Differences among bulimic subgroups, binge eaters, and normal eaters in a female college population

MISSING IMAGE

Material Information

Title:
Differences among bulimic subgroups, binge eaters, and normal eaters in a female college population
Physical Description:
x, 148 leaves. : ;
Language:
English
Creator:
Ousley, Louise Byrne, 1957-
Publication Date:

Subjects

Subjects / Keywords:
Food habits   ( lcsh )
Bulimia   ( lcsh )
Appetite disorders   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1986.
Bibliography:
Includes bibliographical references (leaves 141-147).
Statement of Responsibility:
by Louise Byrne Ousley.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000879776
notis - AEH7561
oclc - 14937681
System ID:
AA00003391:00001

Full Text












DIFFERENCES AM1NG
BUITMIC SUBGROUPS, BINGE EATERS, AND NORMAL EATERS
IN A FEMALE COLLEGE POPULATION














BY

IDUISE BYRNE OUSLEY


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR
THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA


1986



















This dissertation is dedicated to all my past, present and future
clients, in the hopes that this and other research will provide
the basis for improved treatment and prevention of their eating
disorders.














ACKNCWMIEDGEMENTS


Space and decorum have somewhat inhibited the length and

content of this section; nevertheless, I will attempt to convey my

thanks as briefly and genuinely as I am able. First, to my

chairman and advisor, Franz Epting, and to my committee members,

Dorothy Nevill, Dave Suchman, Connie Sheehan, and Bob Ziller: I

appreciate your time, support and your faith in me. To Scott

Whitely without whom this project would not have been possible, I

send my blessings. Those who helped with the design and analysis

of the project, especially Mike Furlong and Carol Geer, deserve a

round of applause (at least). Darlene Rhoades has been wonderful

as my typist and crisis intervenor. My friends and family have

loved and supported me throughout, knowing just when to ask "How's

it going?" and when to not ask. To those who are dearest, espe-

cially Jill and Gary, and who have helped me live and laugh

throughout this project: I love you; don't ever change.















TABLE OF CONTENTS


UST OF TABLES


ABSTRA . ......... . .

CHAPTERS

I INIRODUCTION . . . ..... .


Statement of the Problem .
Importance of the Problem .


iii

vi

ix



1


3
4


II REVIEW OF THE LITERATURE .


Defining and Classifying Bulimia .. .*. 6
Incidence and Parameters of Bulimic Symptams
(Binge Eating) in the Obese . . 11
Incidence and Parameters of Bulimic Symptams
in Anorexia Nervosa ... ..... 13
Incidence and Parameters of Binge Eating in
the Nonanorexic, Nanobese (Normal Weight)
Population . ......... . 15
The Bulimic Syndrame in the Nonanorexic,
Nonobese (Normal Weight) Population . 21
Description of Weight and Symptom History . 25
Psychological Profile of Bulimics . . 29
Summary . . . . . 40
Hypotheses . . .. . .. 47


III MEHOD . . . . . .

Subjects . . . . .
Materials . . . . .
Procedure. . . ..
Data Analysis .................

IV RESULTS . . . . . .

Eating Behavior and Weight-Related
Characteristics . . . .
Major Findings: The EDI and the CBSI . .
SODI Results . . . . .


51

51
51
56
59

61


61
62
100


ACKNOWIEDGEMENTS . . .


. . .










Table of Contents (con't)


V DISCUSSION . . . . .

Summary of Results . . . ..
Relation to Previous Research . . .
Limitations and Implications of the Study .


APPENDIX

A

B

C

D


ES

WEIGHT MANAGEMENT QUESTIONNAIRE . .

SELF-OTHER DIFFEENIATICN INSTRUMENT . .

EAMY HISTORY QUESTIONNAIRE . . .

CONVERSION OF RAW SCORES TO PERCENTILE SCORES
FOR THE OMNTE BORDERLINE SYNDROME INDEX .


REFERENCES . . . . . .

BIOGRAPHICAL SK=EC . . . . .


109

109
112
118


126

131

136


140

141

148


.*











LIST OF TABLES


Table

2-1

4-1

4-2


4-3


4-4

4-5


4-6


4-7


4-8


4-9


4-10


4-11

4-12

4-13

4-14

4-15


4-16


4-17


EDI SCALES

WEIGHT-RELATED CHARACTERISTICS (OF 813 RESPONDENTS)

FREQUENCY OF TOTAL SAMPLE MEETING THE CRITERIA
FOR BULIMIC SYNDROME

ANALYSIS OF VARIANCE SUMMARY TABLE FOR DRIVE FOR
THINNESS

ANALYSIS OF VARIANCE SUMMARY TABLE FOR BUIJMIA

ANALYSIS OF VARIANCE SUMMARY TABLE FOR BODY
DISSATISFACTION

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
INEFFECTIVENESS

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
PERFECTIONISM

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
INTERPERSONAL DISTRUST

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
INTEROCEPTIVE AWARENESS

ANALYSIS OF VARIANCE SUMMARY TABLE FOR MATURITY
FEARS

MEANS, STANDARD DEVIATIONS FOR DRIVE FOR THINNESS

MEANS, STANDARD DEVIATIONS FOR BULIMIA

MEANS, STANDARD DEVIATIONS FOR BODY DISSATISFACTION

MEANS, STANDARD DEVIATIONS FOR INEFFECTIVENESS

MEANS, STANDARD DEVIATIONS FOR INTEROCEPTIVE
AWARENESS

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
THE CBSI

MEANS, STANDARD DEVIATIONS FOR THE CBSI

vi


Page

46

63


64


67

68


69


70


71


72


73


74

75

76

77

78


79


81

82









Tables (con't)


List of

Table

4-18


4-19

4-20


4-21


4-22


4-23


4-24


4-25


4-26

4-27

4-28

4-29

4-30


4-31

4-32

4-33


4-34


4-35


ANALYSIS OF VARIANCE SUMMARY TABLE FOR
THINNESS

ANALYSIS OF VARIANCE SUMMARY TABLE FOR

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
DISSATISFACTION

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
INEFFECTIVENESS

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
PERFECTOTCISM

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
INTERPERSONAL DISTRUST

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
INTEROCEPTIVE AWARENESS

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
FEARS

MEANS, STANDARD DEVIATIONS FOR DRIVE FC

MEANS, STANDARD DEVIATIONS FOR BULIMIA

MEANS, STANDARD DEVIATIONS FOR BODY DIE

MEANS, STANDARD DEVIATIONS FOR INEFFEC)

MEANS, STANDARD DEVIATIONS FOR INTEROCE
AWARENESS

ANALYSIS OF VARIANCE SUMMARY TABLE FOR

MEANS, STANDARD DEVIATIONS FOR THE CBS]

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
SELF-PARENT VARIABLE

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
SELF-SISTER VARIABLE

ANALYSIS OF VARIANCE SUMMARY TABLE FOR
SELF-BRDTHER VARIABLE


R THINNESS



;SATISFACTION

TIVENESS

SPITVE


THE CBSI

E


vii


DRIVE FOR


BULIMIA

BODY















MATURITY


97

101

102


105


106


107









List of Tables (con't)

Table

4-36 MEANS, STANDARD DEVIATIONS FOR SELF-PARENT,
SELF-SISTER, SELF-BRMOIER VARIABLES

5-1 EDI SCALE MEANS FROM FOUR STUDIES

5-2 CONTINUUM OF SEVERITY OF BINGING AND BULIMIC
SYMPiaS


viii


108

115














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy


DIFFERENCES AMDNG
BULIMIC SUBGROUPS, BINGE EATERS, AND NORMAL EATERS
IN A FEMALE COLLEGE POFUIATION


BY
IDUISE BYRNE OUSLEY

May 1986

Chairman: Franz Epting
Major Department: Psychology



The purpose of this study was to test the hypotheses that a)

women who fit the DSM-III criteria for the syndrome of bulimia are

more disturbed than those who have no eating problems; b) wamen

with the syndrome of bulimia have more intrapersonal and interper-

sonal difficulties than do women who have only the symptom of

bulimia (binge eating); c) wamen who have the symptom of bulimia

are more preoccupied with thinness, have more problems with

controlling their eating and have a more negative body image than

those who have no eating problem; and d) those with the syndrome

of bulimia who purge using laxatives, vomiting or spitting out

food (purging bulimics) have more intrapersonal and interpersonal

difficulties than those with the syndrome who do not purge (non-








purging bulimics). These hypotheses were tested using the Eating

Disorder Inventory (EDI) and the Conte Borderline Syndrcme Index

(CBSI) with a randomly selected group of college women from a

nonclinical population. Another instrument, the Self-Other

Differentation Instrument was developed by the author and included

as an exploratory measure in this study.


Partial support was found for the first three hypotheses.

The bulimic syndrome group (purging and nonpurging bulimics

combined) had greater preoccupation with thinness, greater body

dissatisfaction, more compulsive eating, more confusion of emo-

tional and hunger/satiety cues, and more borderline personality

features than either the symptom or control groups. The bulimic

syndrome group also had greater feelings of ineffectiveness than

the control group. The symptom (binge eating) group had greater

preoccupation with dieting and greater body dissatisfaction than

the control group. While the hypothesized differences between the

purging and nonpurging bulimic groups were not confirmed, other

unhypothesized differences were found which suggested that purging

bulimics are more disturbed than nonpurging bulimics. The impli-

cations of the results for research and treatment are discussed

and a continuum is presented to describe the symptoms associated

with the worsening of disordered eating patterns.














CHAPTER I
INTRODUCTION


The term bulimia (from the Greek, "ox appetite") is used to

refer to both a symptom and a psychiatric syndrome of disordered

eating found primarily in young white females. The symptom of

bulimia (also called binge eating, compulsive eating, or gorging) is

characterized by the rapid consumption of large quantities of food

in a relatively short period of time. The bulimic syndrome is a

pattern of episodes of binge eating followed by methods used to

compensate not only for the binge calories but also for the guilt,

depression, and feelings of loss of self-control. The measures most

frequently used to compensate for the binge include purging (self-

induced vomiting, laxative or enema abuse), restricting (fasting or

restrictive dieting), and excessive, compulsive exercise. Bulimia

was first given status as a psychiatric syndrome with the 1980

publication of the DSM-III (American Psychiatric Association, 1980).

Since then therehas been a proliferation of scientific and popular

' literature written on the subject. Demands for treatment of bulimia

have risen sharply in the last decade. Although the mental, emo-

tional, and physical complications of the bulimic syndrome are less

life-threatening than those of its close relative, anorexia nervosa,

the incidence of bulimia in young women is estimated to be nearly

ten times that of anorexia (American Psychiatric Association, 1980).









While bulimia has had official status as a psychiatric syndrome

only five years, it has appeared throughout the past century in the

medical and psychiatric literature as a symptom in both anorexic and

obese patients (Casper, 1983). Although the number of reports of

binge eating in normal weight women increased during the 1950's and

60's, bulimic symptoms were seen primarily as an extension of the

anorexia nervosa syndrome. In 1959 Stunkard identified "binge

eating" as a disturbed eating pattern characteristic of obese

patients. Bruch (1974) made one of the first attempts to identify

binge eating and vomiting as a pattern in her normal weight pa-

tients. She named it the "thin fat person syndrome" and described

it as a cycle of binging and self-induced vomiting or starving

accompanied by preoccupations with weight and food. Boskind-Lodahl

(1976) subsequently coined the term "bulimarexia" to refer to an

eating disorder of gorging and purging and/or fasting separate from

anorexia and commonly found in normal or near normal weight women.

She cited the sociocultural pressures to achieve perfection through

thinness as responsible for the increasing development of bulimarex-

ia among young women.

A number of studies have reported a rise in the past decade in
the incidence of bulimia in both clinical (patients/clients) and

nonclinical populations of normal weight or overweight (nonobese)

women (Boskind-Indahl and White, 1978; Casper, 1983; Fairburn and

Cooper, 1983; Halmi, Falk and Schwartz, 1981; Johnson, Lewis and

Love, 1984(b); Mitchell and Pyle, 1982; Pyle, Mitchell, Eckert,

Halverson, Neuman, and Goff, 1983; Sinoway, 1983; Stangler and

Printz, 1980). While reviews of the literature (Casper, 1983;









Halmi, 1983; Johnson, Lewis, and Hagman, 1984 (a); Mitchell and Pyle,

1982) cite several inconsistencies in the criteria used to determine

the incidence of bulimia, the most conservative estimate of diagnos-

able bulimia in the general female population 16-25 years of age is

5% (Johnson.et al., 1984a).

The syndrome of bulimia is reportedly a significant problem

among women in college, with incidence reports ranging from 7.8%

(Pyle et al., 1983) to 19% (Halmi et al., 1981). The "typical"

bulimic woman is college educated, perfectionistic, self-critical

and comes from a family with high achievement expectations (Hazard,

1984; Johnson et al., 1984a). Bulimic symptoms are reported by as2

many as 47% of college women (Sinoway, 1983). Related problems such

as chronic restrictive dieting, negative body image, and chronic

weight preoccupation are all commonly found among college women

(Boskind-Loaahl and Sirlin, 1977; Carter and Moss, 1984; Garner,

Olmstead and Garfinkel, 1983; Nagelberg, Hale, and Ware, 1984). It

is clear that there exists among college women a spectrum of prob-

lems with eating and weight ranging_ in severity from occasional

bringing and body image concerns to a syndrome of frequent bringing

and purging or restricting. In addition there is evidence of

subgroups within the syndrome of bulimia which are thought to differ

along behavioral and psychological continue (Johnson et al.,

1984(a); lacey, 1982; Nagelberg et al. 1984; Sinoway, 1983).

Statement of the Problem

The present study first investigates the differences among

those who have no identified eating problem (normal eaters), those

who only have the symptom of bulimia (binge eating), and those with









the syndrome of bulimia (a pattern of binge eating followed by some

method of compensatory weight control). Second, this study address-

es the issue of sub-types in the syndrome of bulimia. Specifically

the present study separates those with the DSM-III syndrome of

bulimia into two subgroups, purging bulimics (those who purge by

vomiting, using laxatives, or spitting food out) and nonpurging

bulimics (those who do not purge). The behavioral and psychological

characteristics of these two subgroups are compared to each other,

to those who have the synptam of bulimia (binge eaters), and to

normal eaters. In addition, this study is the only large-scale

epidemiological study of female college students which systematical-

ly randomized its sample, and so will provide useful information

about a variety of eating disorder behaviors and weight control

habits of college wamen.

Importance of the Problem

The recent rise in concern over the prevalence of bulimia is

reflected in the increased coverage of the subject in the popular

and professional literature, as well as the increased number of

self-help organizations and inpatient and outpatient treatment

programs. Because the identification of bulimia as a clinical

syndrome is relatively new, there are gaps in knowledge about its

etiology, epidemiology, treatment and prognosis. There remain

inconsistencies in and controversies over the terminology and

criteria used to identify bulimia. Some have suggested that the

problems in classifying bulimia have contributed to the sensation-

alization and/or the overidentification of bulimia in the college

population (Carter and Moss, 1984; Johnson et al., 1984(a);








Nagelberg et al., 1984). Many authors have cited the need for

research that would provide information leading to greater accuracy

and consistency in the description and classification of bulimia

(Abraham and Beumont, 1982; Garner, Olmstead, and Garfinkel,

1983(a); Groh, 1984; Halmi, 1983; Katzman and Wolchik, 1984;

Lepkowsky, 1983; Mitchell, Hatsukami, Eckert and Pyle, 1985;

Mitchell and Pyle, 1982; Sinoway, 1983).

It is the intention of the present study to provide information

that will further the development of meaningful psychological and

behavioral descriptions of bulimia in a female college population.

Delineating the significant psychological characteristics of symptom

and syndrome groups will facilitate the development of appropriate

treatment plans. In addition, the present study extends the current

research by a) studying bulimic symptom and syndrome groups from a

large, randomly chosen nonclinical sample; b) adding to the studies

(Katzman and Wolchik, 1984; Nagelberg et al., 1984; Pyle et al.,

1983; Sinoway, 1983) which have compared groups of binge eaters and

binge-purgers; and c) adding to the information provided by the one

study (Sinoway, 1983) which investigated the differences between

binge-fasters and binge-purgers in a nonclinical sample.














CHAPTER II
REVIEW OF THE IITERATURE

In the following section, the literature will be reviewed on

the following topics:

1. The definition and classification of bulimia, with special

attention to the confusion and controversy around the symptom

and the syndrome, the DSM-III criteria, the criteria used to

identify bulimia in research, and the identification of sub-

groups within the bulimic syndrome;

2. The incidence and parameters of bulimic symptoms within the

obese population;

3. The incidence and parameters of bulimic symptoms within the

anorexic population;

4. The incidence and parameters of bulimic symptoms within the

normal weight population; and

5. The bulimic syndrome in the normal weight population, including

the incidence rate, demographic, family and psychological

profiles, and weight and symptom history description.

Unless otherwise indicated, one may assume subjects used in

reviewed studies were women.

Defining and Classifying Bulimia

There has been a good deal of confusion in the professional

literature surrounding the terminology and criteria used to identi-

fy, classify and diagnose bulimia as a symptom and a syndrome. The









term bulimia first appeared in literature which reported symptoms of

binging in anorexic patients (Beumont, George, and Smart, 1976;

Casper, Eckert, Halmi, Goldberg, and Davis, 1980; Garfinkel,

Moldofsky and Garner, 1980; Russell, 1979; Vandereycken and

Pierloot, 1983). Numerous other terms have been used since to

describe either the symptom or the syndrome, or both. The symptom

of bulimia has been referred to as "binge eating" or "ccupulsive

eating" (Abraham and Beumont, 1982; Dunn and Ondercin, 1981; Halmi

et al., 1981; Hawkins and Clement, 1980; Katzman and Wolchik, 1984;

Mitchell et al., 1981; Ondercin, 1979; Wardle and Beinart, 1981;

Wolf and Crowther, 1983). Some authors, however, use either "binge

eating" or compulsivee eating" to refer to both a symptom and a

syndrome (Dunn and Ondercin, 1981; Ondercin, 1979; Wardle and

Beinart, 1981; Wermuth, Davis, Hollister, and Stunkard, 1977). The

syndrome has been called "bulimia nervosa" (Allerdis, Florin and

Rose, 1981; Carter and Moss, 1984; Casper, 1983; Fairburn and

Cooper, 1982, 1983; Russell, 1979), "bulimarexia" (Boskind-Lodahl

and White, 1978), "dsyorexia" (Guiora, 1967); and "the dietary chaos

syndrome" (Palmer, 1979). Although the terms vary, the syndrome

named is always described as a pattern of binge eating accompanied

by some combination of purging, fasting, or excessive exercise and

dieting. Some authors include as necessary syndrome criteria low

self-esteem (Boskind-Lodahl and White, 1978) and a preoccupation

with weight and food (Palmer, 1979).

Halmi (1983) presented four reasons to classify clinical

syndromes: 1) to allow clinicians to communicate about clinical

phenomena; 2) to conduct and replicate research on clinical phename-









na; 3) to elucidate clinical phenomena and allow classification to

be revised, deleted, or otherwise changed; and 4) to properly design

and evaluate treatment. Diagnostic accuracy in research and treat-

ment is assured only by the establishment of criteria which ade-

quately describe the clinical syndrome.

The diagnostic criteria listed in the DSM-III to classify the

syndrome of bulimia are as follows:

1. Recurrent episodes of binge eating (rapid consumption of a

large amount of food in a discrete period of time, usually less

than two hours);

2. At least three of the following:

a. consumption of high-caloric, easily ingested food during a

binge

b. inconspicuous eating during a binge

c. termination of such eating episodes by abdominal pain,

sleep, social interruption, or self-induced vomiting

d. repeated attempts to lose weight by severely restrictive

diets, self-induced vomiting, or use of cathartics and/or

diuretics

e. frequent weight fluctuations greater than ten pounds due

to alternating binges and fasts

3. Awareness that the eating pattern is abnormal and fear of not

being able to stop eating voluntarily;

4. Depressed mood and self-deprecating thoughts following eating

binges;








5. Bulimic episodes are not due to anorexia nervosa or any known

physical disorder (American Psychiatric Association, 1980, p.

71).

Several authors have cited problems with the DSM-III criteria.

Halmi (1983) notes that the criteria were written before there was

information on demographic valuables or on the course or prognosis

of the syndrome. The criterion excluding bulimic anorexics from the

diagnostic category has been questioned, especially in light of

evidence that they may be more similar to normal weight bulimics

than to restricting anorexics (Garner et al., 1985a). Many studies

have suggested the addition of criteria of frequency and/or severity

of binging/purging/restricting symptoms to better identify clinical-

ly significant levels of bulimia (Garner et al., 1983; Groh, 1984;

Johnson et al., 1984(a); Mitchell et al, 1985; Nagelberg et al.,

1984; Sinoway, 1983). Halmi (1983) and others (Johnson et al.,

1982; Lepkowsky, 1983; Weiss and Ebert, 1983) have warned against

assuming the bulimic syndrome is homogeneous with respect to psycho-

pathology and severity of the illness. Studies using the DSM-III

criteria, however, have supported the synptom-syndrome distinction

and the validity of establishing the bulimic syndrome as a separate

diagnostic category (Halmi et al., 1981; Hazard, 1983; Katzman and

Wolchik, 1984; Mitchell et al., 1985; Sinoway, 1983; Weiss and

Ebert, 1983).

The criteria used in research to identify either the bulimic

syndrome or bulimic symptoms have varied, even with the introduction

of the DSM-III diagnostic category. The use of a variety of crite-

ria has made reports of incidence rates difficult to interpret.









Most investigations of the incidence and parameters of the syndrome

(Allerdis et al., 1981; Fairburn and Cooper, 1983; Halmi et al.,

1981; Johnson et al., 1984(b); Johnson et al., 1982; Katzman and

Wolchik, 1984; Mitchell et al., 1985; Pyle et al., 1983); or the

symptoms (Dunn and Ondercin, 1981; Hawkins and Clement, 1980;

Ondercin, 1979; Wolf and Crowther, 1984) have employed self-report

questionnaires. A few have used interviews or both interviews and

questionnaires to identify individuals with the bulimic syndrome

(Carter and Moss, 1984; Garner et al., 1985(a); Nagelberg et al.,

1984; Pyle et al., 1982; Russell, 1979; Stangler and Printz, 1980;

Weiss and Ebert, 1983). While interviews generally produce lower

rates of incidence of the syndrome, it is possible that the confi-

dentiality of the questionnaire affords subjects more room to

accurately self-disclose without fear of embarrassment (Nagelberg et

al., 1984). The most rigorous definitions of the bulimic syndrome

have used a modified DSM-III diagnosis, including criteria for binge

amount and/or frequency; and/or frequency of compensation methods;

and/or frequency of weight fluctuations (Johnson et al., 1982;

Katzman and Wolchik, 1984; Mitchell et al., 1985). It has become

evident, however, that the amount or type of food eaten is not as

salient to the definition of a binge as are the feelings of guilt

and loss of self-control (Johnson et al., 1985; Wardle and Beinart,

1981). The presence of binge eating as a single synptnm or as part

of the syndrome is generally identified by endorsement of questions

such as "Do you ever engage in periods of uncontrolled excessive

eating commonly called binge eating or binging?" (Hawkins and

Clement, 1980; Wolf and Crowther, 1983), or "Do you frequently









consume large amounts of food in a short period of time other than

meals?" (Katzman and Wolchik, 1984).

The definition and classification of bulimia as a symptom and

as a syndrome is in its infancy. While early researchers have

concentrated on simply describing symptomatic behaviors and their

psychological concomitants, current and future research must address

the complex task of refining the diagnostic criteria for bulimia.

Differentiating clinical and subclinical levels of severity of

pathology, determining the appropriateness of the exclusion of

bulimic anorexics, and identifying subgroups within the syndrome,

are but a few of the tasks facing those studying and treating

bulimia.

Incidence and Parameters of Bulimic Symptoms (Binge Eating)
in the Obese
The study of obesity has frequently searched for psychological

or eating disturbances to "explain" the "causes" of obesity.

Reviews of the research (Wardle and Beinart, 1981; Wooley, Wooley,

and Dyrenforth, 1979) report no consistent psychological, emotional,

or eating pattern differences in obese and other populations. Binge

eating patterns have been found, however, in a proportion of over-

weight and obese individuals.

Stunkard (1959) first used the term binge eating to refer to an

eating pattern he observed in three of 40 obese individuals in a

treatment clinic. He described the binges as periods of "orgiastic"

eating which occurred under stress but with no apparent regularity.

A similar report from Hamburger (1951) on 18 obese patients found

binge eating episodes were triggered by the act of eating itself, as









if the patients were addicted to food. Kornhaber (1970) described a

"stuffing syndrome" of binge eating, depression, and withdrawal in

obese patients. Kornhaber theorized that the binge eating was

triggered by the deterioration of the patient's defense mechanisms

under stress. Studies have found no significant differences between

the obese and nonabese in the type of triggers or amount of food

eaten in a binge, nor in the frequency, duration, or emotional

consequences of binge eating (Gormally, Black, Daston, and Rardin,

1982; Loro and Orleans, 1981; Polivy and Herman, 1985; Spencer and

Fremouw, 1979; Wardle and Beinart, 1981; Wenrmuth et al., 1977; Wolf

and Crowther, 1983).

Two studies of the incidence and severity of binge eating in

those seeking treatment for obesity (Gormally et al., 1982; Loro and

Orleans, 1981) have reported at least 20% of such individuals do not

binge. Neither study found any correlation between degree of

obesity and binge eating severity. Gormally et al. (1982) found of

112 male and female patients 55% were judged to have moderate

problems with binging and 23% had severe bringing problems. Of 280

participants in a weight loss program at Duke University, Loro and

Orleans (1981) found 28.6% binged at least twice per week and 22.1%

binged once per week. The two studies related severity of binging

to early onset of obesity (Loro and Orleans, 1981) and to feelings

of lack of self-control, guilt and self-hate (Gormally et al.,

1982). In addition, other studies (Spencer and Fremouw, 1979; Wolf

and Crowther, 1983) have linked severity of binge eating not to

weight classification but to degree of conscious restraint exerted

over eating.








In summary, the study of binge eating in the obese population

has found binging to be unrelated to degree of overweight, to be

similar to nonobese in the frequency, duration and amount of food

per binge, and to have triggers and emotional consequences similar

to those in the nonobese. Once thought to be the characteristic

eating pattern and main cause of obesity, it is clear that binge

eating is a common pattern found across weight classifications.

/" Incidence and Parameters of Bulimic Symptoms in Anorexia Nervosa

Bulimic symptoms have been reported frequently in the litera-

ture on anorexia nervosa patients. Binge eating was found in 25% of

patients seen by Bruch (1973), 47% of hospitalized patients studied

by Casper et al. (1980), and in 48% of patient histories studied by

Garfinkel, Moldofsky, and Garner (1980). The variability of fre-

quencies may be attributed to diagnostic differences across studies

or to the anorexics' reluctance to acknowledge such a loss of

self-control, or to the tendency of more current researchers to

elicit specific information about binging and vomiting (Wardle and

Beinart, 1981).

Binge-eating in anorexia, unlike that of obesity, is followed

by compensation methods of restricting, purging, and/or exercising.

Several studies have investigated the characteristics of anorexics

who exhibit symptoms of bulimia, including binge eating and the use

of vomiting or purging to compensate. Beumont et al. (1976) com-

pared the histories of anorexic patients who restricted intake to

lose weight ("dieters") with patients who relied primarily on

vomiting and laxative abuse to lose ("vamiter/purgers"). The

dieters were significantly more introverted and less sexually









experienced, had a lower rate of premorbid obesity and had a better

prognosis than the vomiter/purgers. The rate of binging was

markedly different, with 43% of purgers reporting binge eating

ccampared to only 18% of the dieters.

Russell (1979) studied the histories of 30 patients who binged

and purged, of which 26 had met a true or cryptic diagnosis of

anorexia. He compared them with 30 "true" anorexics without bulimic

symptoms. The 30 binge/vamiters, whose illness Russell called

"bulimia nervosa," were on the average heavier, more sexually

active, more acutely depressed and more regular in their menstrual

periods. Russell found binging was triggered by emotional distress

and/or by ingestion of "forbidden" foods. Prognosis was worse for

the vomiter/purger group than the "true anorexic" group, and thera-

peutic attempts to encourage weight gain were ineffective as were

attempts to relieve the "powerful, intractable" urges to binge.

Problems with impulse control, including drug abuse, suicidal

behavior, and kleptomania, have been found more frequently in

anorexics with bulimic symptoms (Casper et al., 1980; Garfinkel et

al., 1980; Garner et al., 1985(a); Strober, Salkin, Burroughs, and

Morrell, 1982; Vandereycken and Pierloot, 1983). Casper et al.,

(1980), in a study of 105 hospitalized anorexics, found those who

binged daily had elevated scores on the MMPI schizophrenia, psycho-

pathic deviate, paranoia, depression and psychasthenia scales.

Garner et al. (1985a) compared three groups of patients: 59 anor-

exics and 59 bulimics with comparably severe bulimic symptoms and 59

restricting anorexics. The two bulimic groups evidenced a similar

pattern of impulse dyscontrol and labile moods which was signifi-








cantly different from the pattern of restricting anorexics. In

addition, both bulimic groups had a greater personal and familial

predisposition to obesity than did the restricting group.

In conclusion, there exist in the group of patients with

anorexia nervosa significant differences between those who only

restrict, and those who binge and/or purge. There are authors who

have proposed that the diagnostic distinction of anorexia and

bulimia is not entirely accurate; many patients fulfill both sets of

criteria within the course of their illness (Casper et al., 1980;

Garner et al., 1985(a); Holmgren et al., 1983; Russell, 1979;

Vandereycken and Pierloot, 1983). Garner et al. (1985a) point to

the similarities of those with bulimic symptoms across weight

classifications and they suggest that anorexia and bulimia should

not be differentiated solely on the basis of weight. Further

studies are needed to understand the psychological profiles within

groups of anorexics and bulimics.

Incidence and Parameters of Binge Eating in the Nonanorexic,
Nonobese (Normal Weight) Population

The following is a discussion of the literature on the symptom

of bulimia (binge eating) in what is usually called the "normal

weight" population of women. Normal weight is defined as nonobese

(obese is 20% or more above standard body weight for height and age)

and nonanorexic (anorexic is 75% or less of standard body weight for

height and age). Studies generally employ as a standard body weight

range for a medium frame, corrected for height and age. Reviews of

the literature on bulimia in normal weight women usually treat arti-

cles on the symptom and articles on the syndrome as if they were a








hmogenous group. The present review has separated the symptom and

syndrome literature and will present evidence to support the

continued differentiation of binge eaters from syndrome bulimics.

Ondercin (1979) published a study of 279 college women recruit-

ed from introductory psychology classes. Episodes of binge or

compulsive eating (overeating not in response to hunger) were

reported by 78% of the women. Ondercin separated the group into

three levels: those who identified themselves as "definite" compul-

sive eaters (high), "sometime" compulsive eaters (medium), and "not"

compulsive eaters (low). Eighteen percent of the group were high

compulsive eaters, 51% were medium, and 30% were low. A chi square

analysis found significant differences among the groups: High

compulsive eaters tended to eat more often in response to unpleasant

affective states, to eat when not hungry, to experience guilt after

overeating, to think more about food, to use food to reduce tension,

and to diet more than the medium or low groups. The low compulsive

eaters tended to binge only a few times a year compared to once or

more a week for the high group. The greatest weight dissatisfaction

and most frequent weight fluctuations were reported by the high

compulsive eaters. Unfortunately no data were reported on the use

of fasting, exercise, or purging to control weight, so it is possi-

ble that some of Ondercin's subjects were diagnosable bulimics

The Binge Scale, a nine item self-report measure of behavioral

and psychological aspects of bringing, was developed by Hawkins and

Clement in 1980. They surveyed 247 normal weight women, 110 normal

weight men, and 26 overweight (110-120% of standard body weight)








wcmen on a college campus. Hawkins and Clement reported 79% of the

women and 49% of the men had engaged in binging. Binge eating

severity was related to dieting concern and dissatisfaction with

physical image for men and wcmen. Binge eating began for the

majority of the subjects between the ages of 15 and 20 and most

could not pinpoint the causes of their binges. Severity of binging

in women was related to major life changes in the last month and

female bingers reported more guilt after binging and more preoccupa-

tion with thoughts of food than did male bingers. The amount of

deviation from ideal weight was not related to binge eating severity

for males or females when degree of dieting concern was statistical-

ly controlled.

Ondercin's (1979) and Hawkins and Clement's (1980) findings

have been supported by two other studies. Dunn and Ondercin (1981)

using Ondercin's (1979) scale compared 23 high and 23 low compulsive

eaters to each other and to a control group on several psychological

tests. They found the low group was similar to the control group,

while the high group showed a greater need for approval, more inner

tension and suspiciousness, greater guilt proneness and less self-

control and emotional stability than either the low or control

groups. Wolf and Crowther (1983) used Hawkins and Clement's (1980)

Binge Scale to identify mild, moderate, and severe binge eaters in a

group of 255 women, half of whom were normal weight (within 90-110%

of standard body weight) and half of whom were overweight (more than

10% over standard body weight). Using a canonical correlation

procedure they found severity of binge eating was independent of

degree of deviation from standard body weight. The following









variables were positively related to binge eating severity: preoc-

cupation with food, concern about dieting, fear of loss of control

over eating, increased body image dissatisfaction, low self-esteem,

and amount of stress experienced in the last year. From these four

studies it is clear that while binge eating is a common practice

among normal weight college women, severe binge eating is generally

related to increased dietary concern, problems with self-esteem and

body image, stressful life circumstances, and eating in response to

unpleasant emotional states.

Wardle and Beinart (1981) reviewed the literature on binge

eating and concluded that while dietary restraint and binging are

not of themselves abnormal, differences might be found in the

binging habits of clinical (patients or clients in treatment) and

non-clinical populations. Pyle et al. (1983) surveyed 1355 male and

female students in freshman English classes at a large midwestern

university. Their questionnaire included questions that allowed

them to identify students who met DSM-III criteria for bulimia. In

addition, the same questionnaire was administered to 37 bulimic

outpatients in treatment at an eating disorder clinic. Binging

habits and weight control methods were compared between the bulimic

patients and female bulimic student group, and between the female

bulimic students and the female nonbulimic students. The results

provide some insights into how the binging habits of nonbulimic

women compare with a clinical and non-clinical sample of bulimic

women.

Pyle et al. found 4.5% of women surveyed met DSM-III criteria

for bulimia (a criterion of at least weekly binging was added). Of








the nonbulimic women 57.4% admitted to binge eating, 17.2% at least

weekly. Bulimic students reported significantly more frequent 24

hour fasting, more frequent impulse-related problems (alcohol/drug

abuse and stealing), and a greater tendency to be overweight than

nonbulimic students. All bulimic patients, 77.8% of bulimic stu-

dents, and 62% of nonbulimic students reported a fear of fat.

Feeling depressed after binge eating was reported by 39.1% of

nonbulimic students. The use of weight control methods such as

laxatives, diuretics, vomiting, and exercise were reported by 47% of

nonbulimic students, although the majority used them less than

weekly. Over half of the bulimic students and all of the bulimic

patients induced vomiting at least weekly. Bulimic patients had a

higher frequency (48.6%) of previous treatment for anorexia or

bulimia than either the bulimic students (8.8%) or the nonbulimic

students (1.3%). Pyle et al. indicate that while binge eating, fear

of fat, self-induced vomiting, laxative and diuretic abuse, and

guilt from overeating are all found in a number of college women,

there are significant differences between binge eaters and either

clinical or nonclinical groups of bulimics.

Nagelberg et al. (1984) explored the differences among at least

weekly bingers (N=14), binge-vomiters (N=10), and controls (N=7) all

recruited from a sample of 244 women in introductory psychology

classes. Bingers had more restrained eating and more compulsive

eating than controls. Purgers differed significantly from bingers

by scoring lower on a measure of self-discipline and regard for

social demands.








A study by Katzman and Wolchik (1984) found similar differences

between bingers and controls but found greater dissimilarity than

Nagelberg et al. between bingers and bulimics. Of the 80 female

introductory psychology students used as subjects, 30 were identi-

fied as meeting DSM-III criteria for bulimia, with added criteria of

a 1200 calorie per binge minimum and more than 8 binges/month.

Binge eaters (N=22) were those with more than 8 binges/month but who

did not satisfy all DSM-III criteria, and controls, (N=28) were

those with no binge eating problem. Binge eaters had more compul-

sive eating and a greater preoccupation with dieting than did the

controls. All three groups' average ideal weights were thinner than

the standard body weight for their age and height. The bulimic

group exhibited a greater preoccupation with dieting, lower self-es-

teem, poorer body attitude, greater depression and a greater need

for approval than the binge eaters. Bulimics also had a greater

average number of calories per binge, a higher incidence of previous

psychological treatment, a greater interest in treatment, more life

disruption from their eating problems, and a more frequent history

of anorexia nervosa. Katzman and Wolchik concluded the syndrome of

bulimia is associated with a significantly greater amount of psycho-

logical disturbance than is the symptom of bulimia. They issued a

warning against applying results from studies of binge eaters to

bulimics.

Finally, a study by Sinoway (1983) investigated the differences

between nonbingers and bingers in a female college population.

Instead of screening subjects on the basis of binging frequency, the

women were classified as bingers if they ate a certain amount of









food during a binge and if they felt unable to stop or control their

binging. Of the 1172 subjects who answered, 11.6% were bingerss

only" and 13.7% were "binge-revisionists" (they binged and dieted,

fasted, or purged). Several discriminant function analyses were

performed to determine which variables best discriminated among the

groups. Degree of dietary restraint/concern and eating in response

to emotions were the most salient discriminative variables, with

binge-revisionists scoring highest, bingers next highest, and

controls the lowest. On the basis of these results, Sinoway con-

cluded that those who have binge eating problems are easily differ-

entiated from those who do not, and those who binge can be easily

separated into two groups: bingers-only and binge-revisionists.

In conclusion, the research supports the distinction between

bingers and nonbingers, and between those who binge and those with

the syndrome of bulimia. Bingers-only exhibit more preoccupation

with dieting and more compulsive eating habits than do nonbingers.

Severity of binge eating has been related not to weight classifica-

tion but to eating in response to emotions, increased preoccupation

with dieting, problems with self-esteem, body dissatisfaction,

stressful life circumstances, and early onset of weight problems.

Those identified as meeting DSM-III criteria for bulimia show

greater behavioral and psychological problems than either bingers or

nonbingers. The following section will further address the inci-

dence and parameters of the syndrome of bulimia.

The Bulimic Syndrome in the Nonanorexic, Nonobese
(Normal Weight) Population

This section will be divided into five subsections: incidence,

demographics, weight and symptom history, psychological profile, and








family profile. Each subsection will describe the major findings of

research studies in that topic, and where appropriate will separate

the findings in the clinical population from those in the nonclini-

cal population. Following the five subsections, this section of the

literature review will be summarized and the hypothesis of the

present study will be proposed.

Incidence

As has been previously mentioned in the present paper, inci-

dence rates of the bulimic syndrome have differed widely, primarily

due to the use of a variety of population samples, sampling methods,

and criteria to define the syndrome. The majority of studies have

used college student samples. Halmi et al. (1981) surveyed 355

female and male summer school students with a self-report question-

naire of eating and weight control habits using the DSM-III crite-

ria. Thirteen percent were identified as bulimic, 87% of whom were

female (19% of the total female population), and 13% of whom were

male (5.9% of the total male population). The diagnostic criteria

did not include any minimum binge or purge frequency. A more

conservative questionnaire used by Pyle et al. (1983) to survey 1355

freshmen college students found 4.1% of the total population (7.8%

of the females and 1.4% of the males) merited a probably diagnosis

of bulimia. When a criterion of at least weekly binge eating was

added, 2.1% (0.4% of males and 4.5% of females) met all criteria.

And the added criterion of weekly binging and purging reduced the

frequency to 0.6% of the total (1.0% of the female and 0.3% of the

male population). Johnson et al. (1984b) used criteria similar to

those of Pyle et al. in a survey of 1268 female high school

students.








Using all DSM-III criteria and added criterion of at least weekly

binge eating, Johnson et al. certified 4.9% as bulimic. They

qualified their findings however, by noting that the conservative

criteria probably resulted in a number of false negatives, thus

considerably reducing the reported incidence of bulimia.

Freshmen college students at a large university were surveyed

by Sinaway (1983), who found of 1172 women, 13.7% reported binging

followed "always" or "often" by purging, fasting, or dieting. Binge

frequency was not a criterion for identification of bulimia.

Stangler and Printz (1980) reviewed the DSM-III diagnoses from the

records of 500 patients at the University of Washington Student

Psychiatric Clinic and reported 3.8% of the sample were diagnosed

bulimic. Not unlike the findings of Halmi et al. (1981), 89.5% of

those diagnosed bulimic were wcmen, while 10.5% were men. Carter

and Moss (1984) screened for bulimia with a self-report

questionnaire of eating problems, then interviewed those with a

probable diagnosis of bulimia. Although 16.7% of 162 female

introductory psychology students reported binge eating, only 2.45%

were diagnosed bulimic through subsequent interviews. Nagelberg et

al. (1984) suggested that while interviews produce a more

conservative estimate of the incidence of bulimia, the anonymity of

a questionnaire may afford more accurate report from subjects, who

generally are acutely embarrassed and ashamed about their binging

and purging.

A rare study of women outside the college environment was

conducted by Fairburn and Cooper (1983). A questionnaire was given

to 40 normal weight female patients under age 40 at a family plan-

ning clinic in England. Of the 384 respondents, 1.9% filled all








three criteria for bulimia nervosa: 1) powerful, intractable urges

to overeat; 2) avoidance of fattening effects of overeating by the

use of vomiting or purgatives; and 3) a morbid fear of fat. No

criteria for frequency of binge eating or purging were used.

Demographics

The demographic profile of the bulimic woman is remarkably

consistent across studies. Fairburn and Cooper (1983) surveyed 669

women who responded to a British magazine requesting volunteers for

a study of women who induce vomiting to control their weight. Of

the 499 respondents who filled the criteria for bulimia nervosa,

24.8% were students, 20.7% were married, 5.1% were housewives, 12.9%

were in the medical professions, and 3.4% were in the food profes-

sions. The average age was 23.8 years. Johnson, Stuckey, Lewis,

and Schwartz (1982) conducted a similar survey of women who had

requested information from an eating disorder clinic. Of the 509

respondents, 316 met the DSM-III criteria with an added criterion of

having binged at least once in the last 6 months. The majority were

single, Protestant, college-educated, and lived with parents or a

partner. The average age was 23.7 years and nearly half were

currently students. Their families of origin usually had more than

one child and their fathers were usually professionals. Other

studies (Sinoway, 1983; Weiss and Ebert, 1983) have also found

bulimic subjects tend to be white, single, between the ages of 22

and 26, college-educated, and from middle to high income families. -

Demographic profiles of bulimics from outpatient treatment

programs are similar to those from nonclinical populations. The

majority of patients are female, white, single, between the ages of








22 and 26, college-educated, and from middle to high income family

with at least 2 children (Abraham and Beumont, 1982; Allerdis et

al., 1982; Mitchell et al., 1985; Pyle, Mitchell, and Eckert, 1981).

Hazard (1983) compared bulimic and normal women and found the

bulimic group's family social status was higher because of the

greater educational and occupational level of their mothers. No

other study has reported this demographic difference between

bulimics and nonbulimics.

Description of Weight and Symptom History

Nonclinical population

Bulimics in community and college nonclinical samples (i.e. not

recruited primarily from treatment sittings) tend to be within

normal body weight range, to have a history of being overweight and

to begin eating disordered behavior between the ages of 15 and 18.

Fairburn and Cooper (1983) found 83.2% of their magazine ad study

and 86% of their family clinic study subjects were 85 115% of

standard body weight for their height and age. The average age at

which binge eating began in the magazine ad subjects was 18.4 years,

with a mean duration of binging of 5.2 years. Vomiting began at an

average age of 19.3 years with a mean duration of 4.5 years. Nearly

one half (45.2%) had a highest body weight since menarchs of >115%

of standard body weight, and 43% had a lowest body weight of <85% of

normal body weight. Significantly fewer of the family clinic sub-

jects had a history of overweight. Binging and vomiting were less

frequent in the family clinic sample with 1% vomiting at least

weekly and 7.3% binging at least weekly, while 56.1% of the magazine

ad sample vomited daily and 27.2% binged daily. The rates of








binging and vomiting in two samples studies by Johnson and associ-

ates showed differences similar to those found by Fairburn and

Cooper. Of the female high school samples surveyed by Johnson et

al. (1984b) 17% reported vomiting weekly, 21% binged weekly, and 11%

used laxatives weekly. In the sample of women surveyed by mail,

Johnson et al. (1982) found 51% binged daily, 59% vomited daily, and

25% used laxatives daily. The weight distribution was similar to

Fairburn and Cooper' 's magazine ad sample, with 81.6% within under-

weight to normal body weight range (76-110% of norm), 17.5% over-

weight (>110% of norm), and 0.9% anorexic (<76% of norm). Like

Fairburn and Cooper's subjects, half of Johnson et al's (1982)

subjects had a history of overweight (>110% of norm) and 5% had a

history of anorexic weight.

Halmi et al. (1981) and Pyle et al. (1983) both investigated

the bulimic syndrome in college women. Halmi et al. found while

9.9% of the male and female population had ever induced vomiting,

only 2.5% vomited one or more times per month, and 1.7% vomited one

or more times per week. Pyle et al. reported 1% of all female

students surveyed had weekly binge/purge episodes. Pyle et al.

ccmpared female students with the bulimic syndrome to female bulimic

outpatients. The frequency of daily binging (15.5%) and daily

vomiting (8.9%) in the bulimic students was significantly lower than

the frequency of daily binging (89.2%) and vomiting (79.6%) of the

bulimic patients. The bulimic students used 24 hour fasting more

frequently than they used any other weight control while the pa-

tients used vomiting most frequently.

Weiss and Ebert (1983) compared a nonclinical bulimic sample

with a control sample and found no difference in average height,








weight, or weight history. This conflicts with findings that

bulimics more frequently have a history of overweight than normal

(Fairburn and Cooper, 1983; Halmi et al., 1981; Pyle et al., 1983).

Not unlike previous findings, Weiss and Ebert reported subjects'

average age at the onset of bulimia was 19.7 years with an average

duration of 6.4 years. Bulimics tended to have chaotic eating

patterns, eating fewer times a day with longer times between meals

and a smaller variety of foods than did controls.

Clinical population

Subjects from both clinical and nonclinical samples have

reported binge eating is triggered by negative affective states,

restrictive dieting, and eating forbidden foods (Abraham and

Beumont, 1982; Johnson et al., 1982, Johnson and Larson, 1982;

Mitchell et al., 1985; Pyle et al., 1981; Russell, 1979). Bulimic

subjects from both types of samples also have reported eating fewer

regular meals, fasting more, feeling more guilty and worried after

eating, having a greater history of being overweight, having more

menstrual difficulties, and being more preoccupied with thoughts of

food and eating than nonbulimic subjects (Abraham and Beumont, 1982;

Hazard, 1984; Johnson et al., 1982; Johnson and larson, 1982;

Mitchell et al., 1985; Pyle et al., 1983; Pyle et al., 1981; Weiss

and Ebert, 1983). The age of onset of bulimia found in clinical

samples averages 18 years old (Garner et al., 1985; Mitchell et al.,

1985; Pyle et al., 1981; Russell, 1979), not unlike the average

reported in nonclinical samples.

Several studies of bulimic outpatients have reported average

weekly frequencies of binge eating and vomiting caaparable to those








found in the nonclinical sample of Johnson et al. (1982). Garner et

al. (1985a) reported 46% of normal weight bulimic patients binged at

least daily, 73% vcmited, and 47% abused laxatives. The most common

bulimic pattern found by Mitchell et al (1981) was at least daily

binging accompanied by self-induced vomiting (an average of 11.7

binge/purge episodes per week). Laxative use has been reported less

frequently than vomiting by all researchers, followed in order of

decreasing frequency by exercise, diuretics, and diet pills.

Studies of outpatients have also reported physical complications

from frequent binging and vomiting, including dizziness, headache,

fatigue, sore throats, swelling of parotid glands, and dental

problems (Abraham and Beummnt, 1982; Hazard, 1984; Pyle et al.,

1981; Russell, 1979).

The onset of bulimia has been reported by patients and nonpa-

tients alike to correspond with an increased concern about weight

and dieting, or with a stressful or traumatic event, or with in-

creased difficulty dealing with emotions (Abraham and Beumont, 1982;

Johnson et al., 1982; Mitchell et al., 1985; Pyle et al., 1981).

Johnson and larson (1982) conducted a unique study to compare the

daily food-and-mood related experiences of bulimic patients and

controls. All subjects carried beepers which went off at random

times within 2 hour periods. In addition bulimics recorded thoughts

and feelings before, during and after binge/purge episodes. The

results showed that the average prebinge experience included feel-

ings of tension, increased appetite, and loss of control. During

the binge subjects reported an increase of negative feelings and a

decrease in hunger. The purge was accompanied by less anger and








hunger and greater alertness, guilt, and feelings of adequacy and

self-control. After the purge subjects felt even less anger and

hunger, less alertness, and more self control. Johnson and larson

suggested that prolonged restrictive eating, accompanied by diffi-

culties in regulating affective states, predispose individuals to

binge eat. Terrified at the prospect of weight gain such individu-

als begin to purge after binges. While the binge is at first the

primary reinforcing symptom because it discharges emotions and

satisfies food cravings, the purge eventually becomes more reinforc-

ing because it "undoes" the effects of the calories ingested during

the binge.

Psychological Profile of Bulimics

Nonclinical population

Bulimic women in nonclinical samples have been characterized as

having greater concern about dieting, more compulsive eating, more

interpersonal sensitivity, anxiety, and depression, more negative

body image and self-esteem, more negative relationships with family

members, and greater achievement expectations and need for approval

than nonbulimic women (Johnson et al., 1984 (b); Katzman and Wolchik,

1984; Weiss and Ebert, 1983). Problems with impulse control (evi-

denced in stealing and alcohol/drug abuse) and suicidal thoughts and

behaviors have also been found more frequently in bulimics (Johnson

et al., 1982; Pyle et al., 1983; Weiss and Ebert, 1983). Pyle et

al. (1983) reported that while bulimic college students had a

greater history of stealing and alcohol/drug abuse than nonbulimic

students, they did not have as frequent a history as bulimic outpa-

tients of stealing or treatment for depression and eating disor-









ders. Bulimics from clinical samples may have on the average more

severe behavioral and psychological disturbances than bulimics from

non clinical samples. Even so, Johnson et al. (1982) found that of

a nonclinical sample of 316 bulimics, 14% had made suicidal

gestures, 5% reported serious suicide attempts and 90.1% reported

their thoughts and feelings about themselves were completely

affected by their eating disorder. Of those subjects 56% had not

attempted to seek professional help. Weiss and Ebert (1983)

reported that while bulimics and controls did not differ in number

of life stressors or early separations, bulimics had significantly

more depression, psychiatric hospitalizations and suicide attempts.

Clearly there is a continuum of severity of psychological

disturbance found among bulimic women in nonclinical and clinical

populations. Most of the information on psychopathology of bulimia,

however, has been obtained from samples in treatment settings.

Clinical population

Psychological disturbances common to bulimic patients include

depression, impulse dyscontrol, negative self and body image,

borderline personality features, alienation from others, and a

variety of adjustment problems (Allerdis et al., 1981; Flanagan,

1983; Garner et al., 1985(a); Hazard, 1984; Hutton, 1984; Johnson

and Berndt, 1983; Johnson-Sabine, Wood, and Wakeling, 1984;

Lepkowsky, 1983; Mitchell et al., 1985; Pyle et al., 1981; Russell,

1979). Several authors who have administered the MMPI to bulimic

patients have found clinical deviations or either the D (Depression)

scale or the Pd (Psychopathic deviate) scale or both, indicating








tendencies towards serious depression and problems with impulse

control (Flanagan, 1984; Hazard, 1983; Pyle et al., 1981).

Flanagan (1984) conducted a study of psychopathology in bulimia

by comparing the MMPl and Rorschach profiles of a bulimic group to

two control groups. Flanagan found the bulimic profiles indicated a

lack of self-awareness, an inability to control impulses, and

greater tendencies than controls toward depression and dissociation.

Flanagan theorized that as bulimics dissociate from their feelings,

their cognitive processes loosen and they become unable to meet

their needs or control their obsessive thoughts except through the

discharge of impulse energy, i.e. through binging, purging, steal-

ing, alcohol/drug abuse, etc. According to Flanagan, they became

trapped in the bulimic cycle because they lack both an awareness of

their needs and the ability to effectively meet those needs.

Hutton (1984) used a form of the Role Construct Repertory test

developed by George Kelly (1955) to compare bulimics' and normals'

perceptions of themselves and others. Bulimics' perceptions of

themselves were more closely tied to their body size. On the

average, a bulimic's worst self was associated with herself at her

highest weight, and bulimics considered "fat" a more negative

self-descriptor than did normals. In addition bulimics saw them-

selves as farther away from their ideals of self, female, and male.

Negative self and body perceptions were also found in bulimic pa-

tients by Garner et al. (1985a). Using the Eating Disorder Invento-

ry (Garner, Olmstead, and Polivy, 1983b), Garner et al. (1985a)

found normal weight bulimics had elevated body dissatisfaction,

depression, and lowered feelings of self-efficacy. Compared to a








group of restricting and a group of bulimic anorexics, the normal

weight bulimics showed equivalent or greater disturbance on all EDI

scales except Interpersonal Distrust. Lepkowsky (1983) found

similar personality disturbances in anorexics and bulimics on

measures of personality organization and personality disorders.
Both groups showed more psychopathology than a group of neurotically

depressed female subjects. Bulimics had more borderline personality

features as measured by the Conte Borderline Syndrome Index (Conte,

Plutchnik, Karasu, and Jerrett, 1980) than either the depressed

neurotic or the anorexic group.

Features of borderline pathology such as impulsivity, aliena-

tion, and relationship difficulties have been reported by Mitchell

et al. (1985), Johnson and Berndt (1983), Pyle et al. (1981), and

Russell (1979). Of 275 patients seen by Mitchell et al. (1985),

34.4% had a history of problems with alcohol and drugs and 17.7% had

been treated for chemical dependency. Bulimia had caused problems

with family members for 60.7% of the subjects, 69.5% had problems

with intimate relationships, 53.3% had problems with finances, and

49.6% had difficulty with work because of bulimia. Of 34 patients

seen by Pyle et al. (1981), over half had a history of stealing

since becoming bulimic, 23.6% had been treated for chemical depen-

dency, and 44.1% reported they were currently stealing food.

Russell (1979) reported serious symptoms of depression and a history

of suicide attempts in nearly half of his patients. Johnson and

Berndt (1983) found the average bulimic functioned at a significant-

ly more disturbed level than a control group in work, social situa-

tions, leisure, family activities, and in overall level of adjust-









ment. The bulimic groups' functioning was comparable in disturbance

to a group of female alcoholics.

Virtually all studies of the psychological functioning of

bulimic patients have found symptoms of depression in their samples.

The literature is currently beset by a controversy over whether or

not bulimia is simply a manifestation of a primary affective disor-

der (Hudson, Iaffer, and Pope, 1982; Hudson, Pope, Jonas and

Yurgelew-Todd, 1983b). Support for this proposition has come from

reports of depressive symptoms in patients, from depressive illness-

es in family members of bulimics, and from studies of the use of

antidepressant medication with patients (Hudson, Laffer, and Pope,

1982; Hudson et al., 1983(b); Pope, Hudson, Jonas, and Yurgelew-

Todd, 1983; Strober et al., 1982). Several authors have suggested,

however, that it would be premature to assume on the basis of the

present data that bulimia and depressive disorders show the same

etiology (Brotman, Herzog, and Woods, 1984; Cooper and Fairburn,

1983; Garner, Rockert, Olmstead, Johnson, and Coscina, 1985(b),

Viesselman and Roig, 1985). While a number of patients with bulimia

respond to antidepressants, many do not respond, or respond in part

(i.e. the depression lifts but they continue to binge and purge).

Further investigation is needed to clarify the link between depres-

sion and bulimia.

Two studies have provided important information about mood

states of bulimics. Johnson and Larson (1982) compared the reports

of mood states of bulimics and controls and found bulimics' moods

were significantly more variable and more negative. Bulimics

reported being alone more and feeling more sad, lonely, weak,








irritable, passive, and constrained. The authors suggested that

bulimics binge in order to regulate their fluctuating negative mood

states over which they feel they have no control. They then lose

control over the binging (i.e. cannot stop and begin to gain weight)

and they purge in order to restore their sense of control. Similar

findings were reported by Johnson-Sabine et al. (1984) in a study of

50 outpatients with bulimia nervosa. Subjects' moods were assessed

for 8 weeks by self-report and by interviews. Negative mood states

in bulimics were immediately relieved by singing and purging.

Increases in the number of negative moods were associated with

increases in abnormal eating behaviors. The authors concluded that

the dysphoria found in bulimics is related to abnormal eating

symptoms and generally is not evidence of a primary depressive

illness.

Subgroups

Based on their studies of behavioral and psychosocial charac-

teristics of the normal weight population of bulimics, several

authors have proposed the existence of subgroups within the bulimic

population. Abraham and Beumont (1982) found their sample of

bulimic patients could be divided into groups along a number of

dimensions, including the presence or absence of vomiting, duration

of the illness, and presence or absence of an anorexic history.

When divided into "vomiters" and "nonvomiters," vomiters were found

to have a longer history of binging and to feel more relief from

anxiety and negative feelings after a binge. Halmi et al. (1981)

suggested that self-induced vomiting accompanied by laxative abuse

represents a more chronic and severe form of bulimia. Pyle et al.








(1983) made a similar discovery in their comparison of bulimic

students and bulimic outpatients. They suggested that since the

students (whose functioning was less disturbed than the outpatients)

fasted more compared with the patients who purged more often for

weight control, binging and fasting may represent an earlier more

mild form of bulimia.

Sinoway (1983) hypothesized significant differences among a

nonclinical sample of binge/fasters, binge/dieters, and

binge/purgers. Although Sinoway's hypotheses were not significantly

substantiated, certain trends were found and offered as directions

for future exploration. On the average binge/purgers were more

dysfunctional than the binge/dieters or binge/fasters. Of the three

groups, binge/purgers scored highest on a measure of dietary concern

and restraint; they binged most often, they voiced the most concern

about their binging; they stole binge food most frequently; they

were the most secretive about their bulimic behavior; and they

evidenced the most severe depression after bringing.

Lacey (1982) and Johnson (1985) characterized subgroups of

bulimic patients based on the patients' psychological characteris-

tics. The 70 outpatients seen in Lacey's clinic were separated into

three groups. The first and largest, the "neurotic group," tended

to be ambitious and to have high academic achievements, to have

labile emotions (within a normal range), and to present with depres-

sion which turned to anger as the eating symptoms were removed. The

second group, the "personality disordered group," often needed

inpatient treatment for the underlying psychopathology. This group

tended to have a history of impulse control problems and to present








as emotionally shallow and histmonic. The last and smallest group

Lacey called the "epileptiform group." Their bouts of bulimia were

due to diagnosable epilepsy and the bulimia disappeared when the

epilepsy was controlled.

Johnson (1985) described a method of subgrouping patients along

a continuum of personality structure ranging from borderline

personality organization to identity conflicted. The first sub-

group, the borderline personalities, Johnson described as needing

structured therapeutic interventions to help them develop the

intrapsychic structure they lack. These patients have had multiple

impulse control problems throughout their lives. Their eating

disorders and other impulse problems are the result of their inabil-

ity to control their internal chaos. The second group of patients

Johnson called the "false self" group, consisting of patients less

disturbed than the borderline group but still requiring long-term

individual psychotherapy for full recovery. This group of patients

has developed a pseudcmature adaptation in response to premature

demands to separate and become autonomous. The "false self" has

served as a facade of competence and control which hid the fearful

and needy inner self. Bulimia has become a way for them to lose

their tight control and to regulate their affect without risking

exposure to others. Less disturbed than either the borderline or

false self group is the third group, the identity conflicted pa-

tients. These individuals have the "healthiest" intrapsychic

structure and often become bulimic in reaction to adjustment diffi-

culties encountered in adolescence. Their relationships and affec-








tive states are more appropriate than the previous groups and these

patients are more amenable to short-term symptom-focused therapy.

Family profile

Research into the families of bulimics has been sparse, and

most of it has relied on self-report from samples of outpatients.

There are, however, some important trends in the literature.

Flanagan (1983) reported eating and weight-related problems in 64%

of family members of bulimic subjects. Bulimia in an immediate

family member was reported by 21% of the subjects, and 50% identi-

fied at least one family member with depression. In study of 34

patients Pyle et al. (1981), 67.6% reported obesity in at least one

first-degree relative, over half reported obesity in at least one

parent, 11.7% reported bulimia in at least one family member, and

47% indicated depression in at least one first-degree relative. In

addition, 50% reported alcoholism in at least one first-degree

relative.

Depression and substance abuse disorders in families of

bulimics have been identified in studies by Hudson, Pope and Jonas

(1983a); Stern, Dixon, Nemzer, Lake, Sansone, Smeltzer, Lantz, and

Schrier (1984); and Strober et al. (1982). The study by Strober et

al. (1982) study used an inpatient anorexic bulimic sample and its

generalizabilty to normal weight bulimics and their families may be

limited. In light of findings by Garner et al. (1985a) that normal

weight and anorexic bulimics have comparable profiles, the study by

Strober et al. is included as part of this discussion.

Strober et al. compared the MMPl profiles and clinical inter-

views of the parents of 35 bulimic and 35 restricting anorexics.








Affective and substance abuse disorders were more common in the

relatives of the bulimic group. Bulimics' fathers were character-

ized significantly more than restrictors' fathers by signs of

personality trait disorders: hostility, immaturity, and impulse

control problems. Bulimics' mothers showed greater depression,

hostility, and dissatisfaction with intrafamilial relationships than

did mothers of restrictors. The three parental MMPl scores which

significantly predicted the severity of the daughters' bulimia were

maternal depression, paternal depression, and paternal psychopathic

deviance.

Stern et al. (1984) interviewed a group of bulimic patients, a

group of controls, and both groups' parents to determine the preva-

lence of psychiatric disorders in their first and second degree

relatives. The results indicated a significantly higher frequency

of psychiatric disorders in the relatives of bulimics. Substance

abuse disorders, but not affective disorders, were significantly

more frequent in relatives of bulimics. No significant difference

was found between control and bulimic subjects in their morbid risk

for affective disorders.

Hudson et al. (1983a) studied the family histories of 75

bulimic patients (20 of wham had a previous diagnosis of anorexia

nervosa). They evaluated patients' reports of 350 first-degree

relatives and found the morbid risk factor for affective disorder

was not significantly different from the risk factor for a group of

bipolar disorder patients. Bulimics did have a significantly higher

morbid risk factor for affective disorders than groups of

schizophrenia and borderline personality disordered patients.








Viesselman and Roig (1985) interviewed 39 hospitalized

binge/purgers, 36 hospitalized binge eaters, and 13 hospitalized

anorexics to compare their family histories. Binge/purgers had a

significantly higher frequency of maternal depression, but were no

different from other groups in the frequency of depression or

alcoholism in other family members. Compared to the family histo-

ries of a group of alcoholics, binge/purgers tended to have more

family members with depression, and to have more frequent alcoholism

in their mothers and sisters. Viesselman and Roig concluded that

while the sympts of depression in anorexia and bulimia may resem-

ble those found in affective disorders, the family histories of

eating disordered patients support the diagnostic distinctions

between affective and eating disorders.

While the family histories of bulimics indicate problems with

depression and alcoholism, other studies have found disturbances in

family relationships. Johnson and Flach (1985) administered a

family environment scale to a group of normal weight bulimic pa-

tients and to a group of controls. In addition to reporting a

higher achievement orientation in their families than did the

controls, bulimics rated their families as more conflicted, less

supportive, less encouraging of independence, less encouraging of

open expression of feeling, and less emphasis on recreational,

intellectual, and social activities. The two family variables which

best predicted the severity of bulimia were family disorganization

and high achievement orientation. A similar characterization was

reported by Garner et al. (1985a) who studies the difference between








family ratings of bulimics and restricting anorexics. Anorexic

bulimics and normal weight bulimics perceived significantly more

disturbance in their families than did the restricting anorexic

group. Bulimics rated their families as more conflicted on 6 of 7

family functioning scales, suggesting bulimics perceived their

families as having more impulse problems. In addition, Nagelberg et

al. (1984) found women who binge and purge feel they have less worth

as family members than do controls.

Summary
The studies reviewed thus far represent a relatively new body

of literature on a complex multidetermined syndrome of rising

prevalence in college-age women. The literature contains inconsis-

tencies and gaps, however, that call for a clearer delineation of

the bulimic syndrome. First, inconsistencies in terminology have

led to the use of the term "bulimia" to refer to both a symptom

(binge eating) and a syndrome (a pattern of repeated binge eating

followed by efforts to compensate for those binges using various

weight control methods). Current research suggests that those with

the symptom and the syndrome differ in important ways, but only one

study has carefully and systematically explored those differences

(Katzman and Wolchik, 1984). Other studies (Nagelberg et al., 1984;

Sinoway, 1983) have compared bingers-only and binge/faster-dieters

with those who binge and purge, but they did not use full DSM-III

criteria to define their groups. Further research is needed to

delineate the ways in which those with the DSM-III syndrome of

bulimia differ from those with binge eating, or the symptom of

bulimia.





41


A second gap in the research on the bulimic syndrome is the

absence of a study which has used a large-scale random sample of

college women. Halmi et al. (1981), Pyle et al. (1983) and Sinoway

(1983) all used sizable samples, but they used only summer school or

freshmen students. The incidence of binge eating, of various weight

control methods, and of the bulimic syndrome in the female college

population has yet to be established with a large, randomly chosen

cross-section of female undergraduates.

A third area of the literature which needs exploration is the

existence of subgroups within a nonclinical sample of DSM-III

bulimics. Reports from clinicians (Abraham and Beumont, 1982;

Lacey, 1982; Johnson, 1985) indicate that within the syndrome of

bulimia, patients may be separated into subgroups of purgers and

nonpurgers, or into subgroups of neurotic and borderline personality

features. Sinoway (1983) found trends that suggested binge/purgers

(in a nonclinical population) were more dysfunctional (more impul-

sive about eating and stealing, and had more difficulties adjusting

because of their eating problems) than binge/dieters for faster.

Sinoway's groups, however, were not formed according to full DSM-III

criteria, and the differences found among the groups just approached

statistical significance. The present author's clinical experience

confirms evidence from other authors (Abraham and Beumont, 1982;

Pyle et al., 1983) that within the syndrome of bulimia, purging

bulimics have greater self-esteem deficits, greater interpersonal

difficulties, and more problems with impulse control than do non-

purging bulimics. No study as yet has looked for those subgroup

differences in a nonclinical sample of DSM-III syndrome bulimics.








A fourth area into which further inquiry is needed is the

degree to which bulimics in a nonclinical sample exhibit borderline

personality characteristics. The results of several studies indi-

cate that bulimic women from nonclinical populations have greater

self-esteem deficits, depression, problems with impulse control and

suicidal thoughts and gestures than do nonbulimics and binge eaters

(Johnson et al., 1982; Johnson et al., 1984 (b); Katzman and Wolchik,

1984; Pyle et al., 1983; Weiss and Ebert, 1983). Several studies

have found borderline personality features (impulse dyscontrol,

chronic emptiness, lack of consistent self-identity, impaired

interpersonal relationships, alcohol/drug abuse history) in bulimic

patients (Flanagan, 1984; Johnson and Berndt, 1983; Lepkowsky, 1983;

Mitchell et al., 1985; Pyle et al., 1981; Russell, 1979). Lepkowsky

(1983) using the Conte Borderline Syndrome Index (Conte et al.,

1980) found bulimic patients had significantly greater borderline

personality pathology than anorexic, depressed, or normal groups.

Several authors have indicated that binge/purgers in particular have

more problems characteristic of the borderline personality, includ-

ing affective instability, low self-esteem and depression, and lack

of control over impulses than do normals (Flanagan, 1984; Johnson

and Flach, 1983) or bingers only (Katzman and Wolchik, 1984;

Nagelberg et al., 1984). In summary, while there are indications

that a) bulimics from nonclinical samples exhibit greater difficul-

ties characteristic of the borderline personality than normals or

bingers, and b) purging bulimics have particular difficulty with

problems characteristic of borderline disturbance, no study has yet

tested those suggestions.








The intention of the present study was to extend the current

literature on the syndrome of bulimia to the four areas outlined

above: the synptam-syndrame differences; the incidence of binging,

various weight control methods, and the bulimic syndrome in a random

sample of college women; within the syndrome of bulimia, the differ-

ences between subgroups of purging and nonpurging bulimics; and the

level of borderline personality characteristics in bulimics and in

bulimic subgroups in a nonclinical sample. To investigate these

areas, four groups of nonobese, nonanorexic (normal weight) wamen

were selected from a random sample of college undergraduates.

Subject selection was performed using a questionnaire adapted from

those of other authors to identify bulimic subjects. The following

four groups were studied: a) women with no identified eating

problems (called the normal eater or control group); b) women who

identified themselves as having a problem with binge eating but who

neither met all of the DSM-III criteria for bulimia nor used vomit-

ing or laxatives for weight control (called the binge eater or

symptom group); c) women who fit the DSM-III criteria for the

syndrome of bulimia and who used self-induced vomiting and/or

laxatives for weight control (called the purging bulimic or purger

group); and d) women who fit the DSM-III criteria but did not use

vomiting or laxatives for weight control (called nonpurging bulimics

or nonpurgers). The c) and d) groups were combined to form the

bulimic syndrome group (called the bulimic syndrome). In the

present study, the bulimic syndrome subjects' scores were first

analyzed together, then separately as the purging bulimic and

nonpurging bulimic subgroups.








The groups were compared on the basis of their answers to three

instruments: the Eating Disorder Inventory (EDI) which has 8

subscales (Garnet et al., 1983b), the Conte Borderline Syndrome

Index (CBSI) (Conte et al., 1980); and the Self-Other Differentia-

tion Instrument (SODI) which was developed by the present author for

use in this study and whose results were considered exploratory.

The EDI was used in the present study for several reasons. First,

several scales of the EDI have been shown to discriminate between

those who exhibit symptoms of disordered eating (extreme dietary

restraint, and dieting and weight preoccupation) and those with the

eating disorder syndrome of anorexia nervosa (Garner et al., 1983a).

Second, a considerable amount of evidence has been gathered by its

authors to demonstrate the reliability and validity of the EDI

(Garner et al., 1983 b). Third, it is one of the the best-

constructed measures available of a range of psychological and

behavioral traits common to eating disordered individuals. Fourth,

the EDI is quickly and easily self-administered, which made it an

ideal choice for inclusion in a survey by mail.

Table 2-1 lists the EDI subscales with descriptions of the

psychological and behavioral characteristics they measure. Garner

et al. (1983a) found that both dieters and anorexic patients had

elevated scores on the Drive for Thinness, Body Dissatisfaction, and

Perfectionism scales. Anorexics, however, had elevations on the

other five scales. In the same study, Garner et al. studied a group

of extreme dieters whose scores on seven of the scales were not

significantly lower than those of the anorexic patients. The only

scale on which this group of dieters scored significantly lower than









the anorexics was the Ineffectiveness Scale. Garner et al. conclud-

ed that while certain symptoms occur along a continuum which is

shared by eating disordered individuals and other weight-preoccupied

and diet-preoccupied women, those with eating disorders have

concomitant psychopathology which the other women do not. The

scales identified by Garner et al. (1983a) and the authors of the

EDI (Garner et al., 1983b) which measure the fundamental aspects of

eating disorder pathology are the Ineffectiveness, Interpersonal

Distrust, Interoceptive Awareness, and Maturity Fears scales, with

particular emphasis placed on the importance of the Ineffectiveness

Scale.

Six of the EDI scales were tested for convergent and discrimi-

nant validity for a sample of anorexic patients (Garner et al.,

1983b). The Drive for Thinness Scale correlated highly with mea-

sures of dietary restraint and anorexic-like attitudes. The Bulimia

scale was related to measures of lack of self-control, restraint,

and body dissatisfaction. The Body Dissatisfaction Scale was most

related to dissatisfaction with body parts which mature at puberty;

(i.e. breasts, hips, thighs). The Ineffectiveness Scale was related

to measures of depression, low self-esteem, and external locus of

control. Perfectionism was most related to interpersonal

sensitivity on the Hopkins Symptom Checklist and Interpersonal

Distrust was related to low self-esteem and depression. Comparable

data were unavailable for the Interoceptive Awareness and Maturity

Fears scales.

The Conte Borderline Syndrcme Index (Conte et al., 1980) was

included in the present study as a measure of the characteristics of









Table 2-2

EDI SCALES

Drive for Thinness: excessive concern with dieting and preoccupation
with weight; entrenchment in extreme pursuit of thinness;
ardent wish to lose and fear of gaining weight

Bulimia: tendency toward uncontrollable overeating and may be
followed by impulse to engage in self-induced vomiting

Body Dissatisfaction: belief that specific parts of body associated
with shape change at puberty are too large

Ineffectiveness: feelings of general inadequacy, insecurity,
worthlessness, not having control over one's life; a highly
negative self-concept

Perfectionism: excessive personal expectations for superior
achievement

Interpersonal Distrust: sense of alienation and general reluctance
to form close relationships, unable to form attachments and
feel comfortable expressing emotions toward others

Interoceptive Awareness: lack of confidence in recognizing and
accurately identifying emotions and sensations of
hunger/satiety

Maturity Fears: wish to retreat to security of pre-adolescent years
because of overwhelming demands of adulthood








borderline personality disturbance. The CBSI was found to discrimi-

nate borderline personality patients from normals, depressed outpa-

tients and inpatient schizophrenics. Borderline patients endorsed

more items than the other groups related to impaired object rela-

tions, impulsivity, emptiness and depression, depersonalization, and

lack of self-identity. When the CBSI has been used with eating

disordered patients, bulimics have scored well into the borderline

personality disorder range, beyond anorexic, normal, and depressed

women (Lepkawsky, 1983). In the present study, the CBSI will be

used to determine whether a group of DSM-III bulimics from a

nonclinical sample have more features of the borderline disorder

than either normals or binge eaters, and within the bulimic group,

whether the purging bulimic subgroup has more borderline features

than does the nonpurging bulimic subgroup. Using both the CBSI and

the EDI, the hypotheses in the following section were tested.

Hypotheses
1. The syndrome group will have significantly higher scores than

the control group on all EDI scales and on the CBSI.

The EDI was designed to measure psychological and behavioral

characteristics of eating disorder pathology. It is expected that

bulimic syndrome individuals within a nonclinical population should,

by definition, show significantly greater disturbance on the EDI.

Such differences will also lend credence to the method used in the

present study to identify and group bulimic women.

Bulimic syndrome subjects are also expected to show signifi-

cantly more borderline features, as measured by the CBSI, than

control subjects. Problems in interpersonal relationships, in









affective stability, and in impulse control, have all been found to

a significantly greater degree among bulimic women than among

nonbulimic women (Johnson et al., 1984 (a); Lepkowsky, 1983; Pyle et

al., 1983).

2. The syndrome group will have significantly higher scores than

the symptom group on three EDI scales: Interpersonal Distrust,

Ineffectiveness, and Maturity Fears.

Results from previous studies have suggested that the IA, ID

and MF scales of the EDI are highly correlated with disturbances in

self-esteem and psychological maturity (Garner et al., 1983 (a) and

b). The present hypothesis is based on previously cited evidence

which indicates that bulimic syndrome subjects suffer greater

depression and self-esteem problems than do binge eaters (Katzman

and Wolchik, 1984).

3. The symptom group will have significantly higher scores than

the control group on three EDI scales: Body Dissatisfaction,

Drive for Thinness, and Bulimia.

The present hypothesis is based on research evidence which

suggests that scores on the BD, DT, and B scales are higher in

groups of weight preoccupied women than in control groups (Garner et

al., 1983a). In addition other studies have shown that as a group,

binge eaters tend more than nonbingers to have a more negative body

image, to eat uncontrollably, and to exhibit greater dietary re-

straint (Hawkins and Clement, 1980; Ondercin, 1979; Sinoway, 1983;

Wolf and Crowther, 1983).

4. Within the syndrome group, the purgers will have significantly

higher scores than the nonpurgers on the Interpersonal Distrust

and Ineffectiveness scales.









This hypothesis is based on studies by Abraham and Beumont

(1982), Pyle et al. (1983), and Sinoway (1983), all of which suggest

that subgroups of purging and nonpurging bulimics may exist within

the syndrome of bulimia, and that purgers are more dysfunctional.

It is proposed that the purging bulimics in the present nonclinical

sample will show more dysfunction on the two scales most related to

self-esteem and interpersonal relationships than will the nonpurging

bulimics.

5. The syndrome group will have significantly higher scores on the

CBSI than the symptom group.

Bulimic subjects in nonclinical samples have been shown to have

greater problems with self-esteem, depression, and impulsivity than-

binge eaters (Katzman and Wolchik, 1984; Nagelberg et al., 1984).

It is hypothesized that the bulimics' greater difficulties will be

reflected in higher CBSI scores.

6. The purging bulimic group purgerss) will have significantly

higher scores on the CBSI than will the nonpurging bulimic

(nonpurgers) group.

It is hypothesized that purgers will have significantly more

problems characteristic of borderline personalities than will the

nonpurgers. This hypothesis is based on evidence that purgers show

less impulse control and social compliance than nonpurgers

(Nagelberg et al., 1984). In addition other authors have suggested

that purgers are more severe bulimics than nonpurgers (Halmi et al.,

1981; Pyle et al., 1983; Sinoway, 1983). It is the present author's

experience that purgers in clinical settings have significantly

greater problems than do nonpurgers in impulse control, acquiring





50


and maintaining interpersonal relationships, and regulating their

affect. It is postulated that this difference will also be present

in a nonclinical population.













CHAPTER III
METHOD

Subjects

The subjects in this study were undergraduate woaen at the

University of California, Santa Barbara, who were surveyed by mail

according to the procedure described in this chapter. Participation

was voluntary and subjects' identities were strictly confidential.

No direct compensation was offered but subjects were given the

opportunity to receive a summary of the study's results. In addi-

tion they were given access to information and/or counseling for any

problems they had with eating disorders (see Appendix A for a copy

of the introductory letter).

Materials
Copies of the materials developed by the present author for use

in this study may be found in the appendices in the following order:

Weight Management, Eating and Exercise Habits Questionnaire (WM)

and its introductory letter, Appendix A; Self-Other Differentiation

Instrument (SODI), Appendix B; and Family History Questionnaire

(FHQ), Appendix C.

The WMQ

The WMQ was developed following a general format used by a

number of authors (Halmi et al., 1981; Johnson et al., 1982; Katzman

and Wolchik, 1984; Stuckey, Lewis, Jacob, Johnson and Schwartz,

1981). The format includes an operationalized version of the

DSM-III criteria for bulimia which was used to separate subjects








into the experimental groups. The questionnaire was divided into

four sections, three of which were used to identify the four groups.

Section A asked questions about the subjects' current and past

weight. Section B contained questions related to exercise, dieting,

and other methods subjects had used in the past two months to reduce

and/or control their weight, and the frequency with which those

methods were used. Exercise methods were activities in which

college students frequently engage for weight loss. Dietary methods

represented habits which are common to both eating disordered and

"average" college women. Many of the habits were included on the

recommendation of the University dietician who views their frequent

and/or combined use as an early sign of an eating disorder. Section

C contained information designed to help direct the dissemination of

information in eating and exercise to university women. Section D

included questions specific to the type and frequency of behaviors

and attitudes associated with students' binge eating.

The Eating Disorder Inventory (EDI)

The EDI is a 64 item, self-report inventory designed to assess

a broad range of psychological and behavioral traits common in

anorexia nervosa and bulimia. A considerable amount of evidence has

been collected to demonstrate the validity of the EDI. Significant

correlations were found between the EDI subscales scores of

anorexics and their scores on the Eating Attitudes Test; on Herman

and Polivy's measure of dietary restraint; on body dissatisfaction;

on depression as measured by the BDI; on samatization,

obsessionality, anxiety, depression, and interpersonal sensitivity

as measured by the Hopkins Symptcm Checklist; and on external locus









of control and lack of self-control (Garner et al., 1983b). In

addition reliability coefficients for the eight subscales range from

.82 to .90 for anorexics and from .65 to .91 for normals (Garner et

al., 1983b).

Subjects responded to each EDI item by rating whether the item

applied "always," "usually," "often," "sometimes," "rarely," or

"never." The most extreme responses ("always" or "never," depending

on the keyed direction) received a score of 3; the responses adja-

cent to the extreme received a 2; the next response a 1 and the

three responses opposite the most extreme were scored zero. Scores

for the subscales were the sums of items included in each subscale.

Conte Borderline Syndrome Index

The CBSI, developed by Conte, Plutchnik, Karasu, and Jerrett

(1980), was used to detect the presence of characteristics of the

borderline personality syndrome as defined by the DSM-III and other

investigators (Masterson, 1977). The 52 items of this self-report

questionnaire reflect characteristics such as poor impulse control,

absence of a consistent self-identity, depression, anhedonia,

impaired object relations, depersonalization, and a number of other

symptoms. Subjects indicated by marking "yes" or "no" on items,

whether the statement described the way they usually feel. A

subject's score was a sum of the "yes" responses, each of which

counts as one point. Raw scores may be converted to percentile

scores for comparison with standards for normal and borderline

personalities. The CBSI has a reliability coefficient of .92 and

has been found to discriminate effectively between patients diag-

nosed as having a borderline personality disorder and groups of









normals, depressed outpatients and inpatient schizophrenics (Conte

et al., 1980).

Self-Other Differentiation Instrument

Subjects completed a Self-Other Differentiation Instrument

(SODI) which was adapted from the Repertory Grid Test developed from

George Kelly's personal construct theory. According to construct

theory, a clear sense of self-identity is possible only if an

individual can elaborate a pattern of differences and similarities

between him or herself and others. The Repertory Grid operation-

alizes the concept of self-identity by requiring an individual to

numerically rate a series of people including him or herself on a

list of bipolar dimensions of meaning, or constructs. A differenti-

ation score is obtained by summing the absolute values of the

differences between the self-ratings and the ratings of each of the

other persons on each of the constructs (Adams-Webber, 1984). An

average differentiation score is ccmputed for each person rated

other than the self.

Research and clinical observation have suggested that persons

with eating disorders have difficulty with the developmental task of

separating and individuating from their families (Boskind-White and

White, 1981; Bruch, 1973; Minuchin, Rosman, and Baker, 1978).

Nagelberg et al. (1984) found that compared to normal women,

binge/purgers felt they had less worth as family members. Johnson

and Flach (1985) found bulimics characterized their families as

having a high emphasis on achievement and a low emphasis on indepen-

dence and self-reliance. Their families lacked emotional support

and were filled with chronic anger and other suppressed emotions.









The nature of these families as seen by their daughters presented a

classic double-bind, where the daughters were expected to achieve

but were not encouraged to attain independence, self-reliance, or

assertive self-expression. These conflicting demands, according to

Johnson and Flach, lead to feelings of ineffectiveness and low

self-esteem.

The SODI was developed by the present author to explore the

ways in which bulimic and nonbulimic wcmen differentiate themselves

from significant others. The SODI consists of four sets of ten

bipolar constructs. The first set represents "Yourself;" the second

set, "Parent to Whom You Feel Closest;" the third, "Sister to Whom

You Feel Closest;" and the last, "Brother to Whom You Feel Closest."

The ten constructs were chosen from a pool of constructs developed

by several clinicians specializing in eating disorders. The con-

structs chosen were those which the clinicians decided were most

salient to the self-concept of bulimics. Each of the ten constructs

is measured on a 13-point scale, with 0 as the mid-point and each

endpoint representing either the negative or positive extreme of the

scale. Three scores were obtained for each subject by (a)

converting all scale scores to scores from 1-13, with 1 as the

negative extreme and 13 as the positive extreme; (b) subtracting

each construct score in the Parent, Sister, and Brother sets from

their counterparts in the Self set; (c) summing the 10 difference

scores for the three sets of comparisons (Self-Parent, Self-Sister,

Self-Brother). Three average difference scores were computed for

each of the symptom, control, purging bulimic, and nonpurging

bulimic groups.









Procedure

Sample Selection

These data were collected as part of a joint research project

carried out by the Chancellor's Office, the Student Health Service,

and Counseling and Career Services of the University of California

at Santa Barbara. A randan sample of 1,487 wcmen was selected form

the population of undergraduate women at the University of Califor-

nia, Santa Barbara. They were mailed packets containing the ~M and

the letter of introduction. Packets were mailed again to those who

had not responded after two weeks. The final sample consisted of

813 women, or 54.56% of those surveyed (see Table 2-2 for a summary

of demographic characteristics of the respondents). All data were

collected over the Winter and Spring Quarters of 1985.

Group Selection

Four groups-control, symptom, purging bulimic, and nonpurging

bulimic-were chosen on the basis of their answers to questions on

the M. The following is a description of the criteria for inclu-

sion in each group.

Control (Normal Eater) Group

Subjects were included if they met all of the following:

1. Subject was within normal body weight.

Normal body weight = > lower limit and < upper limit of weight

range.

Upper limit =

[(Height in feet) (20)+(ht. in inches) (5)]+[20%[ht. in

ft) (20)+(Ht. in inches) (5) ]









Iower limit =

[(Height in feet) (20)+(ht. in inches) (5) ]-[20%[ht. in

ft) (20)+(Ht. in inches) (5) ]

2. Subject indicated no past or current problem with anorexia

nervosa, unhealthy underweight, overweight, or obesity.

3. Subject had not attempted to lose and/or control her weight in

the last month by engaging in any one of the following activi-

ties more than once daily: jogging, swimming, walking, aerobic

dance, or other exercise.

4. Subject had not attempted to lose and/or control her weight in

the last month by daily engaging in more than one of the above

activities.

5. In the last month subject used any one of the following to

reduce and/or control her weight less than once a week: eating

on a special diet, going 24 hours without food (fasting), or

skipping meals.

6. In the last month subject did not use any of the following to

reduce and/or control her weight: laxatives, diuretics,

appetite control pills, vomiting, chewing and spitting out

food, or other methods of purging.

7. Subject indicated no problem with binge eating.

Purging and Nonpurging Bulimic Groups (Combined, the purging and

nonpurging groups form the Bulimic Syndrome Group)

Subjects were included if they met all of the following:

1. Subject indicated she has a problem with binge eating.









2. Subject met three of the following five criteria:

a. Prefers always, frequently, or occasionally to binge in

private;

b. Prefers always, frequently, or occasionally to binge on

sweet, easily digested food;

c. Stops binge eating always, frequently, or occasionally by

sleeping, by being interrupted by others, when her stomach

hurts, or when she vomits;

d. In the last month she controlled and/or reduced her weight

by

-Using laxatives, or vomiting, or chewing and spit-

ting out food at least once a month; or

-Eating on a special diet, or fasting, or skipping

meals at least once a week.

3. Subject gained and lost ten or more pounds in the last month

due to binging and dieting or fasting.

4. Subject was within normal body weight range.

5. Subject answered always, frequently, or occasionally to either

of the following:

"I feel I just can't stop eating when I binge;" or

"I feel badly about myself when I binge."

6. Subject indicated no current or past problem with anorexia

nervosa.

7. Purging bulimics

The purging bulimics purgerss) used any of the following at

least once in the last month to reduce and/or control their

weight: laxatives, vomiting, or chewing and spitting out food.









8. Nonpurging bulimics

The nonpurging bulimics (nonpurgers)

-did not use laxatives, vomiting, or chewing and

spitting out food in the last month, and

-had binged at least three times in the last month.

Symptom (Binge Eater) Group

Subjects were included if they met all of the following:

1. Subject indicated she had a problem with binging.

2. Subject did not meet all the criteria for the bulimic group.

3. Subject had not used laxatives or vomiting to reduce and/or

control her weight in the last month.

4. Subject was within normal body weight range.

5. Subject had indicated no past or current problem with anorexia

nervosa.

Of the 813 respondents, 151 were included in the control group,

51 in the symptom group, 73 in the nonpurging bulimic group, and 46

in the purging bulimic group. A second packet consisting of the

EDI, the CBSI, the SODI, and the FHQ, was mailed to the 321 sub-

jects. A response rate of 74.14% (239) subjects was obtained from

this mailing.

Data Analysis

The data were analyzed using MANOVA and ANOVA procedures to

test for overall group effects for the six hypotheses. Hypotheses

1, 2, 3 and 5 were tested using the MANOVA procedure to compare the

control, symptom, and bulimic syndrome groups on the 8 EDI scales.

Hypothesis 1 and 5 were also tested using a one-way ANOVA to compare

the three groups on the CBSI. Hypothesis 6 was tested with a









one-way ANOVA comparing four groups (the purging bulimics,

nonpurging bulimic, symptom, and control group) on the CBSI.

Hypothesis 4 was tested using the MANOVA procedure to compare the

four groups on the EDI scales.

Since eight ANOVAS were performed on the EDI subscales, to

control for any inflation of the alpha level and to insure a conser-

vative interpretation of the results, Dunn's procedure was employed

(Kirk, 1982). This procedure divides the usually accepted alpha

level of .05 by the number of tests performed. Thus, an alpha of

.0063 was set as the required level of significance for the EDI

ANOVAS.

The SODI was analyzed using the MANOVA procedure to compare the

four groups on the three SODI differentiation scores, the Self-

Parent, Self-Sister, and Self-Brother. Using Dunn's procedure, an

alpha level of .017 was employed. Tukey's t-test procedure was used

to determine significant differences between group means. These

analyses were considered exploratory and will be reported as trends

with implications for future research.














CHAPTER IV
RESULTS

This chapter will be divided into three sections, with the

first describing the eating behavior and weight-related

characteristics of the sample; the second describing the results of

the tests of the hypothesis; and the third describing the results of

the SODI. Numbers appearing in the tables have been rounded off to

the second decimal point.

Eating Behavior and Weight-Related Characteristics

The average height, actual weight, and ideal weight of the 813

women in the sample are shown in Table 4-1. Also shown are the

frequencies with which women reported currently having a problem

with anorexia nervosa, deliberately staying unhealthily underweight,

being unintentionally underweight, being overweight, or being obese.

While 90.7% of the sample were within normal body weight range for

their height, 25.1% of the sample claimed to have a current problem

with being either overweight or obese. It is interesting that the

mean ideal weight of the sample is 8.74 pounds less than the mean

actual weight. Body size distortion and body weight dissatisfaction

are common among college women and this sample, it appears, is no

exception.

The frequency with which women in this sample met each of the

criteria for the bulimic syndrome group is presented in table 4-2.

Of the 813 respondents, 265 or 32.6% claimed to have a problem with









binge eating. Three or more of the five criteria for bulimia were

met by 26.57% of the total sample. Two hundred and three subjects

or 24.79% reported they "always, frequently, or occasionally" felt

they couldn't stop eating when they binged, and 248 or 30.50%

reported they felt badly about themselves after a binge "always,

frequently, or occasionally." Only 18 subjects (2.2%) reported

having anorexia nervosa in the past, and 6 (.73%) reported currently

having it. Purging methods were used at least once a month by

11.07% of the sample (90 subjects). The criterion of > 3 binges in

the last month was met by 169 subjects, or 20.79% of the sample.

The purging bulimic group represented 5.66% of the sample; the

nonpurging bulimics, 8.98%; and the symptom, 6.27% of the sample.

The purging and nonpurging bulimic groups together formed the

bulimic syndrome group, which comprised 14.64% of the total sample.

Major Findings: The EDI and the CBSI

The EDI results were first analyzed using the MANOVA procedure

to compare the symptom, control and syndrome groups. They were

again analyzed using the MANOVA to compare the symptom, control,

purging bulimic, and nonpurging bulimic groups. The CBSI was

analyzed in two separate ANOVAS, one comparing the three groups and

the other comparing the four groups. The order of presentation of

these results is as follows:

1. MANOVA results of the three groups on the EDI

2. Tukey's t-test for significant differences between group means

3. Summary of support/nonsupport for hypotheses 1, 2, 3

4. ANOVA results comparing the three groups on the CBSI

5. Tukey's t-tests for significant differences between means









Table 4-1


WEIGHT-REIATED CHARACTERISTICS (OF 813 RESPONDENTS)

Mean Ht (ft/in) Mean Wt. (ibs) Mean Ideal Wt (ibs)


5' 6.93" 127.78 119.04





Yes-past Yes-now
Ever had a problem with: percent percent


Anorexia Nervosa 2.2% 0.73%
Unhealthy Underweight 9.10% 1.23%
Underweight 7.62% 1.85%
Unhealthy Overweight 41.57% 23.99%
Obesity 1.85% 1.11%









Table 4-2

FREQUENCY OF TOTAL SAMPLE MEETING THE CRITERIA FOR
BULIMIC SYNDROME


Criterion N Percent
(of 813)


A. Yes to "Do you have a problem with 265 32.6%
binge eating?

B. Yes to three of more of the following: 216 26.57%
1. Always/frequently/occasionally binges 196 24.11%
in private
2. A/f/o prefers to binge on sweet, 237 29.15%
easily digested food
3. A/f/o stops singing by sleeping, 229 28.17%
being interrupted, stomach pain,
or vomiting
4. In last month used laxatives, or 492 60.52%
vomiting, or chewing and spitting
out food at least once OR
Ate according to a special diet,
or fasted, or skipped meals at least
once a week
5. Has gained and lost at least 10 lbs 72 8.87%
in last month due to binging and
dieting or fasting

C. Within normal body weight range 738 90.78%

D. A/f/o to either "I feel I can't stop 203 24.97%
eating when I binge" OR
"I feel badly about myself after I 248 30.50%
binge"

E. Did not answer yes to "Did you ever 795 97.79%
have a problem with anorexia nervosa?"

F. Did not answer yes to "Do you now have 807 99.26%
a problem with anorexia nervosa?"

G. Purging bulimics: Yes if they met all 46 5.66%
bulimic criteria and in the last month
took laxatives, vanited, or chewed and
spit out food at least once.

H. Nonpurging bulimics: Yes if they met 73 8.98%
bulimic criteria, did not meet purging
criteria, and had > 3 binges in the last
month.









6. Summary of support/nonsupport for hypothesis 1, 5

7. MANOVA results comparing the four groups on the EDI

8. Tukey's T-tests for significant differences between means

9. Summary of support/nonsupport for hypothesis 4

10. ANOVAS results comparing the four groups on the CBSI

11. Tukey's T-tests for significant differences between means

12. Summary of support/nonsupport for hypothesis 6

MANOVA Results for Three Groups on the EDI

The MANOVA comparing the control, symptom, and bulimic syndrome

groups on the EDI was significant (F(16,456)=9.58; p< 0.0001). The

ANOVAS comparing the three groups on each of the scales are present-

ed in Tables 4-3 through 4-10. Significant results were found for

Drive for Thinness ((F(2,235)=70.55; p< 0.0001); Bulimia (F(2,235)=

40.02, p< 0.0001); Body Dissatisfaction F(2,235)=33.29, p< 0.0001);

Ineffectiveness (F(2,235)=11.06, p< 0.0001); and Interoceptive

Awareness (F(2,235)=17.44, p< 0.0001). Trends toward significance

were found for Interpersonal Distrust (F(2,235)=4.87, p< 0.0195) and

Maturity Fears (F(2,235)=4.87, p< 0.0084).

T-Tests for Significant Differences Between Means

The Tukey's t-tests for significant differences between means

of the three groups on 5 EDI variables are presented in Tables 4-11

through 4-15. All t-tests were performed using an alpha level of

.05.

An examination of the group means for Drive for Thinness

indicates that the bulimic syndrome group had significantly higher

scores than the symptom or control groups. The mean of the

symptom group was significantly higher than that of the control









group. A similar pattern (bulimic syndrome > symptom > control) of

means was found for the Body Dissatisfaction scale.

For the Bulimia and Interoceptive Awareness scales, the bulimic

syndrome mean was greater than either the symptom or control group.

No difference was found between the symptom and control group means

for these two scales. A significant difference between the bulimic

and control groups was also found on the Ineffectiveness scale. No

differences were found on this scale either between the bulimic and

symptom groups or between the symptom and control groups.

Summary of Support/Nonsupport for Hypotheses 1,2,3

Partial support was found for the hypothesis that the bulimic

syndrome group would score significantly higher on all EDI scales

than the control group. The bulimic group mean was significantly

higher than the control mean on the Drive for Thinness, Bulimia,

Body Dissatisfaction, Ineffectiveness, and Interoceptive Awareness

scales. Trends in the hypothesized direction were seen on the

Interpersonal Distrust and Maturity Fears scales. There was no

trend toward significance on the Perfectionism scale. Thus as

predicted, bulimic students showed significantly greater concern for

dieting and weight loss, a greater tendency to binge and to confuse

emotional arousal with hunger/society cues, a greater dislike for

their bodies, and a lower sense of personal adequacy than did

students with no eating or weight problems.

A trend that approached significance was found in the direction

of the second hypothesis, that the syndrome group would have signif-

icantly higher scores than the symptom group on the Interpersonal










Table 4-3

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR DRIVE FOR THINNESS


Source SS Df MS F P R-Square


Model 2706.16 2 1353.08 70.55 <0.0001 0.375

Error 4507.12 235 19.18

Corrected 7213.28 237
Total









Table 4-4

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR BULIMIA


Source SS Df MS F P R-Square


Model 807.11 2 403.56 40.02 <0.0001 0.254

Error 2369.89 235 10.09

Corrected 3176.10 237
Total









Table 4-5

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR BODY DISSATISFACTION


Source SS Df MS F P R-Square


Model 3538.26 2 1769.13 33.29 <0.0001 0.221

Error 12488.53 235 53.14

Corrected 16026.79 237
Total









Table 4-6

ANALYSIS OF VARIANCE SUMMARY TABIE
FOR INEFFECTIVENESS


Source SS Df MS F P R-Square


Model 264.50 2 135.25 11.06 <0.0001 0.086

Error 2810.46 235 11.96

Corrected 3074.96 237
Total










Table 4-7

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR PERFECTIONISM


Source SS Df MS F. P R-Square


Model 80.53 2 40.27 2.26 <0.107 0.019

Error 4196.11 235 17.86

Corrected 4276.64 237
Total









Table 4-8

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR INTERPERSONAL DISTRUST


Source SS Df MS F P R-Square


Model 67.29 2 33.64 4.01 <0.019 0.03

Error 1970.65 235 8.39

Corrected 2037.93 237
Total.










Table 4-9

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR INTEROCEP1VE AWARENESS


Source SS Df MS F P R-Square


Model 545.85 2 272.93 17.44 <0.0001 0.13

Error 3677.14 235 15.65

Corrected 4222.99 237
Total









Table 4-10

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR MATURITY FEARS


Source SS Df MS F P R-Square


Model 68.97 2 34.48 4.87 <0.008 0.04

Error 1663.32 235 7.08

Corrected 1732.29 237
Total









Table 4-11

MEANS, STANDARD DEVIATIONS FOR
DRIVE FOR THINNESS


Bulimic Symptom Control
(N=89) (N=42) (N=107)


X SD X SD X SD

8.98 5.59 4.48 4.85 1.52 2.73
(a,b) (a,c) (b,c)

Note: Means sharing a common subscript differ at the .05 level









Table 4-12

MEANS, STANDARD DEVIATIONS FOR
BULMIA


Bulimic Symptom Control
(N=89) (N=42) (N=107)



X SD X SD X SD

4.33 4.88 1.38 2.29 0.29 0.74
(a,b) (a) (b)


Note: Means sharing a cmman subscript differ at the .05 level









Table 4-13

MEANS, STANDARD DEVIATIONS FOR
BODY DISSATISFACTION


Bulimic Symptom Control
(N=89) (N=42) (N=107)


X SD X SD X SD

14.66 7.53 9.81 7.82 6.13 6.86
(a,b) (a,c) (b,c)

Note: Means sharing a common subscript differ at the .05 level









Table 4-14

MEANS, STANDARD DEVIATIONS FOR
INEFFECTIVENESS


Bulimic Synptom Control
(N=89) (N=42) (N=107)



X SD X SD X SD

3.26 4.15 2.0 4.28 0.93 2.27
(a) (a)


Note: Means sharing a common subscript differ at the .05 level









Table 4-15

MEANS, STANDARD DEVIATIONS FOR
INTERICEPTIVE AWARENESS


Group

Bulimic Symptom Control
(N=89) (N=42) (N=107)



X SD X SD X SD

4.58 5.42 2.33 3.24 1.25 2.5
(a,b) (a) (b)


Note: Means sharing a common subscript differ at the .05 level









Distrust, Ineffectiveness, and Maturity Fears scales. These scales

were chosen to represent what was hypothesized to the the greatest

difference between bulimics and binge eaters: self-esteem. While

the means of these scales were in the predicted direction, the only

significant differences found between the syndrome and symptom

groups were on scales not included in the second hypothesis: Drive

for Thinness, Bulimia, Body Dissatisfactions, and Interoceptive

Awareness. This suggests that the bulimic and binge eating groups

in this study differed more on variables related to disordered

eating, body image, and weight preoccupation than on variables

related to self-esteem.

The third hypothesis, that the symptom group would score

significantly higher on the Body Dissatisfaction, Drive for Thin-

ness, and Bulimic scales was supported for Body Dissatisfaction and

Drive of Thinness. Indeed, of the 8 MANOVAS for the EDI scales, the

symptom and control only differed significantly only on two. These

findings compared with the results of the second hypothesis suggest

that the symptom group differs more from the bulimic group than it

does from the control group on variables related to disordered

eating.

ANOVA and Tukeys' T-Tests for the Three Groups on the CBSI

The results of the ANOVA comparing symtpom, control, and

syndrome groups (found in Table 4-16) was significant (F(2,236)=

12.62, p< 0.0001). An examination of the t-tests indicates that the

bulimic syndrome group endorsed significantly more items indicative

of borderline personality characteristics than either the symptom or

control groups (Table 4-17). The bulimics' mean score was in the









Table 4-16

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR THE CBSI


Source SS Df MS F P R-Square


Model 1194.76 2 597.38 12.62 <0.0001 0.097

Error 1117.16 236 47.34

Corrected 12365.93 238
Total









Table 4-17

MEANS, STANDARD DEVIATIONS FOR
CBSI


Group

Bulimic Symptom Control
(N=89) (N=42) (N=107)



X SD X SD X SD

9.10 8.30 5.33 6.12 4.24 5.78
(a,b) (a) (b)


Note: Means sharing a common subscript differ at the .05 level









81st percentile of the range of scores for normals, and in the 3rd

percentile of scores for borderlines (see Appendix D). The con-

trols' mean score was in the 54th percentile for normals and in the

1st percentile for borderlines. The symptom group's mean was in the

60th percentile for normals and in the 1st percentile for border-

lines. No significant difference was found between the symptom and

control group means on the CBSI.

Summary of Support/Nonsupport for Hypothesis 1,5

Both hypotheses 1 and 5 were supported by the finding that

bulimics scored significantly higher than either the controls or

binge eaters on a measure of borderline characteristics. These

characteristics include (among others) poor impulse control, absence

of a consistent self-identity, depression, and impaired interper-

sonal relations. It is interesting to note that on the ED1 the

bulimic group did not significantly differ from either the control

or symptom group on the measure of feeling of inadequacy in inter-

personal relations (Interpersonal Distrust) or on the measure of

feeling overwhelmed by the demands of adulthood. The CBSI scores

indicate however, that bulimics tend to have greater intrapersonal

and interpersonal difficulties than either the symptom or control

groups. While the bulimic group scored higher on the CBSI than 81%

of the normals they were only higher than 3% of diagnosed border-

lines. These results combined with the ED1 data suggest that the

intrapersonal and interpersonal disturbances seen in the bulimic

sample do not indicate significant personality pathology.

MANOVA Results for Four Groups on the EDI

The MANOVA comparing the control, symptom, purging bulimic, and

non-purging bulimic groups or the EDI was significant








(F(24,658)=7.31,< 0.0001). The results of the MANOVAs comparing the

control, symptom, purging bulimic, and nonpurging bulimic groups on

the 8 EDI scales are presented in Tables 4 18 through 4 25.

Significant results were found for Drive for Thinness (F(3,234)=

54.64, p<0.0001); Bulimia (F(3,234)=31.45, p<_0.0001); Body Dissatis-

faction (F(3,234)=25.51, p<0.0001); Ineffectiveness (F(3,234)=8.05,

p<0.0001); and Interoceptive Awareness (F(3,234=15.81, p<0.0001).

T-tests for Significant Differences

The Tukey's t-tests for significant differences between the

group means on the 5 significant EDI variables are found in Tables

4-26 through 4-30. All t-tests were performed at an alpha level of

.05.

An examination of the group means for the Drive for Thinness

scale indicates that a) the purging bulimics had significantly

higher scores than the nonpirging, symptom, and control groups, b)

the nonpurging bulimics had significantly higher scores than the

symptom and control groups; and c) the symptom group had signifi-

cantly higher scores than the control group. This was the only

scale on which all groups differed significantly from one another.

The mean Bulimia scores of the four groups differed signifi-

cantly from one another except the symptom and control means. No

significant difference was found between these two means. For the

Body Dissatisfaction Scale, the only nonsignificant difference among

the means of the four groups was between the nonpurgers and the

symptom group. All other Body Dissatisfaction means differed

significantly from one another. The pattern of significant differ-

ences among the group means on the Bulimia and Body Dissatisfaction









Table 4-18

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR ERIVE FOR THINNESS


Source SS Df MS F P R-Square


Model 2971.33 3 990.44 54.64 <0.0001 0.412

Error 4241.95 234 18.13

Corrected 7213.28 237
Total










Table 4-19

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR BULIMIA


Source SS Df MS F P R-Square


Model 912.95 3 304.32 31.54 <0.0001 0.287

Error 2264.05 234 9.68

Corrected 3176.99 237
Total









Table 4-20

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR BODY DISSATISFACTION


Source SS Df MS F P R-Square


Model 3949.33 3 1316.45 25.51 <0.0001 0.246

Error 12077.46 234 51.61

Corrected 16026.79 237
Total









Table 4-21

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR INEFFECTIVENESS


Source SS Df MS F P R-Square


Model 287.55 3 95.85 8.05 <0.0001 0.094

Error 2787.42 234 11.91

Corrected 3074.96 237
Total









Table 4-22

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR PERFECTIONISM


Source SS Df MS F P R-Square


Model 98.29 3 32.76 1.83 <0.14 0.022

Error 4178.35 234 17.86

Corrected 4276.64 237
Total









Table 4-23

ANALYSIS OF VARIANCE SUMMARY TABLE
FOR INTERPERSONAL DISTRUST


Source SS Df MS F P R-Square


Model 67.79 3 22.59 2.68 <0.046 0.033

Error 1970.14 234 8.42

Corrected 2037.93 237
Total