr and Garfinkel, 1982). The
EAT has been used as a screening instrument for detecting cases of anorexia
nervosa in groups at high risk for the disorder (Garner and Garfinkel, 1980) as
well as identifying abnormal eating patterns among college students (Button
and Whitehouse, 1981; Thompson and Schwartz, 1982).
Goldberg, Halmi, Eckert, Casper, Davis and Roper (1980) devised a 63-item
measure of anorexic attitudes and behavior for use in an inpatient setting. A
factor analysis of the Goldberg et al. (1980) scale revealed 18 factors; however,
8 of these were formed by only one or two items. Four factors were correlated
with rate of weight gain during 35 days of inpatient treatment, although the
association was weak and it may be argued that rate of weight gain in hospital
may be unrelated to ultimate outcome. Finally, Fichter and Kesser (1980) have
developed a self-rating scale for assessing several aspects of anorexic behavior;
however, validation was not reported for English-speaking subjects.
Questionnaires have also been proposed for measuring the symptoms of
bulimia. Hawkins and Clement (1980) developed a "binge eating scale" and
reported that over two-thirds of females and one-half of males in a college
sample engage in episodes of "uncontrolled" excessive eating. Self-induced
vomiting was reported by 3.5% of the females. Halmi, Falk and Schwartz
(1981) found that 13% of a college sample indicated all of the symptoms of
bulimia (87% of these were female and 13% were male), while 10% reported
self-induced vomiting (12% female; 6% male).
While recent measures permit objective assessment of symptom areas, they
either have the disadvantage of being suitable only for inpatient administra-
tion (Slade, 1973; Goldberg et al., 1980), or they tend to be oriented toward
behavioral/symptom parameters of anorexia nervosa (Gamer and Garfinkel,
1979) or bulimia (Halmi et al., 1981; Hawkins and Clement, 1980). They do
not tap psychological dimensions which have been postulated to be more fun-
damentally related to anorexia nervosa and bulimia.
The aim of the present study is to report on the development and validation
of the Eating Disorder Inventory (EDI) which is a multifaceted instrument
designed to assess psychological characteristics relevant to anorexia nervosa
and bulimia. Justification for the development of the EDI is based on the grow-
ing recognition that anorexia nervosa is a multidimensional disorder with con-
siderable psychological variability across the heterogeneous patient popula-
tion (Beumont et al., 1981; Dally and Gomez, 1979; Garfinkel and Gamer,
1982; Garner and Garfinkel, 1980; Palmer, 1979; Strober, 1980; 1981; in
press). There has been particular interest in contrasting subgroups of anorexic
patients who exclusively restrict their dietary intake with those who ex-

,, INT'LIOURNALOF EATING DISORDERS/VOL.2 NO.2/WINTER 1983

 Multidimensional Eating Disorder Inventory

perience episodes of bulimia followed by self-induced vomiting and purgation

(Beumont, George and Smart, 1976; Casper, Eckert, Halmi, Goldberg and
Davis, 1980; Garfinkel, Moldofsky and Gamer, 1980; Russell, 1979; Strober,
1981). Strober (in press) has identified homogeneous subtypes of anorexia ner-
vosa based on clinical symptomatology and personality features derived from
MMPI profiles. Delineation and more precise measurement of psychological
traits differentiating subgroups may have relevance to the understanding and
treatment of both anorexia nervosa and bulimia. Moreover, identifying
distinct psychological typologies may reveal features which discriminate in-
dividuals with anorexia nervosa from those who display significant symptoms
of the disorder, but who may be less psychologically disturbed. The EDI was
devised rather than relying on existing personality measures, since conven-
tional tests do not adequately address the cognitive and behavioral character-
istics often observed clinically in anorexia nervosa.

METHODS

Questionnaire Construction

A large pool of items was generated by clinicians who were both familiar
with the research literature on anorexia nervosa and who had experience
treating patients with the disorder. Items were designed to measure eleven con-
structs; however, only eight of these dimensions met the reliability and valid-
ity requirements for the scale. A short description of the intended item content
of each of the retained eight subscales and the clinical sources from which they
were derived are presented below.
Drive for Thinness * indicates excessive concern with dieting, preoccupa-
tion with weight and entrenchment in an extreme pursuit of thinness. Bruch
(1973, 1978) and others have described this as a cardinal feature of anorexia
nervosa. Items reflect both an ardent wish to lose weight as well as a fear of
weight gain.
Bulimia* indicates the tendency toward episodes of uncontrollable over-
eating (bingeing) and may be followed by the impulse to engage in self-induced
vomiting. The presence or absence of bulimia differentiates subtypes of
anorexia nervosa (Beumont et al., 1976; Russell, 1979; Casper et al., 1980;
Garfinkel et al., 1980) and has been described in women with no prior history
of anorexia nervosa (Pyle, Mitchell and Eckert, 1981; Russell, 1979; Johnson et
*A recent factor analysis of the EAT (Gamer et al., 1982) has revealed two-item clusters of similar content to
that of the EDI Drive for Thinness and Bulimia subscales. While the correlations between the respective similar
subscales is relatively high (Drive for Thinness with EAT "Dieting," N = 18, r=.8O; Bulimia with EAT
"Bulimia and Food Preoccupation," N = 18, r = .85), a substantial amount of the variance between scales is not
shared. Examination of the item content of the similar scales indicates that the EAT factors are broader in focus
than those of the EDI. Furthermore, the subscales of the EDI were theoretically or deductively derived (fol-
lowed by empirical refinement and validation) while the EAT factors were empirically or inductively derived
from an initial pool of items reflecting symptoms of anorexia nervosa. Despite some conceptual and possibly
predictive overlap, the EDI is not intended as a replacement for the EAT. The EAT is a sound measure of a
range of symptoms common in anorexia nervosa while the EDI focuses more on the specific cognitive and
behavioral dimensions which may meaningfully differentiate subgroups of patients, or which may distinguish
those with serious psychopathology from extreme "dieters."

INT'DOURNALOF EATING DISORDERS/VOL.2 NO.2/WINTER 1983 '^

Multidimensional Eating Disorder Inventory

al., in press). Recent studies have found that bulimia is relatively common
among college females (Wardle, 1980; Halmi et al., 1981; Hawkins and Cle-
ment, 1980); however, these studies have not employed well-standardized
measures.
Bod\/ Dissatisfaction reflects the belief that specific parts of the body asso-
ciated with shape change or increased "fatness" at puberty are too large (e.g.
hips, thighs, buttocks). Body dissatisfaction has been found to be related to
other body image disturbances which have been considered a basic deficit in
anorexia nervosa (see Garner and Garfinkel, 1981 for a review). Crisp (1977,
1980) has suggested that dieting in anorexia nervosa is a response to dissatis-
faction with pubertal "fatness" and the symbolic meaning that it has for the in-
dividual.
Ineffectiveness assesses feelings of general inadequacy, insecurity, worth-
lessness and the feeling of not being in control of one's, life. This feature has
been described by some as the fundamental disturbance in anorexia nervosa
(Bruch, 1973; Selvini-Palazzoli, 1978; Strober, 1980; Wingate and Christie,
1978). While there have been attempts to operationalize this construct in terms
of locus of control (Garner et al., 1976; Hood et al., 1982), it has been sug-
gested that the concept of ineffectiveness also includes a negative self-evalua-
tion (self-concept) component which is not addressed by locus of control
(Garner et al., 1982).

Perfectionism indicates excessive personal expectations for superior

achievement. Bruch (1978) has suggested that the struggle to live up to perfec-
tionistic achievement standards is a characteristic theme in anorexia nervosa.
She interprets the patient's typical superior academic performance as an "over-
compliant adaptation" which breaks down in the face of increasing pressures
to succeed. We have described the perfectionism in anorexia nervosa as part of
a "dichotomous" thinking style (Garner et al., 1982). The families in which
anorexia nervosa occurs have been discussed as highly achievement-oriented
(Kalucy et al., 1977; Dally, 1969; Bruch, 1973) and thus may magnify our cul-
ture's emphasis on success.

Interpersonal Distrust reflects a sense of alienation and a general reluctance

to form close relationships and has been identified as important in the develop-
ment and maintenance of anorexia nervosa (Selvini-Palazzoli, 1978; Goodsitt,
1969,1977; Story, 1977; Strober, 1980). It is to be distinguished from paranoid
thinking and relates to an inability to form attachments or feel comfortable ex-
pressing emotions toward others.
Interoceptive Awareness reflects one's lack of confidence in recognizing and
accurately identifying emotions and sensations of hunger or satiety. Bruch
(1962, 1978) and Selvini-Palazzoli (1978) have described this deficiency in in-
teroceptive labelling as fundamental to anorexia nervosa and there is some em-
pirical support for the existence of deficits in this area (Garfinkel and Gamer,
1982).

INT'L lOURNAL OF EATING DISORDERS/VOL.2 NO.2 / WINTER 1983

 Multidimensional Eating Disorder Inventory

Maturity Fears measures one's wish to retreat to the security of the preado-
lescent years because of the overwhelming demands of adulthood. Crisp
(1965, 1980) has suggested that the central psychopathology of anorexia ner-
vosa is an avoidance of psychological maturity through the mechanism of car-
bohydrate avoidance.

Scoring

The test format is similar to that of the EAT (Garner and Garfinkel, 1979)
where Ss respond to six point, forced choice items by rating whether each item
applies "always," "usually," "often," "sometimes," "rarely," or "never." The
scoring is identical to the EAT with the most extreme "anorexic" response
(always or never depending on the keyed direction) earning a score of 3; the
immediately adjacent response 2, the next response 1 and the three choices op-
posite to the most "anorexic" response receiving no score (0). Scale scores are
the summation of all item scores for that particular scale (see the appendix for
instructions to Ss and sample question format).

TABLE 1

Demographic Characteristics (Means) of Anorexia Nervosa and

Comparison Samples

N AGE HT. WEIGHT PERCENT OF

(IN,) (LBS,) AVERAGE
WEIGHT
Total Anorexia
Nervosa 113 21,8 64,1 102,4 80.0
Restricters 48 21,0 63,9 91,9 72,0
Bulimics 65 22,4 64,3 110,1 85,9
Female
Comparison 577 19,9 64,9 124,1 95,5
Male Comparison 166 20.3 70.4 161,1 103,1
Normal Weight 195 20.8 63,9 123,6 95,1
Bulimics
Obese 44 32,5 63,6 171,9 129,9
Formerly Obese 52 36,7 64,5 137,9 99,5
Recovered Anorexics 17 23,9 62,6 108,2 86,9

Subjects

Two groups of Ss participated in the cross-validation of the EDI. The cri-

terion group consisted of three subsamples of female primary anorexia ner-
vosa (AN) patients (total N=113) seen in consultation at the Clarke Institute

INT'LIOURNALOFEATINGOISORDERS/VOL.2 NO.2/WINTER 1983 19

Multidimensional Eating Disorder Inventory

of Psychiatry. The AN group met a modified version of the Feighner et al.

(1972) diagnostic criteria (Garfinkel and Garner, 1982) and was heterogeneous
in that patients were at various stages of treatment when tested, although none
could be considered recovered. At the time of testing they averaged 20%
below expected weight for their age and height according to norms from
Health and Welfare Canada (1954). Approximately one half of the AN sub-
jects were of the "restricter" subtype and the remainder had the complication
of "bulimia." There were significant differences in percent of average weight
(t=6.32, p<0.0001) between the bulimic and restricter subgroups,
although there were no differences in age or the duration of illness (Table 1).
The female comparison group (FC) consisted of three independent subsamples
of female university students (total N = 577) from first and second year
psychology courses. They were tested in their normal class sessions and
although their participation was voluntary, virtually all of the subjects ap-
proached completed questionnaires. Demographic characteristics for the AN
and FC samples are presented in Table 1.

RESULTS

Item Selection and Cross-Validation

The first two independent AN and FC samples were used to select items
from the original pool. Retained items had to meet two statistical criteria.
First, they had to demonstrate validity by significantly differentiating between
the AN and FC groups. Secondly, items had to be more highly correlated with
the subscale to which they were intended to belong than to any other subscale.
Items in the original pool were evaluated independently in each of the first two
AN and FC samples; thus the validity and homogeneity requirements for re-
tained items were replicated. After the questionnaire had been administered to
the first AN and FC sample, it was necessary to generate additional items for
some subscales.* These items were evaluated on the second and third AN and
FC samples. It is important to note that the item content of subscales was
essentially finalized on the second AN and FC samples; although we were pre-
pared to make minor revisions on the third trial, this turned out to be unneces-
sary. The third cross-validation samples merely confirmed the utility of the
previously selected items.
A final requirement was that subscales have coefficients of internal consis-
tency (Cronbach's alpha) above .80 for the AN samples. Although item-scale
correlation coefficients above .40 for the AN groups were considered
desirable, three items with item-total correlations below .40 were retained
because they were considered conceptually important. The average item-total
Additional items were derived for the Interoceptive Awareness and Maturity Fears subscales. Since these
items were administered only to the second and third AN and FC subsamples, final N s for these two subscales
are smaller than for the other six subscales.

20 INT'L JOURNAL OF EATING DISORDERS/VOL.2 NO.2 / WINTER 1983

 Multidimensional Eating Disorder Inventory

correlation was .63 (S.D. = 0.13) indicating substantial within-scale common

variance among items. Since there were no significant differences within
groups across the three cross-validation trials, samples have been collapsed for
presentation here. The final subscale items, item analysis results and reliability
coefficients for the AN and FC groups are presented in Table 2.

TABLE 2

DRIVE FOR THINNESS: Subscale Items, Item-Total Correlations,

Reliability Coefficients for AN and FC groups.
Item Item on Subscale Item-Total
Number Correlation
AN FC
1.* I eat sweets and carbohydrates without feeling .32 .37
nervous.
7. I think about dieting. .76 .66
11. I feel extremely guilty after overeating. .58 .63
16. I am terrified of gaining weight. .76 .67
25. I exaggerate or magnify the importance of .43 .59
weight.
32. I am preoccupied with the desire to be thinner. .66 .74
49. If I gain a pound, I worry that I will keep .70 .66
gaining.
Reliability coefficients (Standardized Cronbach's Alphas) AN = .85 FC = .85

INTEROCEPTIVE AWARENESS: Subscale Items, Item-Total Correlations,

Reliability Coefficients for AN and FC groups

8. I get frightened when my feelings are too .40 .50

too strong.
21. I get confused about what emotion I am .74 .51
feeling.
26.* I can clearly identify what emotion I am .63 .19
feeling.
33. I don't know what's going on inside me. .54 .53
40. I get confused as to whether or not I am .49 .04
hungry.
44. I worry that my feelings will get out of .73 .48
control.
47. I feel bloated after eating a small meal. .22 .06
51. When I am upset, I don't know if I am sad, .56 .23
frightened or angry.
60. I have feelings I can't quite identify. .75 .51
64. When I am upset, I worry that I will start .43 .19
eating.
Coefficients of Reliability (Standardized Cronbach's Alphas) AN = .85FC = .66

* indicates negatively keyed item.

INT'LlOURNALOF EATING DISORDERS/ VOL.2 NO.2/ WINTER 1983

Multidimensional Eating Disorder Inventory

TABLE 2
BULIMIA: Subscale Items, Item-Total Correlations, Reliability
Coefficients for AN and FC groups
Item Item on Subscale Item-Total
Number Correlation
AN FC_
4. I eat when I am upset. 58 .51
5. I stuff myself with food. 79 .53
28.1 have gone on eating binges where I have felt 69 .64
that I could not stop.
38.1 think about bingeing (overeating). 74 .60
46.1 eat moderately in front of others and stuff 75 .67
myself when they're gone.
53.1 have the thought of trying to vomit in order 70 .42
to lose weight.
61.1 eat or drink in secrecy. 73 .60
Reliability Coefficients (Standardized Cronbach's Alphas) AN = .90 FC = .83

BODY DISSATISFACTION: Subscale Item, Item-Total Correlations,

Reliability Coefficients for AN and FC groups.

2. I think that my stomach is too big. .51 .51

9. I think that my thighs are too large. .69 .68
12.* I think that my stomach is just the right size. .66 .58
19.* I feel satisfied with the shape of my body. .50 .65
31.* I like the shape of my buttocks. .69 .68
45. I think my hips are too big. .78 .75
55.* I think that my thighs are just the right size. .73 .78
59. I think by buttocks are too large. .83 .73
62.* I think that my hips are just the right size. .70 .78
Reliability Coefficients (Standardized Cronbach's Alphas) AN = .90 FC = .91

INEFFECTIVENESS: Subscale Items, Item-Total Correlations, Reliability

Coefficients for AN and FC groups

10. I feel ineffective as a person. .70 .53

18. I feel alone in the world. .62 .54
20.* I feel generally in control of things in my life. .65 .59
24. I wish I were someone else. .51 .49
27. I feel inadequate. .76 .61
37.* I feel secure about myself. .75 .66
41. I have a low opinion of myself. .74 .63
42.* I feel that I can achieve my standards. .52 .50
50.* I feel that I am a worthwhile person. .79 .59
56. I feel empty inside (emotionally). .52 .62
Reliability Coefficients (Standardized Cronbach's Alphas) AN=.9O FC = .86

* indicates negatively keyed item.

22 INT'L JOURNAL OF EATING DISORDERS/VOL.2 NO.2 / WINTER 1983

 Multidimensional Eating Disorder Inventory

TABLE 2
MATURITY FEARS: Subscale Items, Item-Total Correlations,
Reliability Coefficients for AN and FC groups
Item Item on Subscale Item-Total
Number Correlation
AN FC
3. I wish that I could return to the security of .72 .12
childhood.
6. I wish that I could be younger. .66 .16
14. The happiest time in life is when you are a .67 .40
child.
22.* I would rather be an adult than a child. .63 .63
35. The demands of adulthood are too great. .27 .25
39.* I feel happy that I am not a child anymore. .82 .62
48. I feel that people are happiest when they are .80 .28
children.
58.* The best years of your life are when you .72 .42
become an adult.
Reliabilitj 1 Coefficients (Standardized Cronbach's Alphas) AN = .88 FC= .65

PERFECTIONISM: Subscale Items, Item-Total Correlations, Reliability

Coefficients for AN and FC groups

13. Only outstanding performance is good

enough in my family. .66 .46
29. As a child, I tried very hard to avoid
disappointing my parents and teachers. .43 .39
36. I hate being less than best at things. .58 .51
43. My parents have expected excellence of me. .63 .52
52. I feel that I must do things perfectly or not do
them at all. .63 .45
63. I have extremely high goals. .58 .41
Reliability Coefficients (Standardized Cronbach's Alphas) A N = .82 FC = .73

INTERPERSONAL DISTRUST: Subscale Items, Item-Total Correlations,

Reliability Coefficients for AN and FC groups

15.* I am open about my feelings. .76 .58

17.* I trust others. .42 .32
23.* I can communicate with others easily. .67 .41
30.* I have close relationships. .56 .49
34. I have trouble expressing my emotions to .60 .53
others.
54. I need to keep people at a certain distance (feel .61 .45
uncomfortable if someone tries to get too
close).
57.* I can talk about personal thoughts or feelings. .71 .62
Reliability Coefficients (Standardized Cronbach's Alphas) AN = .85 FC = .76

* indicates negatively keyed item.

INT'LIOURNAL OF EATING DISORDERS/ VOL.2 NO.2 / WINTER 1983 23

 Multidimensional Eating Disorder Inventory

Response Bias

In order to examine the possible effects of "response set," the mean subtotal
score of all positively keyed items was compared to that of all negatively
keyed items. There were no significant differences within either the AN or FC
groups. For the AN group, negative and positive mean subtotals correlated .74
(p < .001) and for the FC group, the correlation was .67 (p < .001) suggesting
minimal response set bias.

Comparison Samples

Following initial scale validation, the test was administered to other com-
parison groups. A group of 195 women who would meet the diagnosis of
bulimia (BU) according to the DSM-III (APA, 1980) diagnostic criteria, where
episodes of bulimia and vomiting occurred daily to weekly, completed and
returned the questionnaire by mail as part of a large scale demographic
study.* These BU respondents resided throughout North America and were
randomly selected from the complete 1500 subject sample, the inclusion
criteria being: 1) the presence of bulimia, 2) self-induced vomiting one time a
week or more often, 3) female sex, 4) a current age less than 25 (in order to
match the BU sample to the AN and FC groups), and 5) no prior history of
weight loss sufficient to warrant a diagnosis of anorexia nervosa. According to
Russell (1979) and Palmer (1979), these BU women would share many of the
psychological themes with the anorexia nervosa group. A group of 44 female
obese (OB) subjects obtained from a local dieting group completed the ques-
tionnaire as well as 52 female formerly obese (FOB) subjects who had lost
weight from a mean of 130% to 100% of average weight. A small group of 17
recovered AN patients were also administered the questionnaire. The final
comparison group was 166 first and second year male college (MC) students
from the University of Toronto. Demographic characteristics for the com-
parison groups are presented in Table 1.

Criterion-Related Validity Provided by Comparison Samples

An attempt was made to establish criterion-related validity for some of the

subscales by demonstrating that the comparison groups described above
scored in the theoretically expected manner on specific subscales. One-way
analysis of variance and subsequent planned t-tests were employed. As ex-
pected, the bulimic AN patients scored higher than the restricter AN patients
only on the Bulimia and Body Dissatisfaction subscales (p< .01). Also, in ac-
cord with current theory, the BU comparison group had elevated Drive for
Thinness, Bulimia and Body Dissatisfaction scores, and were not significantly
different from the bulimic AN patients on these dimensions.

*Apprciation is extended to Dr. Craig Johnson of the Michael Reese Medical Center, Chicago for making
these data available.

24 INT'L JOURNAL OF EATING DISORDERS/VOL.2 NO.2 / WINTER 1983

 Multidimensional Eating Disorder Inventory

Both the OB and the FOB groups were administered an early version of the
EDI which did not include all of the items on the Interoceptive Awareness and
Maturity Fears subscales. The OB comparison group, as expected, had signifi-
cantly higher Body Dissatisfaction scores (x = 21.1 .88, p < .001) than either
the AN or FC groups. The OB group was also higher than the FOB group
(x=13.2 .95, p < .001) on Body Dissatisfaction and higher (p< 0.001) than
the FC group on the Bulimia (x = 4.6.66) and Drive for Thinness
(x=8.3.79) subscales. The OB and FOB groups did not differ from the FC
group on Ineffectiveness, Interpersonal Distrust or Perfectionism. The FC
group scored significantly higher than males (MC) only on the eating and
body attitude subscales {Drive for Thinness, Bulimia and Body Dissatisfac-
tion) (p <.003), while the MC's were higher than the FC's on Interpersonal
Distrust (p < .01). The mean subscale scores for the AN, recovered AN, FC
and MC groups are presented in Table 3. In keeping with the manner in which
subscale items were selected, the AN group had significantly higher scores
(p < .001) than the FC and MC groups on all subscales. It is notable that the
recovered AN group scored lower than the AN group (p< 0.001) on each sub-
scale and were not significantly higher (at p > 0.05) than the FC group on any
subscale.

Criterion-Related Validity from Other Sources

A question of further interest in determining the validity of the EDI is the

agreement between the self-report patient profiles and the clinical judgements
of experienced clinicians familiar with the patient's psychological presenta-
tion. One psychiatrist and one psychologist who were either primary thera-
pists or consultants familiar with the patient's background served as raters for
a subgroup (N = 49) of patients who had completed the EDI. For each patient,
the therapist-consultant was instructed to "rate the relevancy of each of these
traits or characteristics for this patient compared to other anorexics that you
have treated" on an analogue scale divided into ten centile intervals. The relia-
bility and validity of a similar type of rating method has been advocated by
Folstein and Luria (1973) and others. The therapist-consultants were given the
description of subscale content appearing in the methods section of the current
study and they were provided with each patient's total score (sum of scale
scores) percentile rank within the entire anorexic sample. Table 4 indicates the
agreement between therapist-consultant ratings and the AN patients' self-
report subscale scores. All correlations are significant at the p < O.OQl level.
For each subscale the percent of AN and FC groups correctly classified using
a discriminant function analysis is also reported in Table 4. The AN group was
divided into bulimic and restricter subtypes and a discriminant function
analysis correctly classified 85% of AN subjects based upon their Bulimia
subscale score.

INT'L lOURNAl OF EATING DISORDERS/ VOL.2 NO.2 / WINTER 1983 25

Multidimensional Eating Disorder Inventory

TABLE 3

Mean Scale ( standard error) Scores for AN, Recovered AN, FC, and MC groups

Anorexia Recovered Female Male

Nervosa* Anorexia Comparison Comparison
(N = 113) Nervosa (N = 577) N = 166
(N = 17)

Drive for Thinness 15,4(,50) 3,6(1,3) 5,0(,22) 1,6(,24)

Bulimia R = 2,7(,55) 0,3(,14) 2.0(,14) l,0(,14)

B = 10,8(,69)

Body Dissatisfaction R = 14,2(l,0) 6,3(1,5) 10,2(,32) 3,9(,39)

B = 17,4(l,0)

Ineffectiveness 14,4(,75) 3,1(,95) 2,0(,15) 1,6(,24)

Perfectionism 10,0(,47) 6,5(,88) 5,2(,16) 6,2(,30)

Interpersonal Distrust 7,7(,49) 1,9(,56) 2,2(,12) 3,1(,24)

Interoceptive Awareness+ 12,5(1,1) 2,1(,86) 2,9(,47) 1,4{,39)

Maturity Fears+ 6,0(,99) 1,1(,46) 2,5(,33) 2,7(,44)

* R = Restricters (N = 48); B = Bulimics (N = 65); when bulimic and restricter subgroups do not
differ significantly on subscale scores, only total group means are reported.
+ N = 35 for AN, N = 13 for Recovered AN, N=82 for FC, N = 68 for MC

Convergent and Discriminant Validity

Since several subscales of the EDI overlap conceptually with available psy-
chological tests, convergent and discriminant validity could be determined for
subsamples of AN patients. Table 5 presents the correlations for the AN group
between EDI subscales and tests measuring: "anorexic" attitudes (EAT, Garner
and Garfinkel, 1979), "restraint" (Herman and Polivy, 1975), overall body dis-
satisfaction (modified from Berscheid et al., 1973), dissatisfaction with bodily
regions associated with changes at maturation (i.e. breasts, buttocks, hips and
abdomen), locus of control (Reid and Ware, 1973), self-control (Reid and
Ware, 1973), feelings of inadequacy (Janis and Field, 1959), depression (BDI,
Beck, 1978), physical anhedonia (Chapman, Chapman and Raulin, 1976),
somatization, obsessionality, anxiety, depression and interpersonal sensitivity
(Hopkins Symptom Check List, HSCL, Derogotis, Lipman, Rickels, Uhlen-
huth and Covi, 1974).
Because of the likelihood of such a large number of intercorrelations yield-
ing significant findings by chance (Type I Error), an alpha level of p < 0.001
was chosen for each correlation. With 120 comparisons, a Bonferroni-type ad-

INT'L lOURNAL OF EATING DISORDERS/ VOL.2 NO.2 / WINTER 1983

 Multidimensional Eating Disorder Inventory

TABLE 4

Criterion Validity: Correlations of Anorexia Nervosa (AN)

Patients' (N=49) Subscale Scores with Clinicians' Ratings and
Percents of Anorexia Nervosa and Female Comparison (FC) Croups
Correctly Classified in Discriminant Function Analysis

Subscale Clinicians' Rating* Percent of AN

and FC Groups
Correctly Classified
Drive for Thinness .53 89.9
Bulimia .57 88.3
Body Dissatisfaction .44 87.6
Ineffectiveness .68 92.0
Perfectionism .47 89.2
Interpersonal Distrust .56 90.0
Interoceptive Awareness .51 93.1
Maturity Fears .43 88.3
All Subscales 91.7
* All correlations significant (p < 0.001). With a Bonferroni-type adjustment and 8 individual
comparisons, the family-wise error rate is 0.01 (Myers, 1979).

justment yields a family-wise error rate of .11 (Myers, 1979. Percent of

average weight, established according to the Health and Welfare Canada
(1954) norms, was only significantly related to the Bulimia (N = 113, r=.33,
p < O.Od) and the Body Dissatisfaction (N=113, r=.36, p < 0.001) subscales.
The Drive for Thinness scale had the highest correlation with the EAT and
Restraint scale; Bulimia was most related to lack of self-control, restraint and
body dissatisfaction; Body Dissatisfaction was most related to dissatisfaction
with the maturational regions (breasts, buttocks, hips and abdomen); Ineffec-
tiveness was related to low self-esteem, depression and external locus of con-
trol; Perfectionism was most related to interpersonal sensitivity on the HSCL
and Interpersonal Distrust was related to low self-esteem and depression. Data
on the revised Interoceptive Awareness and Maturity Fears subscales were on-
ly available for a small group of Ss and their correlations with other measures
were generally low and all non-significant (Table 5).
The EAT was also administered to 55 female college students who had com-
pleted the EDI. The total EAT score correlated highest with the Drive for Thin-
ness subscale of the EDI (r=.88, p< 0.001).

INT'L JOURNAL OF EATINC DISORDERS/ VOL.2 NO.2/ WINTER 1983 27

 Multidimensional Eating Disorder Inventory

Intercorrelations among Subscales

Table 6 presents the intercorrelations between EDI subscales for the AN and
FC samples. Although these correlations are presented mainly for descriptive
purposes, it is worth noting that there are 56 individual correlations. With a
comparison-wise alpha of 0.001 and a Bonferroni-type adjustment, the family-
wise error rate is 0.05 (Myers, 1979). For the AN group, only 9 of the 28 corre-
lations were significant, with no coefficients reaching .60. Because of the large
FC sample size, most subscales were significantly correlated even though the
coefficients were generally low (only three were above .60).

Construct Validity

Nunnally (1967) has indicated that construct validity can only be deter-
mined by a series of experiments which demonstrate that the theoretical con-
struct and operational measures are indeed related to one another. The con-
gruence between clinicians' ratings and patients' subscale scores provides some
evidence of construct validity. Moreover, the demonstration of convergent
and discriminant validity for subscales as well as their ability to differentiate
between the AN and FC groups contribute to the construct validity of the sub-
scales of the EDI.

TABLE 5

Convergent and Discriminant Validity: Correlations Between Subscales

and Other Psychometric Instruments for AN Patients

MEASURES (N) IEDI SUBSCALES

DT B BD I P ID IA-I- M +
EAT (50) .51* .05 .32 .28 .19 .08 .22 - . 0 2
Restraint (52) .50* .44* .42* .21 .22 .03 .00 - . 2 1
Body Dissatisfaction (35) .29 .41* .55* .25 .23 .23 .16 - . 0 3
Dissatisfaction with (35) .45* .44* .69* .38 .07 .30 .15 - . 3 1
Maturational Regions
Self-Esteem (33) - . 1 6 -.17 -.14 .76* .04 .57* .21 .15
Locus of Control (66) .23 .28 .29 .44* .28 .26 .35 - . 0 5
Lack of Self-Control .18 .53* .14 .11 .02 - . 1 5 .15 - . 2 9
BDI (64) .26 .25 .27 .50* .12 .29 .50 - . 0 5
Physical Anhedonia (61) .10 --.15 .07 .05 .01 .29 .32 .16
HSCL (Total) (65) .33 .22 .31 .57* .39* .19 .31 .07
Somatization .32 .18 .19 .29 .19 .21 .40 - . 3 7
Obsessionality .30 .29 .08 .42 .37 - . 0 4 .15 .48
Anxiety .27 .12 .28 .45* .33 .26 .28 - . 0 3
Depression .26 .11 .24 .62* .35 .27 .34 .09
Interpersonal Sensitivity .27 .23 .20 .43 .51* .06 .27 - . 1 4
* p < 0.001 for each comparison, for family of 120 comparisons, p < 0.11
+ data were available on only 23 anorexia nervosa subjects

INT'LlOURNALOF EATrNC DrSORDERS/ VOL.2 NO.2/ WINTER 1983

 Multidimensional Eating Disorder Inventory

TABLE 6

Intercorrelations between subscales

ANOREXIA NERVOSA (N=113) +

1) Drive for Thinness .17 .55* .28* .15 .26 .32* - . 2 5

2) Bulimia .55* .30* .06 .09 - . 1 1 .04 - . 0 3
3) Body Dissatisfaction .58* .45* .20 .08 .14 - . 0 1 - . 2 8
4) Ineffectiveness .29* .37* .28* .28* .59* .45* .50
5) Perfectionism .16* .13* .02 .17* .17 .30 .10
6) Interpersonal Distrust .16* .18* .16* .41* .11 .47* .09
7) Interoceptive Awareness .69* .78* .50* .82* .29 .36* .16
8) Maturity Fears .05 .03 - . 1 4 .08 .19 -.05 .03

NORMAL COMPARISON GROUP (N = 577)-l- + (below diagonal)

* p < 0.001 for each comparison, for family of 56 comparisons p < 0.05
+ N = 3 5 for subscales 7 and 8
+ + N=82 for subscales 7 and 8

DISCUSSION

The current study proposes the Eating Disorder Inventory (EDI) as a new
measure of attitudinal and behavioral dimensions relevant to anorexia ner-
vosa and bulimia. Results indicate that the EDI is a reliable (internally consis-
tent) and valid test within the limits imposed by the samples selected for this
initial study. Eight subscales were deductively derived and then empirically
validated with the aim that they differentiate, with very little overlap, between
patients with anorexia nervosa and females from a college comparison group.
Subscales indicate two levels of disturbance considered relevant to patients
with anorexia nervosa. The first three subscales (Drive for Thinness, Bulimia
and Body Dissatisfaction) assess attitudes and/or behaviors related to eating
and body shape. While disturbances in these areas are central to anorexia ner-
vosa, they may also exist in other groups of dieters. The remaining five sub-
scales of the EDI measure traits which have been identified by clinical theorists
as fundamental aspects of the psychopathology of anorexia nervosa. These
subscales have been labelled Ineffectiveness, Interpersonal Distrust, Interocep-
tive Awareness, Perfectionism, and Maturity Fears, and their content has been
described in detail in the methods section of this report. Despite some consen-
sus regarding the significance of these areas of disturbance in the pathogenesis
of anorexia nervosa, there have been few attempts to operationalize or objec-
tively assess them.

INT'L lOURNALOF EATINC DISORDERS/VOL.2 NO.2 / WINTER 1983 29

 Multidimensional Eating Disorder Inventory

Appendix

EDI

Name: .Date:
Age: _
Present Weight: .Height:. .Sex:.
Highest Past Weight: _ (lbs)
(excluding pregnancy)
How Long Ago? _ . (months)
How Long Did You Weigh This? . (months)
Lowest Past Adult Weight: (lbs)
How Long Ago? . (months)
How Long Did You Weigh This? . (months)
What Do You Consider Your Ideal Weight To Be? (lbs)
Age at Which Weight Problem Began (if any)
Father's Occupation:
Instructions:
This is a scale which measures a variety of attitudes, feelings and behaviours.
Some of the items relate to food and eating. Others ask you about your feel-
ings about yourself. THERE ARE NO RIGHT OR WRONG ANSWERS SO
TRY VERY HARD TO BE COMPLETELY HONEST IN YOUR ANSWERS.
RESULTS ARE COMPLETELY CONFIDENTIAL. Read each question and
place an (X) under the column which applies best for you. Please answer each
question very carefully. Thank you.

i 1. I eat sweets and carbohydrates without

feeling nervous.
2. I think that my stomach is too big.
3. I wish that I could return to the securi-
ty of childhood.
4. I eat when I am upset.

INT'LJOURNALOF EATING DISORDERS/VOL.2 NO.2/WINTER 1983

 Multidimensional Eating Disorder Inventory

Evidence for the validity of the subscales comes from several sources other
than their ability to differentiate between anorexia nervosa and comparison
groups. First, there was good agreement between patients' self-report profiles
and the clinical judgements of experienced clinicians. This type of criterion
validity is often difficult to demonstrate because it is influenced by both pa-
tient and clinician sources of error. Secondly, a small group of clinically
recovered anorexia nervosa patients scored similarly to college females on all
subscales. Thirdly, varying degrees of convergent and discriminant validity
were established for the anorexia nervosa sample depending on the number of
other tests administered and their relevance to the hypothesized constructs
underlying each subscale. For example, the subscales tapping attitudes toward
eating and shape were highly related to other available measures of these
dimensions and less related to personality traits. On the other hand, subscales
aimed at psychological functioning usually had minimal correlations with
other measures of eating. Moreover, the subscales assessing "personality"
traits had logical patterns of correlations with other trait measures. The Inef-
fectiveness subscale was most highly correlated with feelings of inadequacy,
depression and external locus of control; Perfectionism with a measure of in-
terpersonal sensitivity; Interpersonal Distrust with low self-esteem and depres-
sion. Percent of average weight was only significantly correlated with the
Bulimia and Body Dissatisfaction subscales.
The EDI differentiated between individuals with anorexia nervosa and
obesity as well as formerly obese subjects recruited from a weight-loss pro-
gram. It may be concluded that the EDI does not simply measure concern with
weight or dieting but rather indicates more disturbed attitudes. It could be
postulated that the EDI would be sensitive to gross psychopathology reflected
in the minority of obese individuals for whom being overweight is an overt
manifestation of more pervasive emotional disturbance. This impression has
been confirmed for a small number of obese patients whom we have tested and
treated but not described in this report.
The EDI was administered to a large group of normal weight women with
bulimia but who had not experienced sufficient weight loss to meet a diagnosis
of anorexia nervosa. Similar to the anorexic sample, the bulimic group scored
significantly higher than normal on EDI subscales related to Bulimia, Drive for
Thinness and Body Dissatisfaction. In an independent comparison of bulimic
anorexia nervosa patients and bulimic normal weight women with no history
of anorexia nervosa, we have found that both groups score similarly on all EDI
subscales except Maturity Fears on which the anorexia nervosa patients
display greater psychopathology (Gamer, Garfinkel and O'Shaughnessy, sub-
mitted).
Garner, Polivy, Olmsted and Garfinkel (submitted) have used the EDI to
compare the psychological characteristics of weight-preoccupied college
women with those of patients with primary anorexia nervosa. While the
groups were similar in Drive for Thinness and. Body Dissatisfaction and
Perfectionism, the most salient difference was that the anorexia nervosa sub-

INT'L lOURNALOF EATING DISORDERS/VOL.2 NO.2/ WINTER 1983 31

Multidimensional Eating Disorder Inventory

jects were significantly more pathological on the Ineffectiveness and In-

teroceptive Awareness dimensions of the EDI. This is consistent with Bruch's
(1962, 1973) postulation that these traits are part of the core disturbance in
anorexia nervosa.
While the EDI may have utility as a prognostic screening instrument,
several cautions must be emphasized. First, like all self-report instruments, the
EDI is vulnerable to distortion due to response style bias and inaccurate re-
porting by the subject. Moreover, the defensive structure of the patient may
invalidate or distort self-evaluation data. However, self-report information
has the inherent advantages of economy, actuarial scoring and access to
phenomenological infonnation that may be unavailable through conventional
clinical assessments or more oblique tests. Secondly, the EDI was empirically
refined based on its capability in differentiating between a criterion group and
non-clinical samples. Thus it may lack external validity, in that elevated scale
scores among non-clinical subjects cannot be assumed to reflect the same psy-
chopathology inferred for the patient group. Thirdly, the EDI should not be
considered to represent an exhaustive sampling of psychopathological charac-
teristics of anorexia nervosa. Other personality features such as obsessional-
ity, rigidity, compliance, depression, introversion, interpersonal sensitivity,
poor ego strength, excessive control and others have been identified (Crisp et
al., 1979; Strober, 1980; Smart, Beumont and George, 1976; Stonehill and
Crisp, 1977; Wingate and Christie, 1978). Fourthly, while some of the psycho-
logical dimensions assessed by the EDI may have primary significance, the
presence or magnitude of others may simply be by-products of the disorder. It
is not justified to infer a causal relationship between the attributes observed
and the development of anorexia nervosa. Finally, the EDI should not be
employed as the sole means of screening for or diagnosing anorexia nervosa.
Several recent reports have apparently misinterpreted these to be intended
purposes of the Eating Attitudes Test (Button and Whitehouse, 1981;
Williams, Hand and Tarnopolsky, 1982). Although the EDI or EAT may be
useful as preliminary screening tools for identifying groups in which a high
proportion of subjects have formal eating disorders, these must be confirmed
by clinical diagnosis; psychological tests should be an adjunct, not a replace-
ment, for clinical judgments.
Despite these qualifications, the EDI represents an attempt to isolate and ob-
jectively measure specific psychological features viewed as important in the
anorexia nervosa syndrome and which have been previously described. Both
reliablity and validity data are presented which support its use as an
economical instrument for evaluating behavioral and psychological traits
common to individuals with eating disorders.

The authors are grateful for the assistance of Dr. Paul Garfinkel, Ms.
Martha O'Shaughnessy, Ms. Vanessa Boratto, Ms. Victoria Mitchell, Ms.
Shirley Kartuz and Ms. Wendi Rockert. This research was supported by the

32 INT'LIOURNALOF EATING DISORDERS/VOL.2 NO.2/WINTER 1983

 Multidimensional Eating Disorder Inventory

Ontario Mental Health Foundation (grant 810) and a Medical Research

Council of Canada Scholarship (Dr. Gamer). An abbreviated version of this
paper was presented at the American Psychological Assocation Annual
Meeting, Los Angeles, December, 1981.

REFERENCES
American Psychiatric Association (1980) Diagnostic and Statistical Manual of Mental Disorders (DSM 111).
Washington, D . C , American Psychiatric Association.
Berecheid, E., Walster, E. and Hohmstedt, G. (1973) The happy American body: A survey report. Psychology
Today Nov: 119-131.
Beck, A.T. (1978) Depression Inventory, Center for Cognitive Therapy, Philadelphia.
Beumont, P.J.V., Abraham, S.F., Argall, W.J. and Simpson, K.G. (1981) The sexual behavior and attitudes of
girls and women with anorexia nervosa. Psychological Medicine 2:131-140.
Beumont, P.J.V., George, G.C. and Smart, D.E. (1976) "Dieters" and "vomiters and purgers" in anorexia ner-
vosa. Psychological Medicine 6:617-622.
Bruch, H. (1962) Perceptual and conceptual disturbances in anorexia nervosa. Psychosomatic Medicine
24:187-194.
Bruch, H. (1973) Eating Disorders. New York, Basic Books.
Bruch, H. (1978) The Golden Cage. New York, Basic Books.
Button, E.J. and Whitehouse, A. (1981) Subclinical anorexia nervosa. Psychological Medicine 11:509-516.
Casper, R.C, Eckert, E.D., Halmi, K.A., Goldberg, S.C and Davis, J.M. (1980) Bulimia. Archives of General
Psychiatry 37:1030-1035.
Chapman, L.J., Chapman, J.P. and Raulin, M.L. (1976) Scales for physical and social anhedonia. Journal of
Abnormal Psychology 85:374-382.
Crisp, A.H. (1965) Some aspects of the evolution, presentation and follow-up of anorexia nervosa. Proceedings
of the Royal Society of Medicine 58:814-820.
Crisp, A.H. (1977) The differential diagnosis of anorexia nervosa. Proceedings of the Royal Society of Medicine
70:686-690.
Crisp, A.H. (1979) Personality, body weight and ultimate outcome in anorexia nervosa. Journal of Clinical Psy-
chiatry 40:332-335.
Crisp, A.H. (1980) Anorexia Nervosa: Let Me Be. New York, Grune and Stratton.
Crisp, A.H. (1981) Anorexia Nervosa at a normal weight! The abnormal normal weight control syndrome. The
International Journal of Psychiatry in Medicine 11:203-234.
Dally, P.J. (1969) Anorexia Nervosa. New York, Grune and Stratton.
Dally, P. and Gomez, J. (1979) Anorexia Nervosa. London, Heinemann.
Derogatis, L., Lipman, R., Rickets, K., Uhlenhuth, E.H. and Covi, L. (1974) The Hopkins symptom checklist
(HSCL): A self-report symptom inventory. Behavioral Science 19:1-15.
Feighner, J.P., Robins, E., Guze, S.B., Woodruff, R.A., Winokur, G. and Munoz, R. (1972) Diagnostic criteria
for use in psychiatric research. Archives of General Psychiatry 26:57-63.
Fichter, M.M. and Kesser, W. (1980) Das Anorexia Nervosa Inventar zur Selbstbeurteilung (ANIS). Arch
Psychiatr Nervenkr 228:67-89.
Folstein, M.F. and Luria, R. (1973) Reliability, validity, and clinical application of the visual analogue mood
scale. Psychological Medicine 3:479-486.
Garfinkel, P.E. and Gamer, D.M. (1982) Anorexia Nervosa: A multidimensional perspective. New York,
Brunner/Mazel.
Garfinkel, P.E., Moldofsky, H. and Gamer, D.M. (1980) The heterogeneity of anorexia nervosa. Archives
of General Psychiatry 37:1036-1040.
Gamer, D.M. and Garfinkel, P.E. (1979) The eating attitudes test: An index of the symptoms of anorexia ner-
vosa. Psychological Medicine 9:1-7
Gamer, D.M. and Garfinkel, P.E. (1980) Socio-cultural factors in the development of anorexia nervosa. Psy-
chological Medicine 10:647-656.
Gamer, D.M. and Garfinkel, P.E. (1981) Body image in anorexia nervosa: Measurement, theory and clinical
implications. International Journal of Psychiatry in Medicine 11:263-284.
Gamer, D.M., Garfinkel, P.E., and O'Shaughnessy, M. (Submitted for publication) A comparison between
bulimia in anorexia nervosa and bulimia in normal weight women.
Gamer, D.M., Garfinkel, P.H. and Bemis, K.M. (1982) A multidimensional psychotherapy for anorexia ner-
vosa. International Journal of Eating Disorders 1:3-46.
Gamer, D.M., Garfinkel, P.E., Stancer, H . C and Moldofsky, H. (1976) Body image disturbances in anorexia
nervosa and obesity. Psychosomatic Medicine 38:227-237.
Gamer, D.M., Olmsted, M.P., Bohr, Y. and Garfinkel, P.E. (in press) The eating attitudes test: Psychometric
features and clinical correlates. Psychological Medicine.

INT'LIOURNALOF EATING DISORDERS/VOL2 NO.2/WINTER 1983 33

 Multidimensional Eating Disorder Inventory

Gamer, D,M,, Polivy, J,, Olmsted, M,P., and Garfinkel, P.E, (Submitted for publication) A comparison between
anorexia nervosa and weight preoccupied women,
Goldberg, S,C,, Halmi, K.A,, Eckert, E.D., Casper, R.C., Davis, J.M. and Roper, M. (1980) Attitudinal
dimensions in anorexia nervosa. Journal of Psychiatric Research 15:239-251,
Goodsitt, A, (1969) Anorexia nervosa. British Journal of Medical Psychology 42:109-118,
Goodsitt, A, (1977) Narcissistic disturbances in anorexia nervosa. In Adolescent Psychiatry, Vol. 5. Edited by
Feinstein, S.C, and Giovacchini, P.L, New York, Jason Aronson.
Halmi, K.A., Falk, J.R. and Schwartz, E. (1981) Binge-eating and vomiting: A survey of a college population.
Psychological Medicine 11:697-706,
Hawkins, R,C. and Clement, P.F. (1980) Development and construct validation of a self-report measure of
binge eating tendencies. Addictive Behaviors 5:219-226,
Health and Welfare (Canada): Canadian average weights for height, age and sex (pamphlet) (1954). Nutrition
Division of the Department of Health and Welfare,
Herman, C P . and Polivy, J, (1975) Anxiety, restraint and eating behavior. Journal of Personality 84:666-672.
Hood, J., Moor, T.E, and Gamer, D.M. (1982) Locus of control as a measure of ineffectiveness in anorexia
nervosa. Journal of Consulting and Clinical Psychology 50:3-13,
Janis, LL, and Field, P.B, (1959) Sex differences and personality factors related to persuasibility. In Personality
and Persuasibility. Edited by Houland, CJ. and Janis, LL, New Haven, Yale University Press,
Johnson, CL., Stuckey, M,K., Lewis, L.D, and Schwartz, D,M, (in press) Bulimia: A descriptive survey of 509
cases. In Anorexia Nervosa: Recent Developments. Edited by Darby, P.L., Garfinkel, P.E., Gamer,
D.M. and Coscina, D,V. New York, Allan R. Liss.
Kalucy, R.S., Crisp, A.H, and Harding, B, (1977) A study of 56 families with anorexia nervosa. British Journal of
Medical Psychology 50:381-395.
Myers, J.L, (1979) Fundamentals of Experimental Design. Boston, Allyn and Bacon.
Nunnally, J.C (1969) Psychometric Theory. New York, McGraw-Hill Book Company.
Palmer, R.L. (1979) Dietary chaos syndrome: A useful new term? British Journal of Medical Psychology
52:187-190.
Pyle, R.L,, Mitchell, J.E, and Eckert, E.D. (1981) Bulimia: A report of 34 cases. Journal of Clinical Psychiatry
42:60,64.
Reid, D.W, and Ware, E.E. (1973) Multidimensionality of intemal-extemal control: Implications for past and
future research. Canadian Journal of Behavioral Science 5:264-271,
Russell, G.F.M, (1979) Bulimia nervosa: An ominous variant of anorexia nervosa. Psychological Medicine
9:429-448.
Selvini-Palazzoli, M. (1978) Self-Starvation From individual to family therapy in the treatment of anorexia
nervosa. New York, Jason Aronson,
Slade, P,D, (1973) A short anorexic behavior scale, British Journal of Psychiatry 122:83-85.
Smart, D,E., Beumont, P.J.V. and George, G.C.W, (1976) Some personality characteristics of patients with ano-
rexia nervosa. British Journal of Psychiatry 128:57-60.
Stonehill, E, and Crisp, A.H, (1977) Psychoneurotic characteristics of patients with anorexia nervosa before and
after treatment and at follow-up 4-7 years later. Journal of Psychosomatic Research 21:187-193,
Story, I, (1977) Caricature and impersonating the other: Observations from the psychotherapy of anorexia
nervosa. Psychiatry 39:176-188,
Strober, M. (1980) Personality and symptomatological features in young, nonchronic anorexia nervosa patients.
Journal of Psychosomatic Research 24:353-359.
Strober, M, (1981) The significance of bulimia in anorexia nervosa: An exploration of possible etiological factors.
The Intemational Journal of Eating Disorders 1:28-43.
Strober, M. (in press) An empirically derived typology of anorexia nervosa. In Anorexia Nervosa: Recent De-
velopments. Edited by Darby, P.L,, Garfinkel, P.E., Gamer, D.M, and Coscina, D,V. New York Allan
R. Liss.
Thompson, M,G, and Schwartz, D.M, (1982) Life adjustments of women with anorexia nervosa and anorexic-
like behavior. Intemational Joumal of Eating Disorders 1:47-60.
Wardle, J, (1980) Dietary restraint and binge eating. Behavioral Analysis and Modification 4:201-209.
Williams, P., Hand, D, and Tamopolsky, A, (1982). The problem of screening for uncommon disorders
a comment on the Eating Attitudes Test, Psychological Medicine 12:431-434.
Wingate, B.A. and Christie, M,J, (1978) Ego strength and body image in anorexia nervosa. Joumal of Psycho-
somatic Research 22:201-204.

INT'UOURNALOF EATING DISORDERS/VOL.2 NO.2/WINTER 1983

View publication stats