dissertation submllted in partial fulfillment of the … · 2004-09-01 · exclusive licence...

PREVENTING DISORDERED EATING;

PROMOTING HEALTHY AlTlTUDES AND BEHAVIORS:

A SCHOOL-BASED PROGRAM

by

Rafael David Richman

B.Sc., University of Toronto, 1988

M.A., Simon Fraser University, 1993

DISSERTATION SUBMllTED IN PARTIAL FULFILLMENT OF

THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

in the Department

of

Psychology

O Rafael D. Richman, 1997

SIMON FRASER UNIVERSITY

June 1997

Ali rights reserved. This work may not be reproduced in whole or in part, by photocopy

or other means, without permission of the author.

National Library 1+1 OfCrnada Bibliothéque nationale du Canada

Acquislions and Acquisitions et Bibliographic Services services bibliographiques

395 Wellington Street 395. rue Wellington Ottawa ON K1A ON4 Omwa ON KtA ON4 Canada Canada

The author bas granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, han, distribute or seil reproduire, prêter, distribuer ou copies of this thesis in microfonn, vendre des copies de cette thèse sous paper or electronic formats. la fome de microfiche/fh, de

reproduction sur papier ou sur format électronique.

The author retains ownershp of the L'auteur conserve Ia propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantid extracts fiom it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.

Preventing Disordered Eating iii

ABSTRACT

Keywords: disordered eating; primary prevention; community psychology;

children; school programs; eating disorders.

A prirnary prevention and education program aimed at preventing disordered

eating and promoting healthy attitudes and behaviors was implemented in 13

grade five, six, and seven classes in eight urban primary schools. Ten classes

from three similar schools served as a no-treatment comparison group. The

program consisted of six weekly sessions focusing on disordered eating, social

pressures to be thin, harrnful effects and myths about dieting and weight, coping

with the influences of family and friends, healthy eating habits, and increasing

self-esteem. Questionnaires on knowledge (KQ), body esteem (BES), body

satisfaction (BSM), dieting, gaining, and maladaptive eating attitudes and

behaviors (ChEAT) were administered to both groups pior to, following, and six

months after the prevention program. The dependent measures were

categorized into three primary content areas: Knowledge (KQ), Affective (BSM &

BES), and Behavioral-Attitudinal (ChEAT, dieting, & gaining). Follow-up

questionnaires were given to seven (four treatment and three comparison) of the

classes. Overall results indicated the program was successful and effective in

inducing change. Statistically significant ANCOVA values were obtained for the

Cognitive, Affective, and Behavioral-Attitudinal domains at posttest. Further

analysis revealed statistically significant ANCOVA values on the KQ, BES,

Preventing Disordered Eating iv

ChEAT, ChEAT factor 1, and the dieting scale. The proportion of children

scoring above the ChEAT cutoff criterion decreased in the treatment group from

pretest to posttest, and remained stable and lower than the comparison group at

follow-up. Intervention participants continued to demonstrate superior

performance at follow-up on the Cognitive component, as shown by a statistically

significant ANCOVA value on the KQ measure. A tendency in the positive

direction at follow-up was observed for treatment children's scores on the ChEAT

and BES questionnaires. These values were not, however, statistically

significant. Separate analysis of high risk students suggests changes in the

positive direction for this sub-sample. At posttest, high risk treatment

participants displayed significantly more knowledge and higher body esteem

than comparison students. The present study provides strong support for the

short-term effectiveness of the intervention. Long-terni behavioral and attitudinal

changes possibly could be further enhanced with periodic booster sessions or

minor curriculum modifications. It is suggested that this and similar programs

may serve as practical and feasible tools to combat the increasing prevalence of

disordered eating and related disturbances in children.

Preventing Disordered Eating v

ACKNOWLEDGEMENTS

I would like to thank the members of my supervisory cornmittee: Ron Roesch for

al1 his editorial and moral support, especially when things got tough; Steve Hart

for helping me with the overall structure and conceptualkation for my analysis,

and the complex statistical analysis; and David Cox for his endless energy and

enthusiasm. Thanks to al1 three for their objective editorial comments, general

criticisms, and statistical insights. Thanks to Lori Tarcea for always being there

with the information I desperately needed to complete this project. Joan Foster,

thank you for al1 your statistical and cornputer-related advice. I would also like to

thank Joan Wolfe and Elizabeth Michno for shanhg their cornputer expertise with

me and making my tables, figures, and overall format more pleasing to the eye.

Thank you to Heli Huttenan and Allison Mischel, my dedicated assistants, who

enthusiastically volunteered their time and energy to helping administer and

score the questionnaires. To my parents, brother, friends, and fellow Simon

Fraser colleagu.es, thanks for all the moral and social support along the way.

Last, but definitely not least, I would like to especially acknowledge al1 the

teachers and students who participated in my project. Without their enthusiastic

cooperation, this program would not have been possible.

Preventing Disordered Eatinq VI

TABLE OF CONTENTS

Page

............................................................................. ............. Approval .. ii

.............................................................................................. Abstract iii

.................................................................................. Acknowledgments v

....................................................................................... List of Tables ix

................................................................ ................... List of Figures ... .xi

............................................................................... Introduction ........ .... 1

.......................................................................... Media Influences -3

................................................... Sociocultural Standards of Beauty -5

....................................................................................... Dieting -7

............................................... Prirnary Prevention: What and Why? 10

........................................... Substance Abuse Prevention Programs 14

.................................................... Prevalence of Disordered Eating 15

................................ Initial and Subsequent Epidemiological research 16

Prevalence of Disordered Eating and Maladaptive Attitudes

in Adolescents .......................................................................... 19

Prevalence of Disordered Eating in preadolescents ................ .. ........ 23

Eating Disorders and Disordered Eating Prevention Programs ............. 28

................................... A Model of Prevention and Disordered Eating 38

............................................................................. Present Study 41

Preventing Disordered Eating vii

Page

.............................................................................................. Method 43

. . ................................................................................ Participants 43

................................................................... Dependent Measures 46

............................................................................... Procedures -49

...................................................................................... Design 50

.............................................................................................. Results 54

........................................ Analysis of Posttest Scores (Uncorrected) 54

.............................................. Corrected Analysis (Posttest Scores) 59

..................................... Analysis of Follow-up Scores (Uncorrected) 60

........................................... Corrected Analysis (Follow-up Scores) 62

Effect Size Analysis ..................................................................... 62

Baseline-Group Analysis .............................................................. 64

............................................................. High Risk Group Analysis 64

......................................................................................... Discussion 66

............................ Limitations and Future Revisions ......... ... ..... .. -81

............................................................ Summary and Conclusions 89

Appendix A ............... .. ........................................................................ 91

Appendix B .... .............................................................. 99 .....................

Appendix C ....................................................................................... 101

.................................................................................... Appendix D 104

....................................................................................... Appendix E 107

Preventing Disordered Eating viii

Page

............................................................................. List of References.. 1 10

........................................................................................... Tables.. -1 17

............................................................................................ Figures .155

Preventing Disordered Eating ix

LIST OF TABLES

Table 1

Table 2

Table 3

Table 4

Table 5

Table 6

Table 7

Table 8

Table 9

Table 10

Table 11

Table 12

Table 13

Table 14

Table 15

Table 16

Table 17

Table 18

Table 19

Table 20

Page

...................................................... Ethnicity of Participants 117

................................. Corrected and Uncorrected ANCOVAs 118

Knowledge Questionnaire ANCOVA .................................... 119

....................................... Knowledge Questionnaire Scores -120

............................................. Body Esteem Scale ANCOVA 121

................................................ Body Esteern Scale Scores 122

Body Satisfaction Measure ANCOVA ....................... ... ....... -123

....................................... Body Satisfaction Measu re Scores 124

............................. Children's Eating Attitudes Test ANCOVA 125

Children's Eating Attitudes Test (ChEAT) Total Scores ............ 126

..................................... ChEAT Dieting Subscale ANCOVA 127

......................................... ChEAT Dieting Subscale Scores -128

.... ChEAT Bulimia and Food Preoccupation Subscale ANCOVA 129

ChEAT Bulimia and Food Preoccupation Subscale Scores ....... 130

ChEAT Oral Control Subscale ANCOVA ............................... 131

ChEAT Oral Control Subscale Scores. .................................. 132

Ch EAT Score 220 ............................................................. 133

.............................................................. Dieting ANCOVA 134

Self-Reported Dieting ........................................................ 135

............................................................ Gaining ANCOVA -136

Preventing Disordered Eating x

Table 21

Table 22

Table 23

Table 24

Table 25

Table 26

Table 27

Table 28

Table 29

Table 30

Table 31

Table 32

Table 33

Table 34

Table 35

Table 36

Table 37

Table 38

Page

............................................. Self-Reported Weight Gaining 137

................................................ Body Mass Index ANCOVA 138

Methods of Weight Reducing Used by Boys and Girls .............. 139

Follow-up corrected and uncorrected ANCOVAs ..................... 140

Follow-u p Knowledge Questionnaire ANCOVA ....................... 141

Follow-up Body Esteern Scale ANCOVA ............................... 142

Follow-up Body Satisfaction Measure ANCOVA ..................... 143

Follow-up ChEAT ANCOVA ................................................ 144

Follow-up ChEAT Dieting Su bscale ANCOVA ........................ 145

................................................ Follow-up Dieting ANCOVA 146

............................................... Follow-up Gaining ANCOVA 147

High Risk Knowledge Questionnaire ANCOVA ....................... 148

High Risk Body Esteem Scale ANCOVA ............................... 149

High Risk Body Satisfaction Measure ANCOVA ..................... 150

High Risk ChEAT ANCOVA ................................................ 151

................................................. High Risk Dieting ANCOVA 152

High Risk Gaining ANCOVA ............................................... 153

High Risk Body Mass Index ANCOVA .................................. 154

Preventing Disordered Eating xi

LIST OF FIGURES

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 1 5

Page

KQ pre-post scores .......................................................... 155

BES pre-post scores ......................................................... 155

........................................................ BSM pre-post scores 156

Ch EAT pre-post scores ..................................................... 156

C hEAT Factor 1 pre-post scores ......................................... 157

Pretest. Posttest. and Follow-up KQ .................................... 157

Pretest. Posttest. and Follow-up BES ................................... 158

Pretest. Posttest. and Follow-up BSM .................................. 158

Pretest. Posttest. and Follow-up ChEAT ............................... 159

Pretest. Posttest. and Follow-up ChEAT Factor 1 ................... 159

High Risk KQ .................................................................. 160

High Risk BES ......................... .... ............................... 160

High Risk BSM ................................................................ 161

............................................................ High Risk ChEAT 161

................................................................. High Risk BMI 162

Preventing Disordered Eating 1

INTRODUCTION

Disordered eating and maladaptive attitudes and behaviors associated

with food, diet, weight, and body perception are prevalent in today's western

society (Bruch, 1978; Garfinkel & Gamer, 1982; Hoek, 1993). These concerns

are pervasive among adults, adolescents, and children (Garfinkel & Gamer,

1982), and their ubiquity, especially in fernales, should not be underestimated.

Primary prevention programs are one means of directly addressing and

combating this widespread problem.

In addition to disordered eating, clinical and subclinical eating disorders

are prevalent and have increased in prevalence in recent years (Hoek, 1993;

Jones, Fox, Babigian, & Hutton, 1980). Whereas not too long ago the terms

anorexia nervosa and bulimia nervosa were unfamiliar tu the average layperson,

today they are a part of our vocabulary and, for many, influence our daily lives --

through family, friends, someone we know, or maybe ourselves.

Before discussing primary prevention, the prevalence of disordered eating

and eating disorders, and disordered eating prevention programs, anorexia

nervosa and bulimia nervosa will be defined, and relevant epidemiological data

will be briefly mentioned.

What is anorexia nervosa? The central feature of anorexia nervosa is the

relentless pursuit of thinness associated with the self-perception that one's body

size is too large or fat. Weight gain is dreaded, and dieting and excessive

exercising become means for avoiding fatness. The person with anorexia


nervosa needs to have control over her body; achievement is through the pursuit

of thinness (Garfinkel & Garner, 1982; Wenar, 1994). Further, central to the

diagnosis is a significantly below nomal body weight (DSM-IV, 1994). It is

estimated that anorexia nervosa occurs in 1 % of females (Rice, 1989; Wenar,

1 994).

We now have a working definition; who does this disorder affect? In

repeated observations, anorexia nervosa is overrepresented in the upper social

classes (Garfinkel & Garner, 1982; Hoek, 1993; Levine, 1987). The majority of

anorexics are White females (90% to 95%) who develop the disorder before they

are 25 years of age (Crisp, 1988; Garfinkel & Garner, 1982; Levine, 1987).

There has been a shift toward a higher representation in middle and lower

classes and Blacks (Garfinkel & Garner, 1982). Eating disorders are manifested

in other groups, for example, in males and in older females, however, this occurs

to a significantly smaller degree.

Bulimia nervosa is characterized by recurrent episodes of binge-eating

foilowed by repeated attempts to purge the body of any perceived excess weight

(DSM-IV, 1994). Frequent weight fluctuations and depressed mood are also

cornmon concomitant disturbances. These are, however, not necessary

symptoms of this condition (Levine, 1987). Although not receiving as much

attention, until recently, as anorexia nervosa, it is nevertheless a severe problem

affecting an estimated 5% to 10% of young females (Levine, 1987). Anorexia

nervosa and bulimia nervosa are not mutually exclusive disorders. In other


words, it is not uncornmon to find the two conditions coexisting within a single

person.

Media influences

A concurrent trend, associated with eating disorders and sweeping

through western culture, is the diet craze. Society is placing tremendous

pressure on women to diet as a means for attaining a thinner body shape

(Gamer, Garfinkel, Schwartz, & Thompson, 1980; Gamer, Garfinkel, & Olmsted,

1 983).

Diet advertisements and diet articles are difficult to avoid in the media

(Kilboume, 1993; Wiseman, Gunning, & Gray, 1993). A glance through a recent

newspaper or woman's magazine reveals countless ads for weight-loss clinics,

new diet books, and calorie-wise foods. An unforgettable image is the "before-

weight-loss" versus "after-weight-lossn campaigns. Relatively heavy women, or

occasionally men, are photographed pior to embarking on weight-loss programs

and, some time thereafter, rephotographed after a significant weight-loss. In the

former, the individuals are often portrayed as unhappy with themselves and with

life in general. In the latter, the opposite is portrayed: happy people who are

pleased with their lives. More often than not, the implication is that these altered

body shapes drastically improved al1 realms of the successful dieters' lives. We,

the readers, are led to believe that Our lives will also improve if we subscribe to

their weight-loss programs.

Television commercials and programs convey the same messages. A

recent example was Oprah Winfrey from the Oprah daytime talk show, who

Preventinq Disordered Eating 4

dieted and substantially reduced her weight. She sent her audience a familiar

message: "I look great, 1 feel healthier, and 1 am a better person now that I am

thin." Oprah, proudly modeling her slim proportions, perpetuates the belief

common in fernales (and to a lesser extent in males): The secret to success,

wealth, and happiness lies in the attainment of a thin body. This is only a single

instance of a common theme. Not surprisingly, Oprah later regained all the

weight she lost, plus an additional few pounds!

The Oprah story continues. More recently, she hired herself a full-time

chef who specializes in preparing healthy, low fat cuisine. The chef cooks al1 of

Oprah's meals, adhering to highly nutritionally balanced guidelines. This more

moderate approach to eating, along with a regular exercise routine, has resulted

in Oprah gradually reducing and stabilizing her weight. Oprah's latest regime is

consistent with the message promoted in the nutrition unit of the present eating

disorders intervention curriculum.

Society is given the impression that successful dieting followed by weight

loss makes for a better person. Is this true? According to two prominent

researchers in the field (Polivy & Heman, 1983) this is not necessarily the case:

The ultimate goals - weight loss, slimness, attractiveness, health, and

happiness -- have been taken for granted. Also taken for granted is the

notion that these various goals al1 go together ... there is an accumulating

store of research that indicates that neither health, nor happiness, nor

attractiveness, nor even slimness necessarily follows from dieting ... for

many of us, these various goals may be directly incompatible. (p. 8)


Society's obsession with dieting as a means toward attaining the coveted

thin body shape may be causing, or at least contributing somewhat, to the

current increase in prevalence of eating disorders (Bruch, 1978; Garner et al.,

1983).

Sociocultural standards of beauty

Until Garner et al.'s (1 980) study, there were no empirical data supporting

the correlation between eating disorders and sociocultural factors. The authors

noted a shift, particularly during the 1970s "in the idealized female shape from

the voluptuous, curved figure to the angular, lean look of today" (p. 483). Gamer

et al. (1980) set out to support their hypothesis through studying female body

sizes from two sources: Playboy magazine centerfolds and Miss America

pageant contestants. Both offer an indirect index of contemporary noms for

desirable body sizes in women. Average bust, waist, hip, height, and weight

measurements were calculated for al1 monthly playmates over a 20-year period

(1 959- 1 978). In addition, these average weights were compared to the

American popuiation's mean-female-weights over the same period. Gamer et al.

(1980) found a significant decrease in percent of average weight for the

centerfolds over the 20 years.

Data were also collected for the contestants and the winners of the Miss

Arnerica pageant from the same 20-year period. The pattern paralleled what

Gamer et al. (1 980) found from analyzing the magazine centerfolds: a significant,

gradua1 decline in average weight. Noteworthy is the additional finding that since


1970, winners have weighed significantly less than contestants. Wiseman et al.

(1 992), in a follow-up study, found a continuation of the earlier obsession with

thinness when they measured the physical attributes of Playboy centerfolds and

Miss America coritestants from 1979 to 1988.

Silverstein, Peterson, and Perdue (1 986) studied correlates of the thin

standard of bodily attractiveness for women using an extension of Garner et al.'s

(1 980) methodology, lengthening the time period to 80 years. Pictures of women

from Vogue and Ladies Home Journal - women's magazines chosen because of

their popularity and influence among women -- were measured. Using ratios

between bust, waist, and hips the authors measured curvaceousness.

Silverstein, Peterson, and Purdue (1986) found a very low bust/waist ratio during

the 1920s and during the late 1960s to earfy 1970s. This is indicative of

noncurvaceousness, or slimness.

Silverstein, Perdue, Peterson, and Kelly (1 986) set out to demonstrate the

role of the mass media in promoting a thin standard of attractiveness for women.

They followed three lines of reasoning: first, they demonstrated that 'Yhe current

standard of attractiveness for women portrayed in the media is slimmer than for

men". Second, "that the portrayed standard is slimmer now than in the past".

And third, that the "findings apply to many of the major media" (p. 519).

In the first method, men and women characters from 33 current television

programs were rated using a weight-rating scale created especially for this study.

Silverstein et al. (1 986) rated female characters as significantly thinner than their

Preventina Disordered Eatina 7

male television counterparts. Conversely, fewer of the females were rated as

heavy compared to the males.

Another media source, popular women's and men's magazines, were

analyzed for the number of advertisernents and articles about body shape and

size, dieting, and food-related ads. Women's magazines contained many more

advertisements for diet foods (63 to l ) , articles dealing with body shape and size

(96 to 8), and food ads (73 to 3) than men's magazines (Silverstein et al., 1986).

Women receive a barrage of contradictory messages. Television

characters and diet ads urge them to do whatever is necessary to maintain a

slim physique; and at the same time, women are required to resist the temptation

to indulge in the foods presented to them in magazines and other mass media.

Dieting

A correlate of the cultural emphasis on thinness in women is the diet

craze. Even though repeated dieting constitutes a risk factor for anorexia

newosa and bulimia (Smead, 1985), dieting is, for many individuals, perceived

as the norm:

The current societal preference for a thin physique has spawned a

corresponding societal preoccupation with dieting and weight loss. The

extent of this preoccupation is such that it may now be accurate to regard

dieting and its attendant diet mentality as normative ... normal eating now

requires periodic dieting. (Polivy & Herman, 1987, p. 635)


Abundant evidence supporting the ubiquity of dieting in our society is

found in the media (magazines, books, newspapers, television, films), the diet

industry itself, and in research. Earlier this was illustrated with two anecdotal

examples from magazines and television. There are corroborative psychological

studies.

Garner et al. (1 980) tabulated the number of articles about dieting and

weight loss from six popular women's magazines; advertisements were

excluded. Over a 20-year period (1959 through 1978), the number of diet

articles significantly increased. A follow-up study (Wiseman et al., 1 992)

revealed a continuation of this trend to the mid-1980s.

Wiseman et al. (1 993) also tabulated the number of television

commercials for diet products and diet foods on the three major American

networks fonn 1973 to 1991. They found a significant increase for both,

calculated as a percentage of total commercials, during this time Moreover, the

authors believe that this increase is continuing.

The diet industry is big business today. According to Mazur (1 986), the

growth of the diet industry is well known, and is reflected in the growth in

popularity of diet centers, diet clubs, and diet foods and beverages. Obviously,

this industry has a substantial vested financial interest in perpetuating the thin

body standard. If the obsession with dieting and striving for thinness subsides,

these businesses would stand to Iose billions of dollars in revenue.

In an indirect way, the diet industry's attempts to continue making big

money is related to the increase in eating disorders. This is not to Say they are


solely responsible, or they caused the increase - the opposite may be tnie, the

diet industry may have spmng up as a response to a societal demand - but they

nevertheless have reaped large benefits from promoting a relatively

unsuccessful fight against fat.

Diet books are one influential part of the diet business. Many are popular

and often appear at the top of the best seller lists. An often criticized diet book is

Mazel's (1 980) The Beverly Hills Diet. Apart from 1s populanty in terms of selling

a phenomenal number of copies, the book promotes anorexia-nervosa-like

eating patterns as a cure for fatness.

The Beverly Hills Diet marks the first time an eating disorder - anorexia

nervosa -- has been marketed as a cure for obesity. It is a case of one

disease being offered as a cure for another ... The popularity of her diet

can be seen as yet another symptom of a weight-obsessed culture.

(Wooley & Wooley, 1982, p. 57)

Mazel's (1 980) diet is a prescribed lifelong maintenance plan for a slim

body shape. The central feature is how to accommodate those unavoidable, but

frequent, eating binges. Punishment proceeding a binge episode entails eating

only one type of fruit per day until weight retums to the pre-binge, desirable level.

Wooley and Wooley (1 982) cal1 The Beverlv Hills Diet (1 980) "training in

anorexic psychopathology" (p. 65). That the book sold millions of copies is an

indication of our culture's obsession with dieting and the lengths individuals will

go to obtain and maintain a slirn physique.

Preventing Disordered Eating 1 0

Media influences, sociocultural standards of beauty, and the diet industry,

affect opinions regarding what is an optimal body weight, and consequently

influence self-esteem levels, and attitudes and behaviors toward eating. The

effects of these forces rnay be particularly damaging on younger children who

are still in their formative years, and hence more influenceable and

impressionable. Countering the potentially negative effects of these pervasive

societal influences on children rnay reduce the incidence of disordered eating,

eating disorders, and maladaptive behaviors and attitudes toward body size,

eating, and dieting. Primary prevention programs are a means to achieve this

end.

Primary prevention: What and why?

What is primary prevention? Primary prevention targets the entire

population or high risk groups in an attempt to prevent a problem's or a

disorder's widespread manifestation before it becomes insurmountable. In

contrast to secondary prevention, which focuses on reducing the duration of a

disorder, and tertiary prevention, which focuses on the treatment of individuafs,

the goal of primary prevention is to reduce incidence levels in an entire targeted

population (Cowen, 1983; Mann, 1 978: Rappaport, 1 977).

Albee (1 982) argues cogently in favor of the widespread application of

prirnary prevention strategies in psychology. This approach, he stresses, allows

for the maximum allocation of limited financial resources and social service

agencies toward dealing with widespread psychoiogical problems and disorders

at the individual and societal level. Albee (1982) believes that the favored


alternative, entailing treatrnent after a problem or illness has already manifest

itself, is costly and ineffective when viewed from the big picture.

Discussion regarding primary prevention for disordered eating and eating

disorders has received increasing attention throughout recent years (Chng,

1983; Collins, 1988; Crisp, 1979; 1988; Piran, 1995; Shisslak & Crago, 1987,

1993). Arthur Crisp, a well known and respected author and psychiatrist

speciaiizing in the field of eating disorders, writes in favor of applying primary

prevention to deal with this increasing problem (1 979; 1988). He recommends

developing pilot studies to permit the evaluation of intervention programs, and

outlines a nurnber of goals and specific strategies.

Shisslak and Crago (1 987) strongly encourage the development and

implernentation of eating disorder primary prevention programs, and they outline

exercises and assignments appropriate for conducting such programs. The

same authors (Shisslak & Crago, 1993) comment on the politics of prevention,

and discuss the battle facing advocates of prevention against the powerful diet

industry and big business as an example of the challenges facing those involved

in this field. Three years after they published their earlier article, Shisslak and

Crago (1 990) designed and carried out a successful pilot program with high

school students. Their work is summarized in later sections.

Additional comrnents on the resistance to prevention programs are offered

by Steiner-Adair (1 993). She eloquently and passionately argues that there are

powerfully ernbedded community values that support Western society's ongoing


disordered eating problem, and that preventative efforts will have to contend with

these forces.

Accurate knowledge of causal variables for eating disorders, according to

Vandereycken and Meermann (1984), is crucial when developing prevention

programs for anorexia nervosa. They question the feasibility of these programs,

rernarking that it is difficult to change certain predisposing variables such as

middle-class family characte ristics and sociocult u ral influences. Nevertheless,

Vandereycken and Meermann (1 984) stress that specific attention should be

paid to recognition of pre-anorexics through an evaluation of certain known risk

factors in combination with reliable screening instruments.

Yager (1985) and Katz (1985) disagree with Vandereycken and

Meemann's (1 984) pessimistic opinions. For example, Yager reasons that

sociocultural pressures to be thin could possibly be changed through concerted

public awareness of the detrimental effects of these attitudes and pressure for

change. Katz (1 985) believes that countering sociocultural influences may be

the most appropriate target for eating disorder prevention programs.

Considerations prior to establishing preventive interventions for

disordered eating were outlined by Smead (1985). For instance, she proposes

that material designed to counter the prevailing attitudes toward thinness must

be presented in a careful and thoughtful manner because it faces a strong

adversary and vested interest from the multi-billion dollar diet and weight-loss

industry.

Preventing Disordered Eatinq 13

Gamer (1985) also favors primary prevention for eating disorders and

States that this would be enormously cost-effective in human and economic

terms and probably beneficial in most instances. Potential iatrogenic effects

(such as exacerbating children's preoccupation with their bodies and increasing

dieting behaviors), however, no matter how well-intentioned the program

designers, may be an unforeseen adverse component. In order to minimize this

nsk, Gamer (1985) recommends that special care be taken when planning an

eating disorder prevention program.

Pre- and early adolescents appear to be the most logical recipients for

disordered eating prevention programs. Crisp (1 988) recommends targeting 1 1 - to 16-year-old students, whereas Shisslak and Crago (1 987) encourage

developing programs for students as early as the junior-high-school level.

Moreover, many successful smoking and drug prevention programs have

focused on pre-adolescents (e.g., Luepker, Johnson, Murray, & Pechacek, 1 983;

Perry, Killen, & Slinkard, 1980).

Most children at the ages of 11 - and 12-years-old are old enough to

understand, and therefore hopefully benefit from a disordered eating prevention

program. Preadolescence may be a critical stage in developing attitudes and

behaviors associated with disordered eating and eating disorders. Unhealthy

attitudes toward weight and dieting issues, which rnay have evolved out of years

of parent-, family-, peer-, and society-child interactions, may not actually cause

harm until these attitudes interact with a pubertal weight gain. This may be a

primary contributor to the development of disordered eating and eating disorders.

Preventincr Disordered Eatina 14

The facts that anorexia nervosa is commonly first manifested between the ages

of 14 to 18, that the peak age of onset for bulimia nervosa is 16 (Levine, 1987;

Wenar, 1994), and the belief that unhealthy attitudes may begin to develop early

in childhood, provide reasons that efforts toward changing students' attitudes

should be optimal before the onset of the pubertal physical growth stage.

Recent studies appear to support this assertion (Gustafson-Larson & Terry,

1992; Mellin, Irwin, 8 Scully, 1992).

Pre-adolescent children, it is believed, are still relatively impressionable

and open minded, and rnay be more willing to Iisten and change their attitudes,

beliefs toward eating, diet, and body shape, than older adolescents and adults.

Substance abuse prevention programs

A growing body of research in the area of substance abuse prevention

demonstrates the feasibility, general success, and potential benefits of applied

prevention prograrns (Flay et al., 1985; Killen, 1985; Perry et al., 1980;

Schaps, Dibartolo, Moskowitz, Palley, & Churgin, 1981). Specifically, in targeting

junior high school students, cigarette smoking prevention programs have been

effective in reducing the incidence of smokers immediately following the sessions

and during long-terni follow-up (Luepker et al., 1983; Perry et al., 1980). These

and other studies may be useful in providing some of the groundwork for

developing and conducting effective disordered eating prevention prog rams.

Schaps et al. (1981) reviewed 127 drug abuse prevention programs,

pointing out the characteristics of the best studies. According to the authors,

these experiments included: good, detailed program descriptions, detailed


demographic information on the participants, use of multiple dependent

measurement techniques, pre- and posttesting, and making sure respondents

understand the questions asked. Applying these features to disordered eating

prevention prograrns may enhance the quality of the studies.

Prevalence of disordered eating and eating disorders

Hilda Bruch, in the preface to her book The Golden Caae (1 978), writes:

For the last fifteen or twenty years anorexia nervosa is occumng at a

rapidly increasing rate. Formerly it was exceedingly rare ... Now it is sa

common that it represents a real problem ... One might speak of an

epidemic illness. (pp. vii-viii)

Research pertaining to the epidemiology and prevalence of disordered

eating and eating disorders may be organized into four general groups:

1. Initial articles describing the prevalence of eating disorders in the general

population, where the authors wished to inform professionals and the lay public

of the existence of eating disorders, and alert them to the seriousness of these

conditions. Attention was directed, in particular, to an adult, fernale population.

2. Subsequent articles demonstrating the trernendous increase in the

prevalence of eating disorders and disordered eating during the previous 15 to

20 years. The focus, again, was mainly on adult and college-age women.

3. Studies illustrating a trend toward development of disordered eating, eating

disorders, and associated behaviors (e.g., dieting, preoccupation with thinness)


in a younger population: high school students in the 13- to 18-year-old age

range.

4. Cuvent data, perhaps the most distressing of all, illuminating the prevalence

of disordered eating , dieting, and maladaptive attitudes and behaviors toward

weight and body-size, among young children. These studies demonstrate that

the insidious precursors for developing disordered eating patterns and eating

disorders are apparently set in motion early in childhood.

Initial and subsequent epidemiological research

Evidence supports the notion of an increasing incidence of eating

disorders in recent years. I will btiefly review three supportive studies: Theander

(1 WO), Kendall, Hall, Haiiey, and Babigian (1 973), and Jones et al. (1 980).

The earliest attempt at documenting the incidence of anorexia nervosa for

a defined population was by Theander (1970). Data were taken from psychiatric

and medical records at two Swedish university hospitals during the years 1931 to

1960. An annual mcidence rate of .24 per 100,000 was obsewed.

In a subsequent study, Kendall et al. (1 973) extracted information on

anorexia nervosa patients from three pçychiatric case registrars: North-East

Scotland, Cambemvall (an area in London), and Monroe County, New York.

Length of data collection varied according to area. The authors acknowledged

an inevitable degree of under-reporting the incidence of the eating disorder

owing to their assumption that nearly al1 anorexics in the designated areas were

reported to the register. Nevertheless, anorexia nervosa increased in incidence

in al1 three register areas.


Jones et al. (1 980) criticized the two previous attempts for their probable

underestimation of the disorder's true incidence. Measurements used in

estimating anorexia nervosa from Theander's (1 970) and Kendall et aLts (1 973)

studies only included patients who entered hospitals in the formal health care

system for mental health care. The Kendall et al. (1 973) study is further

criticized by Jones et al. (1 980) for not including general hospital records. This

additional data would extend anorexia nervosa incidence to individuals in the

medical health care system in addition to the psychiatric health care system.

Jones et al. (1980) studied the epidemiology of anorexia nervosa in

Monroe County, New York from the years 1960 through 1969 and again from the

years 1970 through 1976. To overcome the aforementioned shortcornings and

obtain a more accurate measure, the authors investigated anorexia nervosa

incidence th roug h psych iatric case register and general hospital records. A

significant increase in the total num ber of individuals first diag nosed anorexic

was found in the years 1970 to 1976 when compared to the previous decade

(1 960 to 1969). The numbers per 100,000 population nearly doubled from .35 to

.64.

The evidence that anorexia nervosa is increasing comes from (a) a

general agreement by leading authorities in the field that there has been an

increase (e.g., Bruch, 1978), and (b) empirical evidence (Jones et al., 1980;

Kendall et al., 1973; Theander, IWO). There are, however, significant

arguments suggesting the contrary: that the increase is more apparent than real

(Schwartz et al., 1 982).


Schwartz et al. (1982) discussed four altemate expianations for the

apparent increase in anorexia nervosa. First, the trend rnay be rnay be attributed

to better record keeping. Second, a substantial increase in females under the

age of 30 resulting from the post-World-War-Two baby boom rnay be associated

with a higher incidence of eating disorders. This is most apparent when

incidence rates are recorded per 100,000 of the population as a whole. This

hypothesis is testable, for example, by calculating the proportion of anorexie-

fernales-under-30 per 100,000 to females under 30 in the population as a whole.

Prevalence in any arbitrarily chosen group could be tested in this manner.

Research in this area could be useful in that it rnay possibly predict a decrease in

prevalence when these babies of baby-boomers grow older and approach middle

age.

Third, the increase could be due to eating disorders being glamorized, not

chastised. Playgirl's 1975 "Golden Girl's Disease" article is an example of this

glamorization (Garner et al., 1983). Last, Schwartz (1 982) hypothesizes that the

increase could be a function of the confusion between the number of referrals to

centers and individuals specializing in eating disorders (an apparent increase),

and an actual increase.

An additional argument, related to Schwartz's (1 982) third point, is that the

higher incidence rate rnay be a by-product of the increased attention given to,

and recognition of, anorexia as a psychological disorder with potentially severe

consequences. The current heightened awareness rnay encourage individuals

to turn to psychiatric and medical hospitals, eating disorder clinics, or


psychologists for professional care, whereas in the past the same individuals

may have remained at home, untreated. Although impossible to discredit

entirely, it is unlikely that this possibility accounts for a substantial percentage of

the increase in eating disorders.

Prevalence of disordered eating and maladaptive attitudes in adolescents

Nylander (1 971) conducted a landmark study with the aim of discoverhg

how common it is that young people feel fat and how often they diet. In total,

2370 Swedish boys and girls, aged 14 to 21, received questionnaires.

Somewhat disturbingly, most of the girls in Nylandefs survey stated that they

had at some time felt fat and many responded that they presently felt fat. In both

cases, the percentage increased with age. Specifically, 25% of the 1 Cyear-old

girls and 50% of the 18-year-old girls considered themselves fat. Only 8% of the

girls claimeci that they had dieted before the age of 14, but this amount jumped

considerably after age 15.

The boys in Nylandef s (1 971) sample were significantly less likely to Say

that they feit fat or had dieted. The reason given for dieting by most of the girls

that were trying to lose weight, was that they felt fat. Nylander's disturbing

findings spurned many subsequent investigations in the area of adolescent

eating disorders and abnomal eating attitudes and behaviors. Four studies will

be described below.

Pope and colleagues (1 984) assessed the lifetirne prevalence of

disordered eating in three student populations. Their study differed from other

research by not simply assessing current eating-disorders behaviors and


attitudes (i.e., point prevalence). They administered an anonymous

questionnaire covering DSM-III criteria for bulimia and anorexia nervosa to 1060

students at two colleges and a secondary school. Between 1 .O% and 4.2% of

the respondents met the criteria for anorexia nervosa, and 6.5% to 18.6% met

the criteria for bulimia. No males in Pope et a1.k (1 984) study met the DSM-III

criteria. The authors concluded that eating disorders represent a serious health

problem and that their results suggest "alarmingly high prevalence rates of

eating disorders in al1 three student populations" (Pope et al., 1984, p. 49).

Rosen and Gross (1 987) surveyed 1 373 geographically, racially, and

economically diverse high school girls and boys from the northeastem United

States. Students in their study were asked "Are you currently trying to Iose or

gain weight?" Affirmative responses were followed with further probing, trying to

determine the methods of weight reducing and gaining. Options for weight

reducing included: exercising, decreasing calories, fasting, skipping meals,

vomiting, and cutting out junk food, snacks, or sweets. Exercising, increasing

caloric intake, and eating special foods were listed as weight gaining methods.

On the day of Rosen and Gross's (1 987) survey, 63% of girls and 16% of

boys reported that they were trying to lose weight; 9% of girls and 28% of boys

reported that they were trying to gain weight. The four favored methods of losing

weight, in descending order, were exercise (71 % of the girls; 61 % of the boys),

decreasing calories (65% of the girls; 43% of the boys), cutting out snacks (59%

of the giris; 43% of the boys), and skipping meals (34% of the girls; 20% of the

boys). Fasting, vomiting, and other techniques were mentioned less often. The


two most popular methods of gaining weight for boys were exercising (63%) and

increasing calories (60%). Girls preferred increasing their caloric intake (48%) as

a means for gaining weight.

In discussing their results, Rosen and Gross (1 987) concluded that some

fonn of weight modification is common among high school students, and that

these behaviors have increased twofold compared to studies of 15 and 20 years

ago.

Rosen and Gross (1987) also measured the students' actual weight.

They noted, first, that perhaps one of the most striking findings of their study was

that the majority of the students that were actively losing or gaining weight were

already in the normal weight range. In other words, the adolescents were not

over- or underweight. Second, Rosen and Gross (1987) commented that

whereas four times as many girls than boys were trying to lose weight, the

opposite pattern was observed for weight gainers. This finding confirms cultural

stereotypes dictating that girls should be slender and boys should be muscular.

The present study applies the same general questions regarding weight

gaining and losing, to a younger population.

Levine (1 987), in a book on student eating disorders, reviews the extent of

the problem in high school students. He estimates that:

1. Between one and six in every 200 girls will develop anorexia nervosa

between the ages of 12 and 20.

2. Six to ten percent of high school girls are bulimic at any given point in time.


3. At least 40% of White, middle class high school student girls are actively

engaged in losing weight at any given point in time.

4. The peak age of onset for anorexia nervosa is 14 to 18, whereas for bulirnia

nervosa it is 16 to 18.

5. Student eating disorders are increasing in prevalence.

Levine (1 987) arrived at the last estimation by cornparhg the rates from

the earlier epidemiological studies in the 1960s and 1970s to findings from the

1980s. He also delineated goals and guiding principles for developing eating

disorder prevention programs. A lesson plan for grade 7 to 12 teachers,

designed by Levine, provides direct, affirmative actions for his suggestions.

Numerous studies have consistently demonstrated that a large

percentage of adolescents, girls in particular, are dieting, concerned with weight

and image, and using drastic measures to reduce weight (Crowther, Post, &

Zaynor, 1985; Greenfeld, Quinlan, Harding, Giass, & Bliss, 1987; Killen et al.,

1986). For example, Crowther et al. (1985) found that substantial proportions of

adolescents in their sample were using self-induced vomiting, laxatives, and

fasting to control their weight.

Researchers have also documented that anorexia nervosa and bulimia

nervosa occur with a relatively high frequency in adolescent girls (Crowther et

al., 1985; Ledoux, Choquet, & Flament, 1991). Furthermore, Button and

Whitehouse (1 981) argued that a substantial proportion of post-pubescent

females (approximately 5%) develop a subclinical form of anorexia nervosa.


Subclinical cases present serious problems of eating and weight concem, but do

not fulfill the strict criteria for clinical anorexia nenrosa,

Prevalence of disordered eating in pre-adolescents

Determining the age of onset for disordered eating behaviors is

particularly important when developing a primary prevention program. Until

recently, no study targeted pre-adolescents in order to clarify this issue. A

comprehensive literature search led to the discovery of a gradually growing

number of relevant articles from the past few years.

Olsen (1 984, cited in Maloney, McGuire, Daniels, & Specker, 1989)

surveyed teenagers and found that self-reported dieting among adolescents

started as early as eight to ten years of age. Alaming as this may seem, the

author may have underestimated the extent of this problem. Brown and Forgay

(1 987) reported that by age 13, 60% of Arnerican girls have dieted. In a study of

fourth, fifth, and sixth grade children, between 3% and 4% reported self-induced

vomiting (Stein & Reichart, IWO). Salmons, Lewis, Rogers, Gatherer, and

Booth (1 988) found that 12% of girls and 9.9% of boys between 11 - and 13-

years-of-age were dissatisfied with their body shape, while 30% of boys and 35%

of girls were ternfied of gaining weight.

Maloney et al. (1 989) surveyed 31 8 predominantly White, middle-class

boys and girls from grades three through six, randomly selected from elementary

schools in Cincinnati. Their dependent variables included a children's version of

the Eating Attitudes Test (ChEAT), and a demographic and dieting

questionnaire. The overall mean ChEAT score from Maloney et al.'s (1989) data


was eight with a standard deviation of seven. They reported that 6.9% of

students scored in the anorectic range (120), closely matching the findings from

older populations. Moreover, 45% of the children wanted to be thinner, and 37%

responded that they had tried to !ose weight. Common methods of dieting

included exercising (40.3%), restricting calories (1 2.6%), and bingeing (1 0.4%).

Maloney et al. (1 989) concluded that concems about body fat and dieting

are common in 8- to 13-year-olds. They further hypothesized that anorectic

eating attitudes may be set during the preadolescent years but not acted out until

adolescence. Successful interventions, therefore, should focus on shifting

attitudes in relatively young children in order to prevent the development of later

problems.

A replication of Maloney et al.'s (1989) study, with Israeli school-age

children, was undertaken by Sassaon, Lewin, and Roth (1995). The ChEAT

and the demographic and dieting questionnaire were given to 186 grade three to

six and 270 grade seven to eleven students. They found that amoung the grade

three to six students, 8.8% scored above the ChEAT cutoff, and the mean

ChEAT score was 7.5. When asked if they were trying to lose weight, 29.4

percent of boys and 38.8 percent of girls responded positively. These findings,

which are sirnilar to the original study (Maloney et al., 1989), extend the cross-

cultural generalizability of these attitudes and behaviors in children.

Leichner, Amett, Rallo, Srikameswaran, and Vulcano (1 986) looked at

maladaptive eating attitudes in a large sample of Canadian males and females

throughout the province of Manitoba. The original Eating Attitudes Test (EAT-


40; Garner & Garfinkel, 1979) was given to 51 50 students between the ages of

12 and 20 from rural schoofs, urban schools, and the University of Manitoba.

The EAT-40, the first Eating Attitudes Test developed by Gamer and Garfinkel

(1979), was later factor analyzed and condensed to the EAT-26 (Garfinkel and

Garner, 1982). The two tests, which contain 40 and 26 questions, respectively,

are highly correlated and maintain equaI levels of validity and reliability. A score

of 30 and above on the EAT-40 reliably identifies the anorectic population

(compared to a score of 20 on the EAT-26).

From Leichner et al.'s (1 986) total sample, they found that 5% of males

and 22% of females scored in the anorectic range on the EAT, implying

significant abnormal concems and attitudes regarding eating. Specifically, 7.3OA

of 12-year-olds (219 in the sample) scored above the cutoff value, and this

number substantially jumped to 23.7% at age 13, remaining relatively constant

thereafter.

Leichner et al.'s (1 986) findings closely parallel the percentages obtained

by researchers on disordered eating behaviors and attitudes in pre-adolescents

(Maloney, McGuire , & Daniels, 1988; Maloney et al., 1989), and suggest that a

sharp increase in maladaptive eating attitudes occurs between ages 12 and 13.

These data lend further support for developing a prevention program that targets

pre-13-year-old (prepubertal) adolescents and children.

A recent survey and interviews witfi 1600 girls and 1530 boys in grades

five through eight (Childress, 1 991), the largest sample assessed to date,

revealed similar information, consistent with the disturbing trend from Maioney et


al. (1 989) and Leichner et al.'s (1986) research. In this South Carolina

population, Childress (1 991) found that 55% of girls and 28% of boys Say that

they are fat, when in fact only 13% are actually overweight. Among fifth and

sixth grade students, 30% of the girls and 25% of the boys had dieted. Data

from Black and White girls were compared, indicating that feeling fat and wanting

to lose weight were more prevalent among White children.

Brewerton (1992), director of the Eating Disorders Program at the Medical

University of South Carolina, presented Childress's (1 991) data at the annual

meeting of the National Mental Health Association. He added that a significant

number of the children had used fasting, vomiting, diet pills, or diuretics to lose

weight. This recent research furiher substantiates the evidence indicating that

maladaptive eating attitudes and behaviors are ubiquitous among young

children.

Recently there has been widespread interest in cross-cultural

comparisons of disordered eating patterns and eating disorders. Following this

trend, Hill and Bhatti (1 995) measured body shape dissatisfaction and dieting in -

nine-year-old British Asian girls, comparing them to Caucasian children.

Dependent variables consisted of the Body Esteem Scale (BES) and a 7-point

measure of body satisfaction and body shape preference. They found a high

priority for thinness in both groups, even though the Asian girls had a

significantly lower average body weight. Thus, there appears to be some cross-

cultural consistency in dieting and body satisfaction behaviors and attitudes.


In Richman's (1993) pilot prhary prevention program, grade five and six

students received questions on whether they were trying to lose or gain weight.

The results further substantiated the aforementioned surveys, generalizing them

to a Western Canadian, urban population. Richman found, specifically, that

24.7% of his study's 180 participants were trying to lose weight and 9.1 % of

respondents reported that they were trying to gain weight at pretest. Moreover,

30.1 % of females and 19.3% of males were trying to lose weight. Among the

methods reported by Richman's (1 993) students, 68.3% listed exercising, 75.6%

cutting out snacks, 22.0% skipping meals, 36.6% cutting down calories, and

7.3% listed fasting as dieting strategies used at pretest. ChEAT scores above

the cutoff value (120), which is suggestive of anorexia nervosa, were observed in

5.4% of the total sample during pretesting.

Behaviors and concems related to weight among 457 fourth-grade, white

children in 10 rural Amencan schools were assessed by Gustafson-Larson and

Terry (1992). Children were measured on the Body Mass Index (BMI) and other

self-report questionnaires. A high percentage a girls (60.3%) and boys (38.4%)

reported wanting to be thinner. The BMI was significantly associated with the

item indicating a desire for less body fat. As with many other studies, the

authors failed to use the appropriate statistical adjustment (Le., Bonferroi

correction) necessary when analyzing multiple items. Nevertheless, the authors

comment that their findings indicated a need for interventions among growing

children remains valid.


Mellin et al. (1992) studied dieting and disordered eating behaviors in a

group of 9- to 18-year-old, urban females. Dieting was reported by 31% of 9-

year-olds and 46% of 1 0-year-olds. Further, many respondents reported

bingeing, being afraid of becoming fat, and restrained eating. The percentage of

females endorsing the items increased with age. For example, when asked

"how often do you eat huge amounts of food" (bingeing item), 34% of 9-year

olds, 42% of 1 0-year-olds, 48% of 1 1 -year-olds, and 56% of 1 2-year-olds

responded affimatively. In al1 fairness to the children, it rnay be questionable

how they defined the terni huge. What is huge for one child rnay not accurately

resemble a true binge. The researchers note that their study shows that

characteristics of disordered eating may be widespread in urban, middle-class,

prepubescent girls. They state that interventions appear warranted, and further

urge that these programs should target children before middle-school years.

Eating disorder and disordered eating prevention programs

Shisslak and Crago (1 990) developed a pilot eating disorder prevention

project for high school students. The goal of their program was primarily

educational, aiming to teach students and teachers about incidence, symptoms,

and consequences of eating disorders. Participants, 50 sophomore students,

were exposed to the nine-week program during their healtti education class.

Eight information-oriented presentations were made by a psychologist and

teachers, with ample extra time allotted for classroom discussion. The program's

evaluation component consisted of a short questionnaire on eating disorders.


In cornparison to driver education or physical education control groups,

the treatment group correctly answered significantly more questions on the

eating disorder quiz. From their data, Shisslak and Crago (1 990) inferred that

high school students are generally receptive to their eating disorders prevention

program. Shisslak and Crago (1990) also demonstrated the feasibility of this

approach.

Commenting on the seriousness, scope, and significance of eating

disorders, Shisslak and Crago (1990) argue that it is imperative to institute

prevention programs. Although their pilot study was limited to an information-

lecture type design, it did set a precedent and provides encouragement for future

research.

Criticisms of Shisslak and Crago's (1 990) research are fourfold. First,

they did not include pretest measures. The authors questioned students only

after the program ended, and therefore deteminhg whether the treatment and

control groups differed prior to the program's initiation is not possible. Second,

because students were not randomly assigned to the two conditions, the

students in the health education classes (treatment group) may have been more

infomed and educated regarding eating disorders than students enrolled in

driver's education or physical education classes (control group). Students in the

health education classes presumably have more knowledge in health-related

issues such as eating disorders, and differences in the groups rnay have

reflected previous knowledge, not the effects of Shisslak and Crago's (1 990)

program. Third, by failing to address the link between knowledge and attitudes


and behavior, there is no guarantee that any knowledge gained by Shisslak and

Crago's (1990) students would lead to attitude and behavioral changes. Last, by

emphasizing information regarding eating disorders, the authors may have

unwittingly caused an iatrogenic effect (Garner, 1985). In other words, the

knowledge and awareness gained by the participants may contribute to the

development of disordered eating in certain susceptible individuals.

Before the past few years, published, scientific evaluations of disordered

eating and eating disorder interventions for preadolescents were nonexistent.

This paucity is thankfully changing. The recent trend, however, has been

graduai. Several authors recently investigated the efficacy of these programs

(Killen et al., 1 993; Moreno & Thelen, 1 993; Porter, Morrell, & Moriarty, 1 986;

Richrnan, 1993: Rosen, 1989). Their conclusions, at this point, have been

mixed, although contrary to previous reviews (Killen et al., 1993; Smolak &

Levine, 1994) 1 believe that these programs are promising.

Porter et al. (1 986) designed an innovative half-day inoculation program

consisting of art therapy, dance therapy, and music therapy -- al1 believed to be

beneficial in the treatrnent of disordered-eating related problems. For example,

according to the authors, dance therapy has been effective in enhancing body

awareness and correcting biased self-body perceptions. A brief film on anorexia

nervosa and discussion of eating disorders preceded the therapy components,

with the goal of inoculating (exposing the children to a weak dose of unwanted

behavior) the children. Porter et al.'s (1986) inoculation approach aimed at

gently educating students to resist participating in unwanted behaviors (Le.,


eating disorder related behaviors). This contrasts with a fear arousal technique

which attempts to induce behavior and attitude change with fear-enhancing

techniques.

In Porter et a l 3 (1 986) intervention, 44 boys and girls, aged 9 to 16,

volunteered as participants. Pre- and posttest eating disorder behaviors were

measured using the Eating Disorder Inventory (EDI). The authors found a

significant score reduction on the "Drive for Thinness" subscale of the €DI for the

sample as a whole. Scores on the other two €DI subscales ("Perfectionism" and

"Interpersonal Distrusr) decreased, but t-test values did not reach statistical

significance.

Porter et al. (1 986) admit that although their findings are interesting, the

data are preliminary and must be interpreted with caution. The absence of a

control group greatly restricts any interpretations of the results. F urthennore, the

program was brief and no long-term follow-up analysis was conducted. It is

therefore difficult to rule out alternative explanations for the observed changes

and also to predict whether the obtained effects were temporary or more

permanent. For example, demand characteristics or high subject expectations

may have accounted for Porter et al.'s (1 986) findings.

Moriarty, Shore, and Maxim (1990) evaluated an eating disorders

curriculum that was taught in Urban-Canadian elementary and high schools.

The 16-lessons were divided into four sections: diet and eating disorders, male

concems with eating disorders, sociocultural risk factors for eating disorders, and

ways to forestall these sociocultural influences. Their short-terrn analysis


revealed a significant increase in participant's knowledge and a more positive

attitude toward eating disorders at posttest. Although these findings are

promising, the evaluation did focus on measuring gains only on information and

knowledge.

Junior high school students took part in a brief preliminary prevention

program for eating disorders, designed by Moreno and Thelen (1 993). Students

watched a six-and-one-haff minute videotape of a conversation between two

sisters on eating disorders, and this was followed by a 30-minute class

discussion. A short questionnaire was administered to the female students and

a no-treatment comparison-group of girls before, after, and one month following

the intervention. Even though the program was brief, participants' knowledge,

attitudes and behaviora1 intentions regarding some aspects of their eating

behaviors were successfulIy changed frorn pretest to posttest. These effects

were also sustained after one month. Specific findings noted by the authors

were fourfold. First, Moreno and Thelen (1993) found a substantial effect on

their "diet" factor: at posttest the experimental group girls did not see strict

dieting as a good way to control their weight, whereas the control students did.

Second, the expeflrnental group girls gained more knowledge about the

undesirable physical effects of dieting than controls. Third, the intervention

group students exhibited fewer posttest concems with body weight and reported

a reduced Iikelihood that they would radically diet. Last, the girls who received

the program displayed more knowledge about the hanful effects of purging than

the control group.


Moreno and Thelen (1 993) concluded that their program was successful.

In critiquing their own study, the authors recommend booster sessions to

enhance and further sustain theses positive changes. Demand charactenstics -

that the respondents may have indicated sociaily desirable rather than accurate

answers - were discussed as a possible limitation. Moreover, Moreno and

Thelen (1 993) did not directly assess students behaviors. A Bonferroni

correction would have been appropriate, as they analyzed individual items and

did not mention correcting p-values for multiple measurernents. This appears,

however, unlikely to be a problem since their reported statistical values appeared

to have been sufficiently strong and still significant after any rnoderate

adjustments.

Killen et al. (1 993) published the first long-terni, controlled systematic

investigation of a comprehensive eating disorders program for children. Their

1 &lesson curriculum was randomly assigned to 967 sixth and seventh grade

girls and compared to no-treatrnent control classes. The intervention consisted

of three components: development of coping skills to resist sociocultural

pressures for thinness and dieting; promotion of healthy lifestyles through proper

nutrition and exercise; and information on unhealthy dieting practices and

normal pubertal weight fluctuations. The children were exposed to the lessons

through slide presentations and workbook exercises. Dependent measures used

for evaluating Killen et al.'s (1 993) curriculum included the Eating Disorder

lnventory (EDI), the Body Mass Index (BMI), their own knowledge questionnaire,

and a structured clinical interview. These were administered at four intervals


during the two-year period following the lessons. The authors concluded that the

prograrn failed to achieve the hoped for impact, citing the lack of statisticaily

significant effects from pretest to posttest and follow-up. A significant increase in

knowledge and a small change in the BMI for their high-risk subsample,

however, was observed.

Killen et al.'s (1 993) disappointing findings led the authors ta question the

wisdom of providing prevention curriculurns to non-clinically-disordered young

adolescents. Instead, they suggest that efforts and resources may be better

targeted toward high risk students. It is arguable whether Killen et al.'s results

are generalizable and applicable to al1 eating disorders interventions. In

reviewing Killen et al.'s (1993) study, Shisslak and Crago (1993) wonder if the full

effect of the program may not become evident for another year or two. They

seem to imply that Killen's pessimistic conclusions may be premature. Shisslak

and Crago (1 993) also question whether one-shot interventions are effective or

maybe less preferable to yearly, continuing programs. Furthermore, they

contend that the negative results of past programs should not hinder or prevent

our search for future programs. Instead, new research should look at means for

enhancing the curriculums' effectiveness. According to Shisslak and Crago

(1 993), eating disorder interventions are relatively new and require revision.

Their current, unpublished research compares the relative effectiveness of

longitudinal versus one-time interventions, peer group and peer-led versus

teacher led intervention, and the use of naturally forrned versus artificially formed

peer groups.


The disappointing findings by Killen et al. (1 993) and Rosen (1 989)

contrast sharply with the relative success of prevention programs by Richman

(1 993), Shisslak and Crago (1 990), Porter et al. (1 986), and Moreno and Thelen

(1 993). Possible explanations for this apparent discrepancy are offered later

(see Discussion section).

Following the publication of a number of articles describing specific

interventions, some authors have taken a further, and necessary, next step and

offered comrnents and constructive criticism aimed at improving the

effectiveness of primary prevention programs (Piran, 1995; Smolak & Levine,

1 994).

Piran (1 995) believes that the emphasis of current eating disorder

prevention curriculums should be changed from a didactic and knowledge-

dissemination approach toward a more experiential, affective, and personally

relevant approach. In her model, participants would generate their own ideas

and have an impact on the direction of the lessons, thereby making the content

more personally meaningful and powerful. This type of program reflects the

perspective known as qualitative research methodology, which contrasts sharply

with the more cornmonly accepted position in conternporary psychology: the

positivistic paradigm. Piran (1 995) also emphasizes paying more attention to the

relational aspect of the intervention. For example, peer group activities would be

an integral part of the lessons, thereby creating a small group of peers that could

act as a secure buffer against the strong sociocultural forces facing most


children. The inclusion of parents and significant others as part of the

preventative efforts are other ideas put forward by Piran.

Smolak and Levine (1 994) concluded that the effectiveness of prevention

programs aimed primarily at adolescents has been relatively unsuccessful. In

their evaluation, they state that the significant changes attributed to these

interventions has been largely restricted to gains in knowledge, and that effects

on attitudes and intentions regarding dieting and eating have been modest. The

reason, Smolak and Levine propose for the negative findings is that the

programs are too late in reaching the participants. Instead, they argue that

curriculums should be directed toward younger children, specifically elernentary

school students.

In response to Smolak and Levine (1994) there seems to be some

confusion regarding the exact ages of children they address. In their article they

apparently lump children that are in grade five and six in the same group as high

school students. Combining preadolescents and adolescents in the same

category may lead to erroneous overgeneralizations. It is unclear whether

Smolak and Levine are suggesting that programs be directed only toward the

grade four level, or if slightly older children are acceptable candidates for

prevention. In the present study, grade five and six, and some of the grade

sevens are classified as preadolescents. This arbitrary division between

preadolescents and adolescents differs from Smolak and Levine. In the future it

may be necessary to attain agreement in this area.


Richman (1 993), attempting to compensate for the lack of primary

prevention research and improve upon the educational, knowledge-focused

studies, conducted a six-week program with two grade five and two grade six

classes. Pretest, posttest, and follow-up questionnaires on knowledge,

disordered eating, maladaptive attitudes and behaviors, food intake, and body

satisfaction were administered to curriculum-recipients and three no-treatment

comparison classes. Overall results suggested that the intervention was

successful and effective in inducing change in the treatment condition. At the

posttest, treatment children's knowledge (KQ) about eating disorders, dieting,

and nutrition significantly increased compared to the comparison group.

Moreover, a significant increase was found for their level of body-satisfaction

(BSM). The percentages of children trying to lose and gain weight, and the

proportion of children scoring above the Children's Eating Attitudes Test

(ChEAT), al1 displayed a tendency in the positive direction. Specifically, a

smaller percentage of participants reported they were trying to lose or gain

weight, and their mean ChEAT score was lower (Le., indicating less malaptive

behaviors and attitudes) at posttest.

Follow-up analyses, 18 months after the end of the curriculum, also was

highly suggestive of the intervention's positive effect. Program participants

scored significantly different than nontreatment comparison students on the BSM

and ChEAT measures. Their knowledge scores, however, retumed to pretest

levels. It appears that although the information the students learned during the

prograrn may have been forgotten, many of the children who received the


lessons reported feeling more satisfied with their bodies and displayed less

maladaptive attitudes and behaviors regarding eating disorders and dieting than

their cornparison peer group.

The present study is a revision and continuation of Richman's (1993)

original pilot prevention program. Modifications include an expanded sample

size and a few minor c~mculum changes. Regarding the former, approximately

three times as many students participated in the present intervention, and they

came from more schools and classes. Furthermore, more teachers conducted

the sessions. In reference to the latter, a 30 minute video on eating disorders

(from the popular Degrassi junior high television sedes) followed by discussion,

descriptive overheads, and a few new exercises were added. Some minor

elements and exercises from the pilot program were omitted. Further

methodological information is presented in the method section.

Many of the above articles demonstrates the feasibility of conducting

disordered eating prevention research. It is encouraging to finally realize that

researchers are now willing to take the critical step of carrying out a prevention

program instead of merely perpetuating the rhetoric calling for the crucial need

for prevention research.

A model of prevention and disordered eating

lncorporating and synthesizing the previously discussed information on

disordered eating, primary prevention, sociocultural pressures, dieting, and

eating disorders, a model was devised in order to provide an overall framework

for the current intervention. This model includes conceptualizing disordered


eating dong a continuum, hypothesizing about the multiple causes and

maintaining factors associated with disordered eating and related attitudes, and

considering the complex and interactional nature of the variables. (For similar

models see Garfinkel & Garner, 1 982; Wenar, 1994.)

It is believed useful to conceptualize disordered eating and associated

behaviors and attitudes on a continuum from less severe at one end to more

severe and extreme cases at the other end. At the less severe pole would be

individuals with body weights within normal range, minor preoccupation with their

appearance and wanting to be thin, beliefs and attitudes that are relatively

amenable to charqe, ability to function in everyday life, and no diagnosed eating

disorder. At the other pole would be individuals diagnosed with clinical eating

disorders or those suffering frorn debilitating disordered eating behaviors and

attitudes. Their beliefs and fears of becorning fat may interfere with their daily

functioning, pemeate rnany areas of their lives, and be relatively non-amenable

to change. At the severe extreme would be those suffering from life-threatening

conditions and in state of danger. Their condition would be excessive,

debilitating, pervasive, and persistent.

In thinking about the present intervention a two-pronged approach was I

adopted. The primary airn was the reduction of disordered eating and

maladaptive behaviors and attitudes existing at the less severe end of the

continuum, and more applicable to the majority of the population. The

secondary goal, reducing the incidence and future developrnent of eating


disorders and behaviors and attitudes characteristic of the more extreme side of

the continuum, would possibly be relevant to a small minority of the participants.

The etiology of and variables involved in maintaining disordered eating

and maladaptive attitudes and behaviors are considered to be complex and

multifaceted. That is, the circumstances leading to the development of and the

factors associated with the perpetuation of these problems are thought to be

diverse. Some of the conditions hypothesized as important are at different

levels: sociocultural, interpersonal, and intrapsychic, and organic.

Sociocultural influences (as previously covered in depth) encompass

general pressures to be thin, media exposure (television, movies, magazines,

newspapers), and advertisements and product endorsements from the diet

industry. The Interpersonal realm refers to both familial and peer groups.

Familial factors include overt and subtle pressures from parents and siblings to

be thin, diet, and maintain a spcific body shape. Parents and siblings may also

impact children by modeling dieting behaviors and nondeliberately supporting the

sociocultural messages favoring thinness. Moreover certain conditions such as

parental alcoholism and parent conflict would increase the potential risk of

children developing disordered eating problems. As children grow older and

spend more of their time in school and with their peers, maladaptive attitudes

and behaviors prevalent in their friends and classmates rnay play an increasingly

strong part in affecting the way they feel, think and behave. Intrapsychic

variables entail how a person feels and thinks about themselves (i.e., self-

esteem), the person's intemal strengths and abilities to protect themselves from


maladaptive and harmful extemal influences (Le., resiliency and vulnerabilities).

Also relevant here would be the person's behavioral tendencies and

predispositions. For example a need to be perfect and a strong desire to control

self would possibly increase the risk of disordered eating problems. Certain

organic factors may also be involved. Above average inherited body weight,

slower metabolism, and a propensity toward gaining weight, are a few

instances.

The above variables - sociocultural, interpersonal, intrapsychic, and

organic - continuously interact throughout the individual's life development.

Depending on the individual's unique circumstances, these factors may lead

them along any number of pathways, ranging from the development of

disordered eating patterns and eating disorders to a more healthier outcorne.

The present study addresses the sociocultural, interpersonal, and intrapsychic

levels.

Present study

The purpose of the present study was to design and implement a primary

prevention program that facilitated change in participants' attitudes and

behaviors toward eating, dieting, self-esteem, and body satisfaction, and

increased their knowledge of disordered eating and eating disorders. Primary

goals of this program were:

1. lncreasing students' knowledge about disordered eating, eating disorders,

dieting, and nutrition.


2. Modifying participants eating patterns away from unhealthy eating styles and

toward healthier eating habits.

3. Enhancing student's' body- and self-esteem.

4. Encouraging students to resist social pressures and negative influences from

family and friends to be thin.

5. Prompting participants to question societal myths evolving around dieting and

overweight.

Pre-, post-, and follow-up testing cornparing treatment-group and

cornparison-group scores, enabled an assessrnent of the short- and long-term

success of these goals. It was hypothesized that the program would be effective

at inducing change and positively benefit the participants. In particular, three

subhypotheses were formulated. First, that there would be an increase in

disordered eating, nutrition, and diet-related knowledge in the program

participants (Cognitive component). Second, it was hypothesized that there

would be an increase in self-reported body-satisfaction and body-esteem in the

treatment group (Affective component). Last, that there would be a decrease in

maladaptive eating attitudes and behaviors, a positive change in dieting (i.e., a

decrease in the proportion of program participants reporting that they were trying

to lose weight), and gaining behaviors (i.e., a decrease in the proportion of

program participants reporting that they were trying to gain weight; Behavioral-

Attitudinal component) in the treatment group. No significant changes were

expected on these dependent rneasures for the cornparison group.


METHOD

Participants

Fifth, sixth, and seventh grade students from 23 classes participated in

the study. The disordered eating prevention program was received by: one

grade 5, six grade 6, four grade 5/6-split, and two grade 7 classes, frorn a total of

eight urban, schools in the Greater Vancouver region. These students

constituted the treatment condition. Cornparison group classes included: two

grade 5, three grade 516 split, two grade 6, one grade 6/7 split, and two grade 7

classes frorn three schools, also located in the Greater Vancouver region. The

teachers and principals in the treatment condition agreed to conduct the

prevention program, whereas the comparison group condition teachers chose to

participate in the pre-, post-, and follow-up testing, but not the program itself. At

follow-up, four treatment and three cornparison classes agreed to take part.

Data were initially collected from 664 respondents. Parents of nine

students in the treatment group decided against allowing their children to

participate in the intervention. ln the comparison classes, three students chose

not to fiIl out the pre- and posttest questionnaires, and one student decided not

to fiIl out only his posttest. Data were also not tabulated for three students who

did not comprehend most of the questions (i.e., for whom English is their second

language), and three students who missed more than two prevention program

sessions. Other students with poor English reading ability answered the

questions to the best of their abilities with assistance from either their teachers,

the primary researcher, or his assistant.


Pretest questionnaires and the first three lessons only, were completed by

two grade 6 classes in one school. Unfortunately, these students did not finish

the prevention program as their schedule became too busy and they were not

able to fit the remainder of the lessons or the posttest into their schedule. Their

data are only reported in certain pretest demographic sections of the results (see

"extended data set" in results section), and were not included in the overall

statistical analysis.

A grade five teacher, whose class was initially in the comparison group,

later decided to run the disordered eating program. Consequently, after the

originally scheduled posttest, this group was switched to the treatment condition

and received a second post-program posttest. Students from another grade five

class completed an early pretest (pre-pretest). These two classes were

classified as baseline groups and were analyzed separately, in addition to being

included in the overall sample. For these two special cases, data from the

original pretest measures (Le., pre-pretest) were included and used as

covariates in the overall analysis.

The final sample included 605 students: 335 in the treatment and 270 in

the comparison group condition. These children were used for tabulating the

Dieting and Gaining percentages, and mean scores.

Respondents who failed to answer two or more of the primary

questionnaires (KQ, BES, BSM, and ChEAT) at either pretest or posttest were

excluded from the ANCOVA statistical tabulations. The actual number of

respondents included in the analyses varied slightly for each of the questions


and tests, and at pretest, posttest, and follow-up. The missing data were from

students who were absent, for various reasons, from school during pre-, post-, or

follow-up testing, and from ambiguous or unmarked items.

The students were predominantly White, from middle- and upper-middle-

class families. Students ranged in age from 9- to 14-years-old, and the average

age was 10.98 years (treatment: 10.9; cornparison: 1 1.1). Dernographic

characteristics of the baseline samples are presented in Table 1. As shown in

Table 1, the three samples are roughly comparable in ratios of ethnicity and sex.

There were, however, fewer grade seven students in the follow-up classes.

A separate post hoc, posttest analysis was conducted for students

classified as High Risk. This subsample consisted of 34 children, 13 boys and

21 girls, who scored above the cutoff (120) on the ChEAT at pretest.

Of the initial 23 classes, seven teachers were asked to participate in the

follow-up component. This entailed administering the identical (pre-

test/posttest) questionnaires to their students, six months after the program's

conclusion. These particular classes were chosen based upon their earlier

starting date for conducting the intervention. This allowed for completion of the

follow-up within the same school year. All seven teachers agreed to participate -

- four treatment classes and three corn parison classes.

Follow-up data were initially collected from 194 respondents. From this

group, three children chose not to answer the questionnaires. Students who

were absent for either the pretest or foliow-up, or failed to complete two or more

Preventina Disordered Eatincr 46

of the pretest or follow-up primary questionnaires, were omitted from the

analyses. The final follow-up sample included 191 students.

The Extended pretest sample, totaling 664 students, included al1 students

present at pretest. For purposes of this Extended group, no participants were

omitted based on the nurnber of questionnaires they completed.

Dependent Measures

The dependent variables were conceptually organized into three

categories - Cognitive, Behavioral-Attitudinal, and Affective - corresponding to

the three primary hypotheses. This allowed for assessment of students'

knowledge, attitudes and feelings, and behaviors, measured at pre-, posttest,

and follow-up. (see Appendix A for the questionnaires)

Coanitive Measure - Knowledge Questionnaire. The Cognitive

component consisted of the Knowledge Questionnaire (KQ), constructed by the

author (Richman, 1993). The KQ is a 10 question multiple-choice exarn on

general eating disorders, dieting, and nutrition. Questions are based on material

covered in the prevention program. For each item, respondents choose one out

of four possible answers. The items Vary in difficulty. One point is awarded for

each correct answer, with possible scores ranging from zero to a maximum total

of 10 points. Scores above 6 suggest relatively competent knowledge of the

subject area. Test-retest reliability for the KQ in the current study was relatively

stable (1=.59; a=270) over a three- month period.

Behavioral-Attitudinal Measures. In the Behavioral-Attitudinal

component, assessment encompassed weight-losing behaviors, weight-gaining


behaviors, and attitudes related to eating, dieting, and food. The children's

version of the Eating Attitudes Test (ChEAT; Maloney et al., 1988; Maloney et

al., 1989). the dieting, and the gaining questions were used to measure these

qualities.

ChEAT. The ChEAT is a 26-item, self-report questionnaire used to

assess abnormal eating attitudes and behaviors. Testees answer forced-choice

items on a 6-point Likert scale ranging from never to always. Responses yield a

total score and three empirically (factor analytically) derived subscales: dieting

behaviors, bulimia and food preoccupation, and oral control (self-control of eating

and concerns about being oveiweight). Possible total scores range from zero to

78, with scores greater than or equal to 20 suggestive of anorexia neivosa

(Maloney et al., 1988). Sampled on 8- to 13-year-olds, the ChEAT is a

modification of the Eating Attitudes Test (EAT) designed by Garner and Garfinkel

(1979), with simple synonyms replacing the more difficult words. The ChEAT's

test-retest (1=.81) and intemal reliability (1z.76) are comparable to the original

version (Maloney et al., 1988).

The EAT total score is highly correlated with the Eating Disorder lnventory

(EDI; Garner & Olmsted, 1984) drive for thinness subscale (Raciti & Norcross,

1987). Garner et al. (1 987) studied 11 - to 1 Cyear-old ballet students and found

that this subscale was particularly relevant in predicting future eating disorders.

The EAT and ChEAT may therefore have some predictive utility for screening

individuals at high risk for developing eating disorders.


Dieting and Gaining. Students were asked whether they were currently

trying to lose (dieting) or gain (gaining) weight. If the answer to the former

question was affirmative, they were further asked to check off the way(s) they

used to lose weight. The listed options were: fasting, skipping meals, cutting out

snacks, cutting down calories, exercising, vomiting, and other. The format for

this section was based on a questionnaire devised by Rosen and Gross (1 987)

in their study on the prevalence of weight reducing and gaining in adolescents.

Affective Measures. The Affective cornponent included the Body

Satisfaction Measure (BSM; Richman, 1993) and the Body-Esteem Scale (BES;

Mendelson & White, 1982).

Body Satisfaction Measure. Attitudes toward the students' own physical

appearance was estimated with a Body Satisfaction Measure (BSM), constructed

by the author. Respondents indicated, on a 6-point scale, the degree to which

they were happy or unhappy with the way their bodies look. Possible responses

on this one-item continuum ranged from zero to five, with five representing "very

happy with the way my body looksn and zero representing "not happy with the

way my body looks."

Body Esteem Scale. The BES (Mendelson & White, 1982), a 24-item

self-report measure of body satisfaction, questions how students feel about their

own body appearance and how they feel others evaluate their body. Students

answer yes or no to the items. The BES is suitable for young children (designed

at a grade two reading level), demonstrates good split-half reliability (~=.85), and

correlates highly with self-esteem measures (1=.68 with PiersHarris Children's


Self-Concept Scale). The BES and BSM were correlated to check on the

relationship between the two measures, and to test the BSM's criterion validity.

Additional Measure. Body Mass Index. A body-index measure (BMI;

weight:height ratio) was also recorded. This measure of weight (in kilograms)

divided by height-squared (in centimeters) yields a BMI value, a preferred index

of body weight (Killen et al., 1993; see Health Services and Promotion Branch,

Health and Welfare Canada, 1988). Students were asked to estimate their

height and weight. This measure was correlated with the ChEAT (and other

measures). The BMI is a reliable and commonly used measure of body mass

(Killen et al., 1993).

Procedure

Initial application packages were sent to three superintendents from the

Lower Mainland public school system and four principals at private schools.

Superintendents from two of the boards, and one private school principal gave

their assent to conduct the prevention program.

Cover letters (see Appendix B), brîefly describing the disordered eating

prevention program, were then mailed to 29 principals of Bumaby and West

Vancouver elementary schools that were selected from the school board's

listings. The principals were later contacted by telephone and asked if grade

five, six, or seven classes in their schools would be willing to participate. Ten of

the principals responded positively and provided names of teachers in their

schools for further contact.


The offer to teach the disordered eating program was declined by 19

schools. Reasons given for not wishing to take part in the project were that the

principals or teachers were not interested, and/or that they did not have enough

time and available space in their school curriculums.

Design

These efforts led to 13 grade five, six, and seven teachers frorn 10

schools agreeing to run the program, and 10 grade five through seven teachers

who chose not to partake in the disordered eating program, but volunteered their

classes for the cornparison group condition.

Letters were distributed to the students' parents, before the program

commenced, briefly describing the intervention and enabling them to withdraw

their children from participating (see Appendix C). Parents of nine children in the

treatment group decided to pursue this option. Participants in both conditions

received written consent foms prior to completing the questionnaires (see

Appendk D). Before asking the students to sign their names, the researcher

briefly explained the material on the consent forms.

The researcher met with al1 of the teachers in person or spoke with them

on the telephone prior to the program's start. In these informal meetings the

lesson plans were reviewed and questions answered. The intervention itself, a

revision of Rice's (1 989) eating disorders prevention prograrn, encompassed six

modules of approximately one to two hours each, and was run on a weekly basis

by most of the teachers during regular classroorn hours.


There were variations in delivering the prevention program. In one

school, an unforeseen interruption led to approximately a one month break in the

middle of the program for its five classes. As a consequence of these unplanned

circumstances, these students received the first three lessons, then waited for

more than four weeks before partaking in the last three lessons. Lack of

continuity was evidenced and noticed by the researcher, students, and teachers

in these groups. For example, during the posttesting session sorne of the

students mentioned, and were fnistrated by the fact, that they had forgotten

material they had leamed from the beginning of the program. Moreover, it is also

noteworthy that in the aforementioned school each lesson was taught by a

different teacher. This scenario differed from al1 of the other classes where one

teacher taught the entire program.

In another school, two teachers taught the entire program within two

weeks. These teachers began the lessons later in the school year and wanted

to fit the intervention into their class's schedule by condensing the overall length.

Other than the shortened time frame, the lessons were taught similarly to the

other classes.

Pretesting preceded the program by approximately one week, and

posttesting followed the last lesson within one week. A follow-up was conducted

for seven of the classes six months after the end of the lessons. The classes

were selected based on when, during the school year, they started the

intervention. It was only feasible to use the groups who started eariier on during


the school year for further testing. Two classes participated in a pre-pretest

(baseline). Briefly, and in general, the project proceeded as follows:

Week 1 : General Orientation. Pre-test measures, introduction to the

program, and overview of forthcoming sessions.

Week 2: Introduction to eating disorders. Brief background on anorexia

nervosa and bulimia nervosa. Eating disorders defined, followed by an

exploration of what causes them to develop. Video on eating disorders.

Week 3: Social pressures to be thin. Looks at weight obsession in

society, pressures to be thin, and prejudice against fat.

Week 4: So what is healthy? Examines the harmful effects of dieting and

myths about dieting and weight.

Week 5: Influence of family and friends. Focuses on helping students to

identify and resist pressures to diet and lose weight.

Week 6: Eating well. Promotes healthy eating habits by presenting

accurate information and research on nutrition and diet.

Week 7: Self-esteem. Looks at self-esteem in terms of how it develops

and how it can become related to body size.

Week 8: Conclusion. Review and wrap-up. Open discussion on how

students evaluate the program. Comments. Post-test.

Uniformity in running the lessons was encouraged by the researcher.

Some of the teachers, however, decided to add their own minor modifications to

the intervention. For example, some added an extra video about heaithy eating

styles, one teacher read a book about anorexia nervosa to her class, and


another teacher had students draw pictures of themselves. As mentioned

earlier, the duration and frequency of the lessons varied between classes. The

program's overall length ranged from two weeks to more than three months.

A letter (see Appendix E), describing the study's general findings, was

distributed to the parents of the participants after the questionnaire data were

analyzed. lncluded in this letter was the telephone number for an Eating

Disorder Resource Center, intended for any parents with specific concems or

those wishing further services and information.

Proceeding the follow-up, the researcher met with the students in the

participating classes. The initial findings were explained, and the children were

encouraged to ask questions and offer their comments on the prevention

program.

Following the completion of the program, the author spoke with the

teachers who led the treatment groups. They were asked to provide feedback,

criticisms, and recommendations.

Preventing Disordered Eating - 54

RESULTS

Analysis of Posttest Scores (Uncorrected)

The three primary hypotheses were initially analyzed using univariate 2

(condition: treatrnent versus comparison group) X 2 (sex: male versus female)

Analysis of Covariance (ANCOVA) designs for each of the dependent measures.

Posttest means were compared using the students' pretest scores and age as

covariates. In al1 cases the pretest covariates related to the dependent

measures, confiming the rationale for using the ANCOVA procedure.

Differences between the pre-treatment means for the comparison group and

treatment conditions were not statistically significant.

A Bonferroni correction was used to adjust for the multiple hypotheses

and multiple dependent variables. A -1 5 study-wise level of significance was

chosen and divided by the number of primary dependent variables (five). This

resulted in a critical value of .O3 that was applied in determining statistical

significance for the ANCOVA F-ratios. Further Bonferroni corrections were made

when analyzing the three ChEAT subscales, resulting in a .O1 test-wise critical

value .

Unadjusted pretest and posttest means for the treatment and comparison

groups, by condition and sex, on the KQ, BEM, BSM, and ChEAT, are presented

in Tables 4,6, and 8, and 10. Pretest and posttest mean scores are also shown

in Figures 1-5.

Main effects for condition and sex were found on the Cognitive Measure

(KQ), E(l,456)=276.27, ge.0001, and E(1,456)=11.61, p<.001, respectively.


Significant main effects for condition were also found on the BES (Affective),

F(2,441)=12.55, pc0005, and ChEAT (Behavioral-Attitudinal), E(1,460)=5.56, -

ac.02, measures. There were no statistically significant interactions.

Coanitive - Knowledge Questionnaire. Means on the KQ

show that children in the treatment condition (M=6.33) scored significantly higher

than comparison group students (M=3.78) on the posttest. Moreover, females

(M=5.04) in both conditions perforrned better than their male (M=4.00)

classmates on the KQ, as indicated by a significant main effect for sex. Mean

KQ scores increased from lower to higher grades: grade 5 (M=4.1 O), grade 6

(M=4.78), and grade 7 (M=5.02). The complete ANCOVA for the Cognitive

Measure (KQ) is presented in Table 3. Refer tu Table 4 for the detailed

breakdown of KQ means.

Affective Measures - BES. Further analysis of the Affective component

revealed that treatment group means (M=18.80) were significantly higher than

the comparison group means (M=16.49) for the BES at posttest. The ANCOVA

value for condition on the BES was statistically significant, E(1,441)=12.55,

g<.0005. This suggests that boys and girls in the treatment condition, on the

whole, were more satisfied with their physical appearance at post-test than the

comparison group children. There were no statistically significant sex

differences. The complete BES ANCOVA is presented in Table 5. Refer to

Table 6 for a detailed breakdown of the means.


BSM. A main effect for sex on the BSM, E(4.05)= ~c.05, revealed higher

mean scores for boys compared to girls. This suggests that boys, in general,

may have been more satisfied with their body shape than girls.

Behavioral-Attitudinal Measures. Statistically significant ANCOVA main

effects for condition at posttest were also observed on the ChEAT,

F(1,460)=5.56, ~c.02, ChEAT factor 1, E(1,460)=8.07, ~e.01, and Dieting -

measures, E(1,430)=6.09, gc.05. Tables 9, 1 1, 18 display these ANCOVAs.

Mean ChEAT and ChEAT factor 1 scores are presented in Tables 1 0 and 12.

ChEAT. For the ChEAT total score (see Table 9 for the complete

ANCOVA), treatment group means (i&l=6.88) decreased more than the

comparison group (M=8.33) means at posttest cornpared to pretest. Mean

pretest-posttest change for the treatment group (1.78) was substantially higher

than for the comparison group (.53). Students receiving the intervention

demonstrated improvement in their attitudes and behavior related to disordered

eating. These gains were not observed in the comparison group classes. Upon

further analysis, this ChEAT effect appears to be mainly associated with a strong

difference between treatment (M=3.12) and comparison (M=4.10) group means

on Factor 1 (dieting subscale). Mean pretest-posttest changes for the treatment

students (1.32), again, was higher than for the comparison students (.34).

ChEAT scores above the cutoff value (2 20), which is suggestive of

anorexia nervosa, were observed in 7.44% (n=41) of total respondents during

pretesting. Whereas this value dropped from 8.01 % (r~=25) at pretest to 4.1 8%

(n=12) at posttest in the treatment group, the percentage remained relatively


stable in the comparison group: 6.69% (t~=16) at pretest and 6.90% (r~=16) at

posttest. It could be hypothesized that the number of students presenting with

more serious disordered eating-related behaviors and attitudes decreased as a

result of the prevention program. Table 17 shows the percentage of students

scoring above the ChEAT cutoff value by condition and sex.

Dieting. When asked, "Are you currently trying to lose weight?" (dieting

measure), a smaller percentage of children in the treatment condition (23.6%)

reported that they were trying to lose weight at posttest than children in the

comparison condition (36.9%). At pretest both groups reported high levels of

dieting behaviors: 31 .O% of the treatment and 33.9% of cornparison students

answered that they were trying to lose weight. This decrease in the treatment

group was statistically significant, E (1,430)=6.09, p.05. This suggests that

boys and girls receiving the intervention may have reduced their dieting

behaviors during the program, and that this effect was not noticeable in students

who did not receive the program. See Table 19 for the percentages, broken

down by grade and sex.

Gaining. Gaining data show that the percentage of boys who reported

trying to gain weight decreased in the treatment conditions from baseline to

posttest. This finding , however, was not statistically significant. These values

are displayed in Table 21.

Main effects for condition and al1 interaction effects on the BSM, ChEAT

subscale 2 (bulirnia and food preoccupation), ChEAT subscale 3 (oral control),

and gaining failed to reach statistical significance. Mean subscale scores, by


condition and sex are presented in Tables 8, 14, 16 and 21. It may be important

to note that there was a floor effect on the gaining, Cheat factor 2, and ChEAT

factor 3 measures. That is, the scores fell within a lirnited range at the extreme

low end of these scales. This may have considerably decreased their sensitivity

and limited the potential for revealing any statistically significant findings.

Methods of losing weight. Among the 193 students in all pretest

conditions (Le., extended sample) reporting that they were trying to lose weight,

80.8% listed exercising, 66.9% cutting out snacks, 36.8% cutting down calories,

13.5% other, 1 2.4% skipping meals, 10.9% fasting, and 2.1 % listed vomiting as

dieting strategies. Refer to Table 23 for percentages of methods for boys and

girls broken down by grade.

Correlations. A large correlation (Guilford, 1956) was observed between

the pretest BES questionnaire and pretest BSM item (1=.75, ~=525, ~c.005).

The two body-esteem measures appear to be highly related. Moderate

correlations (Guilford, 1956) were noted between pretest dieting and pretest

ChEAT factor 1 @=.-53, r~=531), posttest BES and posttest BSM (1=.69, ~ 4 9 4 ) ~

posttest dieting and posttest ChEAT factor 1 (1=-.55, ~=496) , pretest ChEAT

overall score and pretest BES (1=.45, rp529), and pretest ChEAT factor 1 and

pretest BES (1=.53, n=529). All the correlation coefficients are statistically

significant at the gc.005 level. No large or moderate correlations were noted

between the BMI (body mass index) measure and any of the other dependent

measures. Only small negative correlations between pretest BMI and pretest

BES ( L=-.26), and pretest dieting (p.40) were observed. This suggests that


the students reported height:weight ratios were not highly related to any of the

other dependent measures.

Corrected Analysis (Posttest Scores)

The nature of the design of this study, incomplete nesting of subjects

within classes, presented a potential probfem. There may have been some

dependence within classes (Le., a teacher effect). To examine the possibility of

a differential teacher main effect, the E-values for teacher for each of the

dependent measures were analyzed. A significant effect for teacher was found,

necessitating further analyses (Le., making a conservative statistical correction).

This teacher effect implied that results from students differed significantly more

between classes (different teacher) than within classes (same teacher). In other

words, different results were observed for students with different teachers. This

was possibly due to any number of factors such as different teaching styles,

different levels of motivation, or unique personality characteristics of each

teacher. Whatever the differences, the teacher effect had a significant impact on

the students. Consequently, a re-analysis was perforrned. The procedure

entailed calculating a within-group mean-square error terni (MSE-) for each of

the dependent variables (classes within condition) and substituting this value for

the total within group error tem (MSEtd). This resulted in a corrected F-value.

The corrected F-values yielded the same statistically significant effects as

the uncorrected F-values (Refer to Table 2 for the complete listing of corrected

and uncorrected F-values for the pre-posttest overall analysis). The results and


concfusions did not change following the re-analysis, suggesting that the overall

statistically significant findings at posttest were robust. That teachers differ in

their teaching styles, implementation of lessons and programs, and impact on

students, is a common sense notion. Even taking this factor into consideration,

though, the present intervention remained effective. This further strengthens the

argument favoring the effectiveness of this prevention program and adds to the

robustness and generalizability of the findings to the real world of unique

teachers teaching within different schools.

Analysis of Follow-Up Scores (Uncorrected)

Follow-up analyses entailed cornparisons between means on the follow-

up questionnaires from four of the original treatment classes and three of the

original comparison classes (Refer to Figures 6-1 0). As in the pre-post analysis,

univariate 2 (condition: treatment versus comparison group) X 2 (sex: male

versus female) Analysis of Covariance (ANCOVA) designs were calculated for

each of the dependent measures. Pretest means and students' age were used

as covariates. In al1 cases, with the single exception of the covariate on the

ChEAT factor 2 subscale, the pretest scores were related to the dependent

measures. The follow-up ANCOVA analyses yielded no statistically significant

interactions.

A statistically significant main effect for condition, at follow-up, was found

on the Cognitive measure (KQ), E(1, 142)=45.60, ~c.0001. On the BSM, the

main effect for sex was significant, F(1, 144)=5.30, p.05. The main effect for


condition on the ChEAT factor 1 subscale approached, but did not reach

significance, E(1, 145)=4.50, gc.05.

Coanitive Measure - Knowledge Questionnaire. On the Cognitive

measure, as found in the posttest analysis, treatment group students scored

higher at follow-up (hJt5.73) than cornparison students (b3.81). The treatment

students' gains in knowledge about eating disorders, dieting, and nutrition

appear to have been relatively stable six months after the program's end. See

Table 25 for the complete ANCOVA breakdown.

Affective Measure - BSM. At follow-up, group means for treatment and

comparison males (&4.28) on the BSM were higher than for females (M=3.74),

suggesting that boys in this subsample were more satisfied with their physical

appearance than girls. The BSM follow-up ANCOVA is shown in Table27.

All other follow-up ANCOVA effects failed to reach statistical significance

(see Tables 26, 28-31). Further probing, however, revealed some tendencies in

positive directions for the BES, BSM, and ChEAT in the follow-up data. This

suggests that although not as strong as at posttest, some of the follow-up mean

scores in the treatment condition remained higher than mean scores in the

comparison condition. These tendencies are visually noticeable when

comparing the overall mean scores at pretest, posttest, and follow-up.

Behavioral-Attitudinal Measures - ChEAT. ChEAT scores above the

cutoff value (220) for the follow-up data sample decreased from 6.7% at pretest,

to 2.0% at posttest, and to 3.0% at follow-up in the treatment group.

Cornparison students scores remained relatively stable, increasing from 2.8% at


pretest to 4.1 % at posttest, and then decreasing only slightly to 3.2% at follow-

up. It appears, therefore, that the number of treatment group students scoring in

the anorectic range on the ChEAT decreased after the program and remained

low six months later. See Table 17.

Dieting and Gaining. As the follow-up ANCOVA F-values for condition

on the dieting and gaining measures did not reach statistical significance, it

appears that the number of students in both treatment and comparison groups

indicating that they were trying to lose or gain weight six months after the

program retumed to pretest levels.

Corrected Analysis (Follow-Up)

The same potential problem of a teacher effect, as mentioned above,

applied for the follow-up data. Therefore, the same statistical correction was

computed. Please refer to Table 24 for the corrected and uncorrected follow-up

F-values. Due to the decreased degrees of freedom, and subsequent decrease

in statistical power, the corrected values reduced the strength of initial,

uncorrected findings. In particular, as denoted in Table 24 the error t e m for the

KQ ANCOVA changed from p<.0001 to pc.05. Aside from the reduction in

strength, the results and conclusions did not change following this re-analysis.

Effect Size Analysis

Clinical significance of the intewention's effectiveness was assessed by

calculating mean effect sizes, using the formula for omega squared (G; Keppel,

1982). Tabulations were carried out for the pretest-posttest, and pretest-follow-

up samples.


Posttest. At posttest a large effect (Keppel, 1982) for condition was

found on the Cognitive rneasure (KQ; $=.12), adding further support to the

hypothesis that the treatment group participants outperformed their cornparison

counterparts in knowledge about eating disorders, dieting, and nutrition after the

program.

Small effect sizes (Keppel, 1982) were tabulated for the questionnaires

contained in the Affective component (BES, &.02; BSM &.01). This

indicates that the treatment group students were somewhat more satisfied with

their body shape and had higher body-esteem than the comparison students at

posttest. The Behavioral-Attitudinal component of the present study assessed

by the overall ChEAT score, however, did not show a strong mean effect size.

Follow-Up. At follow-up, cornparisons of the four treatment classes and

three comparison classes yielded a large mean effect size for condition on the

Cognitive measure (KQ, $=.34). This further affirms the earlier assertion that

the gains in knowledge achieved by the treatment students at posttest endured

six months after the program's concfusion.

Overall mean effect sires at follow-up, for condition, on the BES, BSM,

and ChEAT were not large. Mean differences between the comparison and

treatment groups on these measures may not have been clinically significant.

On the ChEAT subscales, small to medium mean effect sizes were

calculated for the dieting subscale (factor 1, &.03, condition X sex interaction),

bulimia subscale (factor 2, $=.027) and oral control subscale (factor 3,


&.01), condition X sex interaction. On ChEAT factor 1, although mean scores

improved for al1 treatment students and not for comparison students, these gains

were substantially greater for females. The effects on ChEAT factors 2 and 3

were reverse to what was expected. This is difficult to interpret (see discussion).

Mean scores for the comparison students on these scales appeared lower (i.e.,

as more adaptive) than for the treatment students.

Baseline-Group Analysis

Analysis of Variances were conducted for the two baseline classes,

comparing the pre-pretest and pretest scores. No main effects for condition nor

interactions reached statisticat significance. This implies that the scores on al1 of

the primary dependent measures (KQ, BES, BSM, and ChEAT) did not change

from pre-pretest to pretest, and further strengthens the conclusions that the

observed positive changes were related to the prevention program, and not

simply due to the passage of time.

High Risk Group Analysis

Post hoc ANCOVAs were conducted on the High Risk subsample.

Statistically significant main effects for condition at posttest were observed on

the Cognitive measure, E(1,32)=19.03, e<.0001, the BES, E(1,29)=l3.2lI

g=.001, and the BM1, E(l , t O)=l2.44, gc.01. Although this sarnple was relatively

small, it appears that high risk treatment subjects scored significantly higher than

their counterparts on the Knowledge Questionnaire. They also reported higher

levels of body-esteem and had a higher mean posttest weightheight ratio

compared to pretest.


Statistically significant main effects were also found for Sex on the BES,

F(1,29)=11.59, gc.01, and BSM, fl1,29)=9.64,pc005. This indicates higher - reported mean scores for males compared to females on rneasures of body-

esteem and body satisfaction. AI1 other main and interaction effects did not

reach statisticai significance. ANCOVAs for the High Risk sample are shown in

Tables 32 through 38. For pretest and posttest mean scores refer to Figures 11-

15.


DISCUSSION

The high prevalence of disordered eating, body shape dissatisfaction,

dieting, and maladaptive eating attitudes and behaviors is widely accepted and

undisputed by most researchers and clinicians in the field (Bruch, 1978; Hoek,

1993; Garfinkel & Garner, 1982). Revelations that these potentially harmful

phenomena are becoming increasingly cornmon and recognized among younger

children is disturbing, and for sorne, difficult to believe.

Accompanying the expanding literature that exposes the high prevalence

of eating disordered behaviors and attitudes, are authors calling for the

establishment of preventative measures to avert or at least reduce the incidence

of these problerns in the general population (Crisp, 1988; Garner, 1985;

Shisslak & Crago, 1987). While many authors write persuasively in favor of

prevention programs, only a srnall nurnber actually follow up and implement their

recommendations in a comprehensive and systernatic manner (Killen et al.,

1993; Richman, 1993). The present study attempts to begin to fiIl this void

through the development and implementation of a primary prevention and

education program that was carried out with grade five, six, and seven classes.

Based on the findings of this study as well as informa1 comments made by

teachers, students, and parents, the present disordered eating prevention and

education program appears largely successful at inducing short-term gains in

participating children. These positive changes were apparent across Cognitive,

Affective, and Attitudinal-Behavioral domains. Longer-lasting changes, at follow-

up, seem more modest, although still promising. The present study illustrates


that not only are intetventions feasible and practical, but they may greatly benefit

preadolescents. Large scale application of prevention programs, in general, and

disordered eating programs, in particular, may aid in reducing maladaptive

behaviors and attitudes in the whole of society, avoiding considerable pain and

suffering. .

The questionnaire data suggest that the treatment group students'

behavior and attitudes shifted in a positive direction on al1 of the primary

dependent measures at posttest. These tendencies were not observed, to the

same significant degree, in the comparison group over the concurrent 3-month

period, or at follow-up. It can therefore be inferred that the positive changes

were affected, to a large extent, by the prevention program. At follow-up, six

months after the intervention's conclusion, these positive shifts were maintained,

although with the exception of the Cognitive component and ChEAT factor 1,

they were not statistically significant. Cornments on this pattern follow in later

paragraphs.

Statistically significant differences between the treatment and comparison

groups were found across the Cognitive, Affective, and Behavioral-Attitudinal

domains. Specifically, main effects for condition were obsenied for the KQ, BES,

ChEAT, and dieting ANCOVAs at posttest. Many of these changes also appear

to be clinically significant, as deterrnined by the effect-size measures.

Data from the Cognitive rneasure (KQ) suggest that the program

participants' knowledge about disordered eating, dieting, and nutrition, increased

considerably more than students not receiving the program. Most of the


students knew little or were misinfomed about this material during the pretest.

The program provided participants with accurate information, and this appeared

to have been leamed during the sessions and retained at posttest in the

treatment group as a whole. This material appears to have been largely retained

at follow-up, six months after the program's conclusion.

Participating teachers, who led the present intervention, also tried to raise

their student's self-esteem, in particular their self-body-esteem, and tried to make

their students more accepting of diverse body shapes and sizes with self and

others. The BSM and BES scores afforded an assessrnent of this segment of

the program (Affective component), and the results suggest that participants'

body self-image was enhanced during the prograrn. At posttest, the students in

the intervention group reported feeling better about their own bodies than did

their comparison counterparts.

Posttest changes in the Affective-Behavioral dornain were supported by

the ChEAT total score, ChEAT factor 1 score, and dieting measure. For

students receiving the intervention, their attitudes and behavior related to

disordered eating improved. These gains were not obsenred in the comparison

group classes. Furthemore, the ChEAT factor 1 and dieting data offer cross-

validated support for the program's specific short-term effectiveness at reducing

the percentage of reported weight loss behaviors in the participants.

Dieting and gaining data show that the percentage of girls who reported

trying to lose weight, and boys who reported trying to gain weight decreased

considerably in the treatment conditions from baseline to posttest. The dieting


and gaining rneasures represent an important element and behavioral test of this

intervention, as most other disordered eating programs and program-evaluations

have not included a behavioral component and have focused pnmarily on gains

in knowledge. Unfortunately, the dieting and gaining effects were not evident at

follow-up. The number of students reporting that they were trying to lose (and

gain) weight at follow-up approximated pretest levels. It appears that weight loss

and gain rnay be a particularly problematic area for children and may be difficult

to pemianently alter with a relatively bnef (non-intensive) intervention. Perhaps

modifications in the curriculum's content, frequency, or duration may improve the

stability of the original positive changes. The percentage of children scoring

above the ChEAT cutoff score (120) decreased considerably in the treatment

condition, while remaining relatively constant in the comparison group. This

pattern implies that the present program may be effective at reducing the

severity of disordered eating behaviors and attitudes in high risk children. This

effect was observed at posttest and, as or more importantly, also at the 6-month

follow-up.

Further analyses with high risk children showed that their scores showed

statistically significant improvement on the Cognitive (KQ) and Affective (BES

and BSM) measures at posttest. High risk students appeared to gain more

knowledge and reported feeling better about their bodies than comparison group

students after the intervention's conclusion. Tendencies in the positive direction

for these children were also observed on the ChEAT and dieting questionnaires.

This suggests that students receiving the treatment reported adopting less


rnaladaptive attitudes and behaviors toward disordered eating, and as a group

reported a reduction in weight-los behaviors. A tentative and cautious

conclusion may be that the present intervention is effective at reducing

disordered eating attitudes, behaviors, and dieting among hig h risk children.

Further systematic research with this special group is needed. In this study the

sample size was too small to draw any definitive conclusions. Research with a

larger number of participants would be helpful in substantiating these daims.

The Analyses of Covariance for the ChEAT subscales 2 and 3, and

gaining measures yielded effects for condition that were not statistically

significant at posttest. On the ChEAT subscales this rnay have been due to a

floor effect. The majority of children endorsed relativeiy few of the items that

load on these two subscales, resulting in low scores. Consequently, there was

not much room for change or irnprovernent on these variables. These factors

may be more suitable for an older population or extremely high Rsk individuals,

and since the majority of the program participants did not fall into this category,

these two scales may not have been appropriate for measuring change.

The correlations between the BES and BSM were large at pretest and

posttest. Both measure body-esteem, and since the two measures are highly

related it may be practical and more efficient to consider administering the one-

item BSM, especially when time or resources are limited. The BSM appears to

offer a quick assessrnent of self-perceived body-esteem. Moreover, the high

correlation between the previously researched and validated BES and the BSM

lends support for the construct validity of the BSM.


Dernographic data showed that this study's sample is comparable to other

Canadian and American student populations. The percentages of young

children who are trying to Iose weight and score high on the ChEAT are

disturbingly high. This finding is consistent with other recent surveys (Childress,

1991 ; Maloney et al., 1989; Rosen & Gross, 1987) and provides further impetus

for the need for more direct preventative action at this young age.

Follow-up results were strong in the Cognitive (KQ) area. Students

appeared to retain most of the information that they leamed during the program,

when retested six months after its conclusion. Gains made in body-esteern (BES

& BSM) also persevered, although not to a statistically significant degree, six

months after the program's end. Both boys and girls, unfortunately, appeared to

have reverted to high levei of trying to Iose weight at follow-up. Modifications to

the program or booster sessions could emphasize the program's anti-dieting

message and review lessons specifically geared towards dieting-reducing

behaviors (e.g., Iesson three). More will be said about this later.

Treatment group means (M=18.69) were higher than the comparison group

(M=16.98) at follow-up on the BES. The same tendency appeared on the BSM

(treatment M=4.08; comparison M=3.84). These mean values imply that the

positive tendency noted at posttest for the Affective component remained

somewhat stable, even if it is not statistically significant at follow-up.

Mean ChEAT scores at follow-up were Iower for the treatment group

(M=6.15) than comparison group (&7.62), suggesting improved attitudes and

behaviors toward eating, food, and dieting and less disordered eating-related


attitudes and behaviors at follow-up in the treatment students. The failure to

reach statistical significance for the ChEAT and ChEAT subscales rnay be

related to the relatively high error variance in conjunction with a smaller sample

size. Since the mean follow-up ChEAT scores were comparable to the mean

pretest-posttest scores for the entire sample, and the latter were statistically

significant, it rnay be that increasing the sample size rnay enhance the statistical

values for these follow-up findings. Regardless of the lack of statistical

significance on the ChEAT, the mean change scores, however, do look

impressive. In this regard, the effect size values rnay be more practically and

clinically informative.

The number and percentage of high risk students (i.e., respondents

scoring 120 on ChEAT) decreased in the intemention group while increasing

slightly in the cornparison group. This finding seems promising, but must be

interpreted with caution, considering the small number of individuals classified

under high-risk status. It seems reasonable that the present intervention, in

addition to benefiting the average, low-risk child, rnay directly prevent or

substantially reduce the risk of developing a severe eating disorder in the small

group of highly susceptible (high isk) children. Further studies rnay choose to

specifically target this sample, using the present prograrn. variations of the

present program, or other similar prograrns. Killen et al. (1 993) favor targeting

high risk students with their particufar disordered eating prevention program.

On the gaining measure (are you trying to lose weight?) the statistical

and clinical analyses at posttest and follow-up rnay have been limited by a floor


effect for girls. Virtually no girls indicated that they were trying to gain weight at

al1 three testing times. This low level of affirmative responses rnay have diluted

the statistical values for the overall sample. More useful information rnay be

gamered by studying only boys on this question. It would also have been

interesting to probe further into the methods used by boys to gain weight. For

example, are the male students using steroids? If yes, is the present

intervention possibly (indirectly) reducing steroid use in males. It is likely that the

lessons on body-esteem and healthy eating may have inadvertently positively

influenced this subgroup of boys. Questions probing the methods used for

gaining weight such as: "are you trying to gain weight by: eating more food;

eating more protein; weight lifting; taking steroids, etc ..." could be added to the

dependent measures.

With the data from the follow-up students, there were no statistically

significant changes on many of the dependent measures six months after the

program ended. Perhaps even more importantly and surprisingly, this same lack

of ANCOVA effects for condition was also observed at posttest for the folfow-up

sub-sample. This trend was noted for al1 the variables, except for the KQ. It

would be unreasonable, in my opinion, to expect statistically significant changes

at follow-up in these children, when the F-values did not reach significance

irnmediately following the program (i.e., at posttest).

Why, then, were the posttest results on most of the dependents

measu res, not statistically sig nificant for this particular su b-sam ple? Are these

students different from the overall sample? This question is difficult to answer.


There was no a priori reason to expect any differences between these classes

and the other non-follow-up groups. Observations of the mean follow-up values

indicate that these treatment-group students did manifest gains in a positive

direction (see Figures 11 -1 5). Perhaps there were meaningful changes at follow-

up, and as mentioned earlier the srnaller sample size reduced the relative

strength of the statistical values. It is possible that the mean changes were not

statistically significant because they were diluted by high error values in

conjunctian with srnaller sample sizes. In other words, there may have been a

lack of statistical power.

When analyzing the effect sizes for the follow-up groups, some of the

values (especially the interactions) were small, but possibly clinically significant.

Therefore, there may be some validity in arguing that this program's long term

effects were noticeable, although to a lesser degree than at posttest. A

consenrative interpretation would be that there may have been some long-term

changes, but it is difficult to reach any definitive conclusions with the present

data. Further research on the long-term effectiveness of this program rnay be

necessary.

Whether or not the initial, post-program, posttest changes were

maintained at follow-up, booster sessions (i.e., additional review-lessons every

few months) may be a valuable future addition to the program. Based on the

mixed findings at follow-up, occasional reviews of the initial lessons by teachers

could be warranted. This may be beneficial for strengthening the impact of the

initial prevention material.


Even if the follow-up findings would have been stronger (i.e.,

unequivocally statistically and clinicaliy significant), booster sessions rnay

nevertheless be helpful. As discussed in the introductoty section, disordered

eating interventions and similar programs are swimming against a strong social

current. That is, various sociocukural and familial pressures will continue to

impact and pressure children and adolescents long after the end of the

intervention. It rnay be unreasonable and unrealistic to expect that a short, six-

lesson, program would be sufficient to alter children's views and behaviors for

the rest of their lives. It seems likely that interventions similar to the one used in

this study can make a significant short-term difference and impact in participant's

lives. The reality is that children are continually battling television, newspaper,

magazines, family, and peer pressures to be thin and diet. Children need to be

re-educated, "de-" and 'Ire-programrned" in order to change deep-rooted, harmful

myths and attitudes they may hold toward weight, body-size, and dieting.

Booster sessions may be helpful in accomplishing this goal.

As discussed in the results section, there waç some variability in the

delivery of the intervention (i.e., a significant teacher effect). In one school,

students in five classes experienced the 6-lesson program in two parts separated

by a one month break in the middle. This, in addition to each lesson being

taught by a different teacher, contributed to a feeling of lack of continuity and

integration among the participants. In other schools, although uniformity in

running the lessons was encouraged by the researcher, some of the teachers

decided to add their own minor modifications to the program. For example, one


teacher added an exva video about healthy eating styles, one teacher read a

book about anorexia nervosa to her class, and another teacher had students

draw pictures of themselves. As mentioned earlier, the duration and frequency

of the lessons varied between classes. The program's overall length ranged

from two weeks to more than three months.

These minor modifications and differences between class effectively

reduce the standardization in the actual implementation of the disordered eating

intervention. This rneans that the treatment program -- by varying in duration,

frequency, and content - was not identically applied by teachers in the

classroom. lnstead of causing difficulties, this phenornenon may add to the

robustness of the findings. The variability in administration may actually

strengthen the external validity of the study by making it more similar to the real

world of school, which is not perfect and involves delays and variations. Dunng

the school year there are often frequent delays, and unexpected circumstances

often anse which disrupt the continuity and flow of ongoing classroom projects.

The strength of the posttest results, and to a tesser degree the follow-up results,

in combination with the variations in the running of the program appears to

indicate that the intervention may potentially be effective in a variety of school

situations. Effectiveness of this prevention program does not seem to depend

on strict adherence to a particular time schedule or a particular spacing of the

lessons. In fact, the intervention may facilitate valuable changes in the

participants when the lessons are taught anywhere from a two week to a three

month span. Moreover, the program seems useful when taught entirely by

Preventing Disordered Eating 7 ï

individual teachers, with different teachers for each lesson, and with minor

changes and additions to its content.

Based upon observations made by the researcher, the teachers' level of

motivation and enthusiasm for the intervention, although generally quite high,

also appeared to Vary frorn class to class. It seems plausible that higher levels

of teacher motivation and enthusiasm may enhance the program's effectiveness.

In observing that most of the participants benefited from the program, this factor

does not appear to be a necessary ingredient for change. Further research

could specifically and systematically manipulate and study the relationship

between teachers' motivation and outcorne. As mentioned before, rather than

being a problem, these differences lend support to the generalizability of the

findings. If the goal of this study is to evaluate and support the feasibility of

conducting the disordered eating prevention program in real schools with real

teachers, then this variability may enhance the robustness of the present

findings.

Following the completion of the lessons, the author met with the teachers

who led the treatment groups. They were asked to provide feedback, criticisms,

and recommendations. In general the feedback was positive, optimistic and

enthusiastic. Many of the teachers were pleased to havs taken part in the

project and expressed their appreciation to the researcher. The teachers also

seemed satisfied that their efforts had an impact on their students. Due to

personal contact with friends, family, and their own battles with eating disorders


and dieting behaviors, the teachers, in general, felt that the intervention was

necessary and worthwhile.

Feedback from the children, at follow-up, was generally very positive.

They responded with great enthusiasm and were generally appreciative of the

opportunity to take part in the disordered eating program. After the intervention

ended, when directly asked by the researcher what they thought of the lessons,

many of the students said that they had leamed a lot, and that the lessons were

fun. It was also clear and apparent by the content of their responses, that the

participants had retained a substantial amount of the material. The researcher

and teachers noticed that many of the children became excited each time they

were tald that it was the time for the prevention prograiri. It was clear that many

also often looked foward to the lessons as an enjoyable change from their other

classroom material.

The students and teachers appeared to have been interested and highly

involved in the sessions. For example, during some weeks a few of the teachers

ended up spending extra hours of class time per week on the program on top of

the scheduled one to one-and-a-half hours. This was in response to the high

level of discussion and interest that was generated by the students.

The intervention's success was also indirectly inferred by positive parental

comments to the teachers. According to some parents, their children began

watching television more critically during and after the program, frequently

commenting on the misinformation presented in weight-loss advertisements.


These students seemed to be more aware of the sociocultural pressures placed

on them to diet and to achieve an unrealistically thin body shape.

During the post-program meetings between the author and the treatment

condition teachers, they were generally pleased with the program. Most of the

teachers indicated that they would be willing to run it again in their classrooms in

the future.

Analysis of the two pre-pretest classes contributes further support for the

success of the intervention. Upon observing the stability in students' scores on

al1 of the dependent measures from the pre-pretest to pretest, it appears that the

later, posttest changes in the treatment groups were not due to the passage of

time or due to exposure to the pretest. The observed changes in the treatment

condition appear to be related to the prevention program, and this test allows

more confidence in ruling out alternative hypotheses.

Three additional possible alternative explanations for the results are:

novelty effect, influence of the primary researcher on the children, and the

unidimensional modality of dependent measures. The first two will be discussed

here, and the latter is included in the next section on future revisions.

The treatment students' changes on the dependent measures could

possibly be explained by a novelty effect. This phenornenon is when apparent

change is caused by the newness, and hence salience, of an event rather than

the event itself. When the novelty wears off, the apparent changes fade. In

reference to the present study, the effects could potentially be associated with

the novelty of the disordered eating program in the participants' lives, rather than


the content of or the intervention itself. In this scenario, once the newness wears

off with the passage of time, the material learned and new attitudes may

disappear and the children may revert to their older, more familiar, thoughts,

attitudes, and habitua1 behaviors. This explanation could be explored though

researching the booster sessions. Theoretically, booster sessions would be

antithetical to this novelty effect. If the initial changes were further enhanced or

at least maintained over the long-term with the benefit of booster sessions, then

this would probably not be due to the program's novelty as this initial period

would fade with time. A more thorough follow-up analysis rnay also elucidate

this potential theory.

The primary researcher attended the pretest, posttest, follow-up, and sat

in on occasional classes. This was done, in most cases, as an observer trying to

be inconspicuous. Any visitor to the classroom, however, cannot be totally

inconspicuous, and may somehow influence the children's ciassroom behavior.

In a few of the classes, the researcher played a more active role, acting as the

psychologist and disordered eating expert, answering questions from the

students and teachers that the teachers felt unable or unqualified to answer.

There was certainly no question that some of the students were trying to impress

their visitor and show-off what they had learned. For example, some of the

students would try extra hard to rnake comments and answer questions when

the researcher was in their classroom.

It is conceivable that the researcher's presence may have influenced the

independent variable (Le., the students' experiences of the lessons) and


confounded the relative purity of the intervention. If so, this effect was most

Iikely minimal and not the key element affecting the measured changes. The

researcher, however, did not attend the majority of the lessons for some of the

intervention-group classes. There appeared to be no strong relationship

behveen the researcher's attendance and the program's effectiveness.

Limitations and Future Revisions

The following section includes further comments and criticisms about the

present intervention, and future revisions are suggested. First, the

unidimensional, self-report nature of the questionnaires is covered. Second, the

equivocal past findings in disordered eating prevention research are discussed.

Third, including the parents as active participants is considered. Fourth,

analyzing the content of specific lessons is recommended. Fifth, a brief look is

taken at the appropriate age of participants for future programs. Last, the

scheduled frequency and spacing of the lessons is discussed.

The questionnaires were all self-report measures, and no attempt was

made to gather quantitative information directly from the students' parents or

teachers. This method was intentionally chosen to maximize efficiency and

minimize the cost, time, and resources needed for the present research project.

Relying on unidimensional measurement modalities, however, may hinder the

validity and confidence of the results. It is conceivable that the respondents

intentionally or unintentionally biased their answers in the direction of socially

favorable results. The students may have tried to answer according to their

Preventincr Disordered Eatina 82

teacher's and researchets expectations. This is a general problem with self-

report measures.

A way to avoid this unidimensionality of assessrnent in future studies

would be to use multiple measurement rnodalities. Time and resources

permitting, including qualitative interviews of the participants, using more direct

behavioral rneasures (properly using the BMI; directly measuring children's

dieting by parents and teachers) would enhance the validity, gereralizability, and

perhaps generate additional informative material about the program and the

participants (see Piran, 1995; and Shisslak & Crago, 1993) This would require

considerably more resources than used in the present program, though. Another

technique to circumvent this problem would be to use or develop more

sophisticated and less transparent questionnaires. The newly constructed KEDS

(Childress, Jarrell, & Brewerton, 1993) measure may be considered as a

preferable alternative scale.

Some of the students may have experienced difficulty comprehending the

questionnaires, particularly the grade five and English-as-a-Second-Language

(ESL) children. This was noted even though careful planning was taken to keep

the questions as understandable and simple as possible. More effort and time

could be spent ensuring that students fully understand al1 of the questions and

difficult words during pre- and posttesting. Perhaps for these children, the

teachers and researchers could orally recite each question to small groups of

students, asking if everyone understands the overall meaning of each of the

words.


As mentioned in the introduction, past eating disorder prevention research

has yielded seemingly equivocal and contradictory results. Some programs

apparently succeed and are deemed effective by the researchers, while others

fail to attain the desired changes. One consistent finding across al1 published

interventions is the gains accrued in knowledge. It is rny belief that changes in

children must be facilitated in a wider domain. Gains in knowledge and

information do not necessariiy lead to changes in the equally or more important

areas of affect, attitudes, and behavior. It is here where the programs appear to

have a difiering impact. What are the reasons for these variations in program

effectiveness? Facts frorn Killen et al. (1 993) and the present study will be

raised in explorhg this issue. These two projects are longer-ten, systematically

applied and evaluated studies. Killen et al. (1 993) failed to find meaningful short-

or long-term changes in their treatment group, whereas the present program and

an earlier pilot study (Richman, 1993) suggested a much more positive outcorne.

How can this be explained?

First, the sarnples are different. Killen et al.'s (1 993) participants were

slightly older (grade six and seven) and only females, whereas in the present

project boys and girls from grades five, six, and seven were chosen. Since this

study's significant results generalized across age and among females, this does

not appear to be a satisfactory explanation for the different outcomes.

Second, in ternis of the overall design, in the present study classes were

nonrandomly assigned to either the treatment or cornparison condition. With the

exception of one grade five class that was switched to the treatment condition,


al1 classes were nested within schools. That is, children within a particular

school either received the program or received no program, and there was Iittle

possibility of a spillover (i.e., children from one treatment class comrnunicating to

children to another nontreatment class about the program). Killen et al. (1 993)

randomly assigned classes to conditions. Although perhaps from an overall

design perspective this appears experimentally superior to nonrandom

assignment, it does create separate potential difficulties. In particular, there is a

reasonable possibility that students frorn one class mingled and talked to

children from another class about the lessons during the intervention period. If

this were occurring with any frequency, then the cornparison groups would not

be pure. The nontreatment children may, in fact, have been indirectly exposed

to many of the elements in the prograrn via discussions from friends in other

ciasses. No mention of this potential confound is discussed by Killen et al.

(1993), and therefore at this point it is impossible to totally rule it out. A

manipulation check , by questioning the students from the classes whether they

discussed the program with friends from other classes, may help answer this

question in the future. It is questionable whether random selection of classes is

desirable, or necessary, in this type of research, since the students are usually

not self-selected. In the present study, selection of schools and classes was

done on the basis of availability through contacting school principals and

teachers, not individual students.

Third, dependent measures differed between studies. Killen et al. (1 993),

with more resources, used self-report questionnaires, direct behavioral weight


measures (BMI) and conducted individual interviews. This more thorough multi-

method assessrnent would be desirable in future studies since the intewiew and

behavioral questions offer cross-validation for the self-report questionnaires.

Killen et al. rely heavily on the €DI and stnictured clinical interview (SCID),

where this project's evaluation is based on the Ch €AT. Killen et al.'s selection of

dependent measures brings up certain questions. For instance, why did they

chose the EDI? Was the SClD altered to make it comprehensible for children?

It is questionable whether the EDI and SClD are valid and appropriate measures

for children, and Killen et al. (1993) do not provide data on the validity of these

measures for this young age group. It is possible then, that the questionnaires

chosen were not sufficiently sensitive to (sense) change in the children. Future

studies should use scales that are validated on children.

Fourth, Killen et al.'s (1 993) curriculum had more lessons (1 8) compared

to the present program (6). Nevertheless, the overall length of time spent on the

lessons, may have been comparable. Whereas Killen et al. (1993) relied heavily

on slide presentations, we included much more direct, non-multimedia exercises

and group work. In Our classes students were encouraged to work together in

groups and the teachers facilitated a lot of class discussions. This different

emphasis may have made a key difference. Many of our activities were fun for

the kids (e.g., crossword puzzles, role-playing, etc ...) and this often increased

their level of interest, participation, and enthusiasm. It is unclear what was the

level of involvement and CO-operation amoung the students in Killen et al.'s

(1993) classes. Moreover, their material may have been drier and less


personally relevant for the students. All these factors are extremely important

when the goal is to change behaviors, feelings and attitudes (see Piran, 1995).

Fifth, and last, in analyzing the questionnaire data, an ANCOVA

procedure was used in the present study, controlling for pretest variability and

thereby reducing the error variance. In al1 cases the pretest covariates were

statistically significant, further validating this process. Killen et al. (1 993) chose

ANOVA instead, and this rnay have weakened the statistical power of their tests.

It would be interesting to retest or reanalyze their data using an ANCOVA

procedure and observe any differences. Upon reviewing their mean values,

however, and comparing them to the values obtained in the present study, it

does not appear that using an ANCOVA would substantially alter Killen et al.'s

(1 993) findings.

At this point it is difficult to corne to any definitive conclusions about the

differential effectiveness of the pu blished disordered eat ing and eating disorders

interventions. More information is needed about the content of the prograrns

and the manner in which the lessons were conducted. Knowledge can be

gained by carefully purviewing research from other successful prevention

programs (dmg and alcohol abuse, suicide, smoking). Researchers need to be

clear, accurate, and systernatic when communicating and writing about what

really happened du ring their program. For example, when the classes were run,

did most of the participants pay attention to their teachers? If yes, were the

students excited and enthusiastic about the lessons or were they generally not

interested and bored?


lncluding the parents as participants in the intervention may be beneficial

at further enhancing the effectiveness of prevention programs for disordered

eating (and other prevention programs for that matter). Engaging the parents

would facilitate generalizing the content of the program beyond the school to the

home and family, both which are crucial in the development of the child

throughout preadolescence and beyond. Parent's attitudes, knowledge, and

behaviors greatly influence their children, and hopefully prevention programs

could catalyze them to act as appropriate role models for their children in al1

areas of living in generai, and dieting, weight and eating disorders, in particular.

For example, parents are integrally involved in passing on negative or positive

attitudes regarding the ovemueight to their children. Furthemore, if they often

diet, they convey strong messages toward their children through their behaviors.

On the negative side, it may be difficult to obtain cooperation of the parents.

Many are busy and rnay not be willing to dedicate extra tirne to this effort,

especially if they consider other issues more relevant and pressing for

themselves or their children. Parents rnay not be open-minded and willing to

adopt new attitudes and behaviors as are their preadolescent children.

These challenges aside, engaging the parents may be well worth the

extra effort. If one of the goals of the program is to improve the participants'

eating habits, then it is imperative to gain the cooperation of the parents. Most of

the children's meals will be prepared and chosen by the parents, or altematively

eaten out at restaurant selected by the parents. Again, they could act as positive

role models, preparing and eating healthy, balanced meals for themselves and


their children, or making healthy choices when eating in restaurants. Time and

effort would be a prerequisite for this, but hopefully active parents will recognize

and value the advantages in their offspring. Some parents may actively oppose

the program, especially those that may have an eating disorder, are overweight,

have issues involving food, or are generally extra-sensitive to this area. Based

on the present findings and experiences during this intervention, the number of

complainants would probably not be substantial and most parents and teachers

would support the program.

In the future, a careful analysis of the content of the present program

would be imperative. Procedures such as dismantling (see Kazdin, 1992),

whereby each component of the intervention is evaluated on its own merits,

could allow us to assess the various components of the curriculum separately.

Moreover, improving certain units in the program may be worthwhile. For

example, the self-esteem component (lesson six) is short cornpared to the other

Iessons. More time and exercises could be devoted to boosting participants'

self-esteem, and this couid be evaiuated. Also, maybe more or less time could

be spent on the knowledge and information material. Careful and systernatic

future research rnay provide answers to these and other questions. We could

also look at the teaching methods used convey information and the Iessons.

Suggestions from Piran (1 995) could be incorporated, such as making the

material more personally meaningful to the participants and adding more group

activities. Guest speakers (women with eating disorders, overweight individuals,

psychologists and other experts) could talk to the students.


Assessing the most appropriate age to target these interventions rnay be

another area for future programs. Smolak and Levine (1 994) argue that younger

children are most amenable to these interventions. According to the data

obtained in this study and Richman (1993), and the casual observations of

teachers and students at varying grade levels, grade six appears to be an

optimal time to effectively reach this audience. Grade four children may be too

young to understand many of the lessons' elements and, on the other side,

grade eight adolescents may be too unwilling to M e n as many children become

unnily and disruptive at this transitional age.

Another consideration raised by Shisslak and Crago (1 993) would be

looking at the relative effectiveness of one-shot versus ongoing interventions. It

may be preferable to expose the students to sirnpler versions of this program at

a younger age and repeat variations of the lessons at age-appropriate levels

every year. Or a combination may be the ideal -- whereby grade six would be

the pinnacle year for the intervention and pre-program preparation would begin

much earlier, on a yearly basis and the program would be followed by periodic

post-program booster sessions. Of course this type of setup would rely entirely

on the full cooperation of teachers, school boards, and principals. This scenario

could be the trend of the future in prevention research and program execution.

Summary and Conclusions

The present prevention and education program illustrates that disordered

eating programs are feasible and can be successfully implemented in the school

system. Conducted on grade five, six, and seven classes, participants in the


present intervention improved their knowledge of eating disorders, dieting, and

nutrition information, and decreased their maladaptive attitudes and behaviors

associated with eating disorders. Moreover, children receiving the program

indicated that they were more satisfied with their body shape, and reported less

weight reducing and gaining behaviors than nontreatment comparison groups.

It is recommended that the present intervention be implemented on a

regular basis in the school system. With a larger number of participating

classes, teachers, and schools, a positive dent rnay be made toward decreasing

the incidence and prevalence of disordered eating and related behaviors and

attitudes.

Preventina Disordered Eatina 9 1

QUESTIONNAIRES


Knowledge Questionnaire

*** Please answer al1 of the questions. CIRCLE the letter next to the best response. If you are not sure about an answer, please make your best guess.***

1. People with anorexia nervosa.. .

a) usually think they are too thin b) feel out of control around food c) often dislike the size and shape of their body d) easily become nervous

2. People with bulirnia nervosa...

a) do not eat very much food b) are usually very thin c) are usually oveweight d) often overeat and then try to get rid of the food by vomiting (throwing UP)

3. About out of every 100 people who have eating disorders (anorexia nervosa and bulimia nervosa) are female:

4. Most overweight (fat) persons ...

a) eat more than normal-weight persons b) eat a lot more than normal-weight persons c) eat about the same as normal-weight persons d) eat a lot less than normal-weight persons

5. Which of the following statements is true:

a) thin people are healthier than fat people b) fat people are healthier than thin people c) fat people are just about as healthy as thin people d) none of the above


6. People with bulimia nervosa ...

a) are very thin b) are overweight c) can be any body size d) are normal-weight

7. People with anorexia nervosa. ..

a) are very thin b) are ovenveight c) can be any body size d) are normal-weight

8. Most diets ...

a) do not work (fail) over the long-term (more than one year) b) do not work (fail) over the short-terni (less than one year) c) lead to permanent weight loss (work well) d) are healthy

9. Most overweight (fat) people ...

a) have more emotional problems than normal-weight people b) have less ernotional problems than normal-weight people c) have no more or less emotional problems than normal-weight people d) are happy al1 of the time

10. Which is the most healthy meal?

a) hamburger, french fries, and milkshake b) green salad, vegetable soup, brown rice, and beans c) barbecued chicken, white bread, milk, and carrots d) chocolate ice-cream and seven-up


Body-Esteem Scale

(circle one)

1. I like what I look like in pictures.

2. Kids my own age like my looks.

3. I'm pretty happy about the way I look.

4. Most people have a nicer body than I do.

5. My weight makes me unhappy.

6. 1 like what I see when 1 look in the mirror.

7. 1 wish I were thinner.

8. There are lots of things I'd change about my looks if I could.

9. I'm proud of my body.

10. 1 really like what I weigh.

1 1. I wish I looked better.

12. 1 often feel ashamed of how I look.

13. Other people make fun of the way I look.

14. 1 think I have a good body.

15. I'm looking as nice as I'd like to.

16. It's pretty tough to look like me.

17. 1 wish I were fatter.

18. 1 often wish I looked like someone else.

19. My classmates would like to look like me.

20. 1 have a high opinion about the way I look.

21. My looks upset me.

22. I'm as nice looking as most people.

23. My parents like rny looks.

24. 1 worry about the way I look.


Dieting Questionnaire

1. Are you trying to lose weight? yes no (circle one)

2. If you answered YES to question #1, place a check mark next to the way(s) you are using to lose weight (you may check more than one):

- fasting (not eating solid food for at least 24 hours)

- skipping meals

- cutting out snackdjunk foodkweets

- cutting down calories

- exercising (some forms of exercise are: jogging, aerobics, swimming, walking)

- vomiting (throwing up)

other (please list: 1

3. Are you trying to gain weight? yes no (circle one)


Body Satisfaction Measure:

Please CIRCLE the number which best describes how you feel:

I am very happy With the way

my body looks

I am happy with the way

my body looks

I am somewhat happy

3 with the way my body looks

I am somewhat unhappy

2 with the way my body looks

I am unhappy with the way

my body looks

I am very unhappy with the way

my body looks


APPENDIX B

COVER LEITER 10 SCHOOL PRINCIPALS

Preventina Disordered Eatinci 100

From: Rizfael Richman C/O Dept. of Psychology Simon Fraser University Bumaby, B.C. V5A 1 S6

oct. **, 1993

To: Principal name of school

Dear Mr. principal:

I am a doctoral student in clinical psychology at Simon Fraser University, presently working on my Ph.D. dissertation. The project involves planning and implementing an eating disorders prevention program for grade five and six students. Components of the program include: a blief introduction to eating disorders, looking aï societal pressures to be thin, examining harmful effects and myths about dieting, improving self-esteem, decreasing prejudices and stereotypes surrounding fat people, and promoting healthy eating habits. Some of the lessons nicely complement, and may be integrated with, the province's Leaming For Living prograrn.

Approval for my research has been given by XX at the XX school district offices.

Phase one of this project, a pilot eating disorders prevention program, was conducted in two XX schools (as my masteh thesis). My findings suggested that the program was very successful in inducing rneaningful short- terni and long-term changes. Further to these promising results, I am now looking for grade five and six classes and teachers who would be willing to participate in the continuation of the program. Therefore, I will be following this fax with a phone call, at which tirne I would be happy to answer any questions you may have. I look forward to speaking with you in the near future. Thank you.

Yours truly,

Rafael Richman, M.A.


APPENDIX C

PARENTAL CONSENT LElTERS


Feb. 4,1994

Dear Parents:

Eating disorders, obsessive dieting, and unhealthy eating patterns are a major concem in pre-adolescents. The incidence of eating disorders among children and adolescents is disturbingfy on the increase. In response to this problem, this school has decided to implement an eating disorders education and prevention program. The primary goal is to prevent eating disorders and related problems from developing in the first place. A further aim is to foster inforrned and accurate attitudes toward obesity and dieting.

Beginning in Iate January your child will have an opportunity to participate in the eating disorders education and prevention program. The lessons include: introduction to anorexia nervosa and bulimia nervosa, examination of the harmful effects of and myths surrounding dieting, enhancement of self-esteem, looking at the rnyths of obesity, discussion of the influence of family and friends on dieting and weight, and suggestions for heaithy eating habits.

Before and after the prograrn your child will be asked to complete some questionnaires on eating attitudes, general knowledge of eating disorders, body esteem, and dieting. The results will be kept confidential (i.e., securely stored), anonymous (i.e., your child's name will not be revealed), and used for research purposes, as part of a psychology Ph.D. dissertation conducted at Simon Fraser University. Your child will be infomed that he/she is allowed to withdraw from the questionnaire component of the program at any tirne.

The project will be coordinated by Rafael Richman and supervised by Dr. Ron Roesch, Professor in the Department of Psychology at Simon Fraser University (tel. # 291 -3370). A copy of the results, when cornpleted, will be availabte by contacting Rafael Richman c/o the Department of Psychology (tel. # 29 1 -3354).

If you have any questions, concems, or if you object to your child participating in the questionnaire component of the program, please contact the XX School at ##. Thank you.


Feb. 4, 1994

Dear Parents:

As part of a psychology Ph.D. dissertation conducted at Simon Fraser University on eating disorders, your child will be asked to complete some questionnaires on eating attitudes, general knowledge of eating disorders, personal eating patterns, and dieting. The results will be kept confidential (i.e., securely stored), anonymous (Le., your child's name will not be revealed), and used for research purposes. Your child will be inforrned that he/she is allowed to withdraw from the questionnaire component of the program at any time.

The project will be conducted by Rafael Richman and supervised by Dr. Ron Roesch, Professor in the Department of Psychology at Simon Fraser University (tel. # 291 -3370). A copy of the results, when completed, will be available by contacting Rafael Richman c/o the Department of Psychology (tel. # 29 1 -3354).

If you have any questions, concerns, or if you object to your child's participation in answering the questionnaires, please contact the XX School at ##. Thank you.


APPENDIX D

STUDENT CONSENT FORMS


You will be taking part in a special program on eating disorders. The goal of the program is to make you aware of the hamful effects of eating disorders and to reduce the risk of future development of unhealthy eating patterns and attitudes.

In addition to answering some questionnaires, the program will include lessons on: eating disorders information, looking at the rnyths of dieting and obesity, self-esteem, talking about healthy eating habits, and how family and friends influence your feelings about your body.

I agree to answer questionnaires for an eating disorders research project run by Mr. Rafael Richman of the Psychology Department at Simon Fraser University. I understand that al1 of my answers will be kept confidential (they will not be shown to anyone). 1 also understand that I may stop answering the questions at any time. If I have any complaints I may speak to my teacher andlor Mr. Richman.

Name Date

Sex: male female

Age:

Height:

Weight:

Grade:

(circle one)


I agree to answer questionnaires for an eating disorders research project nin by Mr. Rafael Richman of the Psychology Department at Simon Fraser University. I understand that al[ of my answers.will be kept confidential (they will not be shown to anyone). I also understand that 1 may stop answering the questions at any time. If I have any cornplaints I may speak to my teacher anaor Mr. Richman.

Name Date

Teacher's name Grade:

Sex: male female (circle one)

Age:

Height:

Weight:


APPENDIX E

FINAL LElTERS TO PARENTS


Dear Parents:

Your child participated in an eating disorders education and prevention program in "Date**. Lessons in the program included: introduction to anorexia nervosa and bulimia nervosa, examination of the harmful effects of and myths surrounding dieting, enhancement of self-esteem, looking at the myths of obesity, discussion of the influence of family and friends on dieting and weight, and suggestions for healthy eating habits.

Before and afier the eating disorders prevention program, questionnaires on eating attitudes and behaviors were completed by your child and hislher classrnates. Based on the answers from these foms, and cornments by the participating teachers and students, the program appeared to be successful. The number of children reporting that they were trying to lose or gain weight decreased during the eating disorders program. Abnormal eating attitudes and behaviors also diminished, and most of the participants were more satisfied with their body shape at the end of the prograrn.

Some of the students, however, reported that they were dieting and were not satisfied with their body shape. Moreover, some of the children's answers on the test of eating attitudes indicated that they have attitudes and behaviors similar to persons suffering from anorexia nervosa and bulimia. Although your child may not necessarily belong in this group, I would encourage you to speak to your child about their attitudes and behaviors related to eating, dieting, and their body satisfaction. If you are concerned and would like further information or services, please contact the Eating Disorder Resource Centre of British Columbia (telephone # 631 -531 3).

Yours truly,

Rafael Richman Dept. of Psychology Simon Fraser University


Dear Parents:

Your child and hisher classrnates participated in the questionnaire component of an eating disorders education and prevention program in **Date**. This involved providing information and answering questions on eating attitudes and behaviors. Based on the answers from these forms, the majority of students appeared to be satisfied with their body shape, satisfied with their weight, and held normal attitudes toward eating.

Some of the students, however, reported that they were dieting and were not satisfied with their body shape. Moreover, some of the children's answers on the test of eating attitudes indicated that they have attitudes and behaviors similar to perçons suffering from anorexia nervosa and bulimia. Although your child may not necessarily belong in this group, I would encourage you to speak to your child about their attitudes and behaviors related to eating, dieting, and their body satisfaction. If you are concerned and would like further information or services, please contact the Eating Disorder Resource Center of British Columbia (telephone #: 631 -531 3).

Yours truly,

Rafael Richman Dept. of Psychology Simon Fraser University


REFERENCES

Albee, G. W. (1 982). Preventing psychopathology and promoting human potential. American Psvcholoaist. 37, 1043-1 050.

American Psychiatric Association (1 994). Diaanostic and statistical manual of mental disorders (DSM-IV)(~~ ed.). Washington, D.C.

Brewerton, T. (1 992, NovemberJDecember). Childhood eating disorders on the rise. East-West Natural Health, p. 17.

Brown, C., & Forgay, D. (1987). An uncertain weil-being: Weight control and self-control. Healthsharina Winter, 1 1-1 5.

Bruch, H. (1978). The golden caae: The eniama of anorexia nervosa. Cambridge, MA: Harvard University Press.

Button, E. J., & Whitehouse, A. (1981). Subclinical anorexia nervosa. PsvchoIoaical Medicine. 1 1, 509-51 6.

Childress, A. C. (1 991, October 22). Study finds many kids worry about weight. Bellinaham Herald.

Childress, A. C., Jarrell, M. P., & Brewerton, T. D. (1993). The Kids' Eating Disorders Survey (KEDS): lntemal consistency, component analysis, and reliability. Eatina Disorders. 1, 123-1 31.

Chng, C. L. (1 983). Anorexia nervosa: Why do some people starve themselves? The Joumal of School Health. 53, 22-26.

Collins, M. E. (1 988). Education for healthy body weight: Helping adolescents balance the cultural pressure for thinness. Journal of School Health. 58, 227-231 .

Cowen, E. L. (1 983). Primary prevention in mental health: Past, present, and future. In R. D. Felner, L. A. Jason, J. N. Moritsugu, & S. S. Farber (eds.) Preventive ~svcholoav: Theoty. research, and ~ractice. New York: Pergamon Press.

Crisp, A. H. (1 979). Early recognition and prevention of anorexia nervosa. Develo~mental Medicine and Child Neuroloav. 21, 393-395.

Crisp, A. H. (1 988). Some possible approaches to prevention of eating and body weightlshape disorders, with particular reference to anorexia nervosa. International Journal of Eatina Disorders, 7, 1-1 7.

Preventing Disordered Eating 1 1 1

Crowther, J. H., Post, G., Zaynor, L. (1 985). The prevalence of bulimia and binge eating in adolescent girls. lntemational Joumal of Eatina Disorders, 4,2942. -

Flay, B. R., Ryan, K. B., Best, A., Brown, K. S., Kersell, M. W., D'Avemas, J. R., & Zanna, M. P.(1985). Are social-psychological smoking prevention programs effective?: The Waterloo study. Joumal of Behavioral Medicine, 8,37-57. -

Garfinkel, P. E., & Gamer, D. M. (1982). Anorexia nervosa: A multidimensional permeciive. New York: BninnerlMazel.

Garner, D. M. (1 985). latrogenesis in anorexia nervosa and bulirnia nervosa. International Joumal of Eatina Disorders. 4, 701 -726.

Garner, D. M., & Garfinkel, P. E. (1979). The Eating Attitudes Test: An index of the symptoms of anorexia nervosa. Psycholoaical Medicine. 9, 273-279.

Garner, D. M., Garfinkel, P. E., & Olmsted, M. P. (1983). An overview of sociocultural factors in the development of anorexia nervosa. In Anorexia nervosa: Recent developments in research (pp. 65-82). New York: Alan R. Liss.

Gamer, D. M., Garfinkel, P. E., Rockert, W., 8 Olmsted, M. P. (1 987). A prospective study of eating disturbances in the ballet. Psvchotheraw and Ps~chosomatics. 48, 1 70-1 75.

Gamer, D. M., Garfinkel, P. E., Schwartz, D., & Thompson, M. (1980). Cultural expectations of thinness in women. Psvcholo ical Reports. 47,483-491

Gamer, D. M., & Olmsted, M. P. (1984). Manual for Eating Disorder lnventorv. Florida: Psychological Assessrnent Resources.

Gamer, D. M., Olmsted, M. P., Bohr, Y. & Garfinkel, P. E. (1982). The Eating Attitudes Test: Psychometric features and clinical correlates. Psvcholoaical Medicine. 12, 871 -878.

Greenfeld, D., Quinlan, D. M., Harding, P., Glass, E., Bliss, A. (1987). Eating behavior in an adolescent population. lntemational Journal of Eating Disorders. 6, 99-1 1 1.

Guilford, J.P. (1956). Fundamental statistics in psycholouv and education. New York: McGraw-Hill.

Gustaisson-Larson, A. M., & Terry, R. D. (1 992). Weight-related behaviors and concems of fourth-grade children. Joumal of the American Dietetic Association. 92, 81 8-822.


Health Services and Promotion Branch: Health and Welfare Canada (1 988). Promotina healthv weiahts: A discussion pa~er. Canada: Minister of National Health and Welfare

Hill, A.J., & Bhatti, R. (1995). Body shape perception and dieting in preadolescent British Asian girls: Links with eating disorders. lntemational Journal of Eatinci Disorders. 17, 175-1 83.

Hoek, H. W. (1 993). Review of the epidemiological studies of eating disorders. International Review of Psvchiatry. 5, 61-74.

Jones, D. J., Fox, M. M., Babibian, H. M., & Hutton, H. E. (1980). Epidemiology of anorexia nervosa in Monroe County, New York: 1960-1 976. Psvchosornatic Medicine. 42, 551 -558.

Kazdin, A.E. (1992). Research desian in clinical psychology (2"' ed.). Toronto: Allyn and Bacon.

Katz, J. L. (1 985). Sorne reflections on the nature of eating disorders: On the need for humility. lntemational Journal of Eatin~ Disorders. 4, 61 6-626.

Kendall, R. E., Hall, D. J., Hailey, A., & Babigian, H. M. (1973). The epidemiology of anorexia nervosa. Psvcholoaical Medicine. 3, 200-203.

Keppel, G. (1 982). Desion and analvsis: A researcher's handbook (2"6 ed.). New Jersey: Prentice-Hall.

Kilboume, J. (1 993). Still killing us softly: Advertising and the obsession with thinness. In P. Fallon, M. A. Katzman, & S. C. Wooley (Eds.), Feminist perspectives on eatina disorders (pp. 395-418). New York: Guilford Press.

Killen, J. D. (1985). Prevention of adolescent tobacco smoking: the social pressure resistance training approach. Journal of Child Psvcholoav and Psvchiatrv. 26, 7-1 5.

Killen, J. D., Taylor, C. B., Telch, M. J., Saylor, K. E., Maron, D. J., & Robinson T. N. (1986). Self-induced vomiting and laxative and diuretic use among teenagers: Precursors of the binge-purge syndrome? Journal of the American Medical Association. 255, 1447-1 449.

Killen, J. D., Taylor, C. B., Hammer, L. D., Litt, I., Wilson, D. M., Rich, T., Hayward, C., Simmonds, B., Kraemer, H., & Varady, A. (1993). An attempt to modify unhealthful eating attitudes and weight regulation practices of young adolescent girls. lntemational Journal of Eatina Disorders. 1 3, 369- 384.

Preventinq Disordered Eating 1 13

Ledoux, S., Choquet, M., & Flament, M. (1 991). Eating disorders among adolescents in an unselected French population. lntemational Joumal of Eatina Disorders. 10, 81 -89.

Leichner, P., Amett, J., Rallo, J. S., Srikameswaran, S., & Vulcano, B. (1986). An epidemiologic study of maladaptive eating attitudes in a Canadian school age population. lntemational Journal of Eatina Disorders. 5, 969- 982.

Levine, M. P. (1 987). How schools can help combat student eatina disorders: Anorexia newosa and bulirnia. Washington: National Education Association.

Luepker, R. V., Johnson, C. A., Murray, D. M., & Pechacek, T. F. (1983). Prevention of cigarette smoking: Three-year follow-up of an education program for youth. Joumal of Behavioral Medicine. 6, 53-62.

Maloney, J. J., McGuire, J., & Daniels, S. R. (1 988). Reliability testing of a children's version of the Eating Attitudes Test. Journal of the American Academv of Child Adolescent Psychiatry. 5, 541 -543.

Maloney, M. J., McGuire, J., Daniels, S. R., & Specker, B. (1989). Dieting behavior and eating attitudes in children. Pediatrics. 84,4820489.

Mann, P. A. (1 978). Community psvcholoav: concepts and applications. New York: The Free Press.

Mazel, J. (1980). The Beverlv Hills diet. New York: Macmillan Publishing.

Mazur, A. (1 986). U.S. trends in ferninine beauty and overadaptation. Joumal of Sex Research, 22,281-303.

Mellin, L. M., Irwin, C. E. Jr., & Scully, S. (1992). Prevalence of disordered eating in girls: A survey of middle-class children. Joumal of the American Dietetic Association. 92, 851 -853.

Mendelson, B.K., & White, D.R. (1 982). Relation between body-esteern and self- esteem of obese and normal children. Perceptual and Motor Skills. 54131, 899-905.

Moreno, A. B., & Thelen, M. H. (1993). A preliminary prevention program for eating disorders in a junior high school population. Joumal of Youth and Adolescence. 22, 109-1 24.

Moriarty, D., Shore, R., & Maxim, N. (1 990). Evaluation of an eating disorder curriculum. Evaluation and Proaram Plannina. 13,407-41 3.


Nylander, 1. (1 971). The feeling of being fat and dieting in a school population: An epidemiologic interview investigation. Acta Sociomedica Scandanavia, 3, 17-26.

Perry, C. L., Killen, J., & Slinkard, L. A. (1980). Peer teaching and smoking prevention among junior high students. Adolescence. 58, 277-281

Piran, N. (1995). Prevention: Can early lessons lead to a delineation of an alternative model?: A critical look at prevention with schoolchildren. Eatinq Disorders, 3, 28-36.

Polivy, J., & Herman, C. P. (1 983). break in^ the diet habit. New York: Basic Books.

Polivy, J., & Herman, C. P. (1987). Diagnosis and treatment of normal eating. Journal of Consultina and Clinical Psvcholoav. 55, 635-644.

Pope, H. G., Hudson, J. f . , Yurgelun-Todd, D. & Hudson, M. S. (1984). Prevalence of anorexia nervosa and bulimia in three student populations. International Joumal of Eatinu Disorders. 3,45-51.

Porter, J. E., Morrell, T. L., & Moriarty, D. (1986). Primary prevention of anorexia nervosa: Evaluation of a pilot project for early and pre-adolescents. CAHPER Journal. 4, 21-26.

Raciti, M. C., & Norcross, J. C. (1987). The EAT and EDI: Screening, interretationships, and psychornetrics. International Journal of Eatinq Disorders, 6, 579-586.

Rappaport, J. (1 977). Cornmunity psvcholoav: - Values, research. and action. New York: Holt, Rinefiart, & Winston.

Rice, C. (1989). Teacher's resource kit. Toronto: National Eating Disorder Information Center.

Richman, H. (1 993). Primarv mevention of eatina disorders. Unpublished master's thesis, Simon Fraser University, Burnaby, British Columbia.

Rosen, J.C. (1 989). Prevention of eating disorders. National Aid Society Newsletter, 12, 1-3.

Rosen, J. C., & Gross, J. (1 987). Prevalence of weight reducing and weight gaining in adolescent girls and boys. Health Psycholoav. 6, 131 -1 47.

Salmons, P. H.,Lewis, V. J., Rogers, P., Gathere, A. J. H., & Booth, D. A. (1988). Body shape dissatisfaction in school children. British Journal of Psvchiatry, 1 53 (suppl. 2L 27-31 .


Sassun, A., Lewin, C., & Roth, D. (1995). Dieting behavior and eating attitudes in Israeli children. lntemational Journal of Eatina Disorders. 17, 67-72.

Schaps, E., DiBartolo, R., Moskowitz, J., Palley, C. S., & Churgin, S. (1981). A review of 127 drug abuse prevention program evaluations. Joumal of Drug Issues. 11, 17-43.

Schwartz, D. M., Thompson, M. G., & Johnson, C. L. (1982). Anorexia nervosa and bulirnia: The socio-cultural context. lntemational Journal of Eating Disorders. 1,2036.

Shisslak, C. M., & Crago, M. (1 987). Primary prevention of eating disorders. Journal of Consultina and Clinical Psvcholoay. 55, 660-667.

Shisslak, C. M., & Crago, M. (1 990). Prevention of eating disorders among adolescents. American Joumal of Health Promotion. 5, 100-1 06.

Shisslak, C. M., & Crago, M. (1 993). Toward a new model for the prevention of eating disorders. In P. Fallon, M. A. Katzman, & S. C. Wooley (Eds.), Feminist ~ers~ect ives on eatina disorders (pp. 41 9-438). New York: Guilford Press.

Silverstein, B., Perdue, L, Peterson, B., & Keliy, E. (1 986). The role of the mass media in promoting a thin standard of bodily attractiveness for women. Sex Roles. 14, 51 9-532.

Silverstein, B., Peterson, B., & Perdue, L. (1986). Some correlates of the thin standard of bodily attractiveness for women. lntemational Joumal of Eating Disorders. 5, 895-905.

Smead, V. S. (1 985). Considerations prior to establishing preventative intervention for eating disorders. The Ontario Psvcholoaist. 17, 12-1 7.

Smolak, L., & Levine, M. P. (1 994). Toward an empirical basis for primary prevention of eating problems with elementary school children. Eating Disorders. 2, 293-307.

Stein, D. M., & Reichart, P. (1 990). Extreme dieting behaviors in early adolescence. Joumal of Earlv Adolescence. 10, 108-1 21.

Steiner-Adair, C. (1 993). The politics of prevention. In P. Fallon, M. A. Katzrnan, & S. C. Wooley (Eds.), Feminist perspectives on eatina disorders (pp. 381 - 394). New York: Guilford Press.

Theander, S. (1 970). Anorexia neivosa: A psychiatric investigation of 94 female patients. Acta Psvchiatry Scandinavica (suppl.). 21 4, 1-1 94.

Preventina Disordered Eatinci 1 1 6

Vandereychken, W., & Meemann, R. (1 984). Anorexia nervosa: Is prevention possible? International Journal of Psvchiatry in Medicine. 14, 191 -205.

Wenar, C. (1 994). Develo~mentat ~svcho~athoioav: From infancv throuuh adolescence (3d ed.). New York: McGraw Hill.

Wiseman, C. V., Gray, J., Mosimann, J., 8 Ahrens, A. (1 992). Cultural expectations of thinness in women: An update. lntemational Journal of Eating Disorders. 11, 85-89.

Wiseman, C. V. Gunning, F. M., & Gray, J. J. (1 993). lncreasing pressure to be thin: 1 9 years of diet products in television commercials. Eatina Disorders, 1 52-61. -8

Wooley, W. O., & Wooley, S. (1982). The Beverly Hills eating disorder: The mass marketing of anorexia nervosa. lntemational Journal of Eatinq Disorders. 1, 57-69.

Yager, J. (1985). Afteword. In R. Hales & A. Frances (Eds.), Psvchiatrv U~date: Annual Review (Vol. 4.. pp. 51 6-521 ). Washington, D.C.: American Psychiatric Association.


Table 1

Ethnicity of Participants

Sample Pre/Post Follow-up Extended

Et hnicity

Caucasian

Indo-Canadian

Asian

Middle-Eastern

First NationdNative Indian

Black

Other

Total

Grade

Five

Six

Seven

Total

Sex

Male

Female

Total

Condition

Treatment

Cornparison

Total

Note. Values are in percentages (number of students in parentheses).


Table 2

Corrected and Uncorrected ANCOVA F-values

SS MS - F for Cond

Dependent Class Cond Class W/I Cond Uncorr Co rr Variable Cond

KQ

BES

BSM

ChEAT

ChEATl

ChEAT2

ChEAT3

Dieting

(71 8.36)

(1 35-24)

(O. 84)

(1 15.96)

(94.13)

(2.19)

(3.23)

(0.70)


Table 3

Knowledge Questionnaire (KQ) Analysis of Covariance

Source - -- -.

Sum of Squares

Mean Square

-- --

condition

sex

condXsex

preKQ

age

al1 covariates

Error


Table 4

Knowledge Questionnaire (KQ) Scores

Condition Pre Post Pre Post Pre Post

Treatment

Boys

M - SD - n -

Girls

M -

SD - n -

Cornparison

Boys

Girls

M - 3.56 3.72 4.47 4.36 5.63 5.37

Preventina Disordered Eatinu 121

Table 5

Body Esteem Scale (BES) Analysis of Covariance

Source Sum of - DF Mean - F Squares Square

condition 1 35.24 1 135.24 12.55**

sex

al1 covariates 7774.77 2 3887.39


Table 6

Body Esteem Scale (BES) Scores

G rade5 Grade6 Grade7


Treatment

Boys

M - SD - n -

G iris

M - SD - n -

Cornparison

Boys

M - SD - n -

Girls

M - SD - n -


Table 7

Body Satisfaction Measure (BSM) Analysis of Covariance

Source Surn of - DF Mean - F Squares Square

condition

sex

condXsex

preBSM

age

ail covariates

Error


Table 8

Body Satisfaction Measure (BSM) Scores

Grade5 Grade6 Grade7


Treatment

Boys

M - SD - n -

Girls

M - SD - n -

Cornparison

Boys

M - SD - n -

Girls

M - SD - n -


Table 9

Children's Eating Attitudes Test (ChEAT) Analysis of Covariance

source Sum of - DF Mean - F Squares Square

condition 1 15.96 1 1 15.96 5.56*

sex 3.85 1 3.85 0.1 8

p reC h EAT 9466.20 1 9466.20 453.70**

al1 covariates 9823.54 2 491 1.77 235.42**

Error 9597.56 460 20.86


Table 10

Children's Eating Attitudes Test (ChEAT) Total Scores


Treatment

Boys

M - SD - n -

Girls

M -

SD - n -

Corn parison

Boys

M - SD - n -

Girls

M - SD - n -


Table 11

Childrens Eating Attitudes Test (ChEAT) Factor 1 Analysis of Covariance

Source Sum of Squares

Mean Square

condition

sex

condXsex

preChEATl

age

al1 covariates

Error


Table 12

Children's Eating Attitudes Test (ChEAT) Factor 1 Scores



Treatment

Boys

M - SD - n -

Girls

M - SD - n -

Cornparison

Boys

M - SD -

n - Girls

M - SD - n -


Table 13



condition 1 .O3 1 1 .O3 0.49

sex 0.1 9 1 0.1 9 0.09

al1 covariates 21 3.03 2 106.52 50.51 *


Table 14




Treatment

Boys

M - SD - n -

Girls

M - SD - n -

Cornparison

Boys

M - SD - n -

Girls

M - SD - n -


Table 15



condition

sex

condXsex

preChEAT3

age

all covariates

Error


Table 16




Treatment

Boys

M -

SD -

n - Girls

M -

SD - n -

Cornparison

Boys

M -

SD - n -

Girls

M - SD 7

n -


Table 17

ChEAT score220 (Follow-up sample only)

pretest posttest follow-up

Condition

Treatment

n -

Corn parison

n -

Note. Values are in percentages (number of subjects in parentheses).


Table 18

Dieting Analysis of Covariance


condition

sex 0.00 1 0.00 0.00

Error 49.58 430 0.1 2

Preventina Disordered Eatinci 135

Table 19

Self-Reported Dieting

Grade 5 6 7

Condition Pre post Pre post Pre post

Treatment

Boys 31.8(14)

n - 44

Girls 22.7(1 O)

n - 44

Total 27.3(24)

Cornparison

Boys 33.3(11)

n - 33

Girls 34.1 (1 5)

n - 44

Total 33.4(26)



Table 20

"Gaining" Analysis of Covariance


condition

sex

condXsex

preGain

age

ail covariates

Error


Table 21

Self-Reported Gaining

Grade 5 6 7

condition Pre post Pre post p re- post

Treatment

Boys 1 1.4(5)

n - 44

Girls i 1.4(5)

n - 44

Total 1 1.4(10)

Cornparison

Boys 21.2(7)

n - 33

Girls 1 2.2(5)

n - 4 1

Total 27.3(12)



Table 22

Body Mass Index (BMI) Analysis of Covariance


condition 0.05 1 0.05 0.03

sex 6.76 1 6.76 3.35

al1 covariates 1 044.34 2 522.1 7 258.73"

E rror 508.60 252 2.01

*p<.o001


Table 23

Methods of Weight Reducing Used by Boys and Girls

Grade 5 Grade 6 Grade 7

Boys Girls Boys G i ris Boys Girls Total

Exe rcise

n - Cut snackdjunk

n - Skip meals

n - Cut calories

n - Fasting

n -

Vomiting

n - Other

n - Total N 29 26 43 60 13 22 193

Note. Results are in percentages. Values are tabulated from pretest questionnaires. Total number of subjects reporting that they were trying to lose weight=l91.


Table 24

Follow-up Corrected and Uncorrected Analysis of Covariance F-values

SS MS - F for Cond

Dependent Class Cond Class WII Cond Uncorr Corr Variable Cond

BES 113.53 47.12 1 3.28 (47.12) 3.69 n.s. 3.55 n.s.

BSM 1.40 0.68 0.14 (0.68) 1.41 n.s. 4.85 n.s.


Table 25

Follow-up Knowledge Questionnaire (KQ) Analysis of Covariance


- - -

condition

sex 0.75 1 0.75 0.30

al1 covariates 24.46 2 12.23 4.89'*'

Error 354.80 1 42 2.50


Table 26

Follow-up Body Esteem Scale (BES) Analysis of Covariance

Source Sum of Squares

Mean Square

condition

sex

condXsex

preBES

age

al1 covariates

Error


Table 27

Follow-up Body Satisfaction Measure (BSM) Analysis of Covariance


condition 0.68 1 0.68 1.41

sex 2.53 1 2.53 5.30*

al1 covariates 83.36 2 41.68 87.29**


Table 28

Follow-up ChEAT Analysis of Covariance


condition 33.03 1 33.03 2.06

sex 16.37 1 16.37 1 .O2

al1 covariates 1046.89 2 523.45 32.69*

Error 2321.51 1 45 16.01


Table 29

Follow-up ChEAT Factor 1 Analysis of Covariance

Source Surn of - DF Mean - F Squares Square

condition 42.76 1 42.76 4.50*

sex 5.54 1 5.54 0.58

al1 covariates 838.60 2 41 9.30 44.16**

Error 1376.92 145 9.50


Table 30

FOI low-up "Dieting " Analysis of Covariance

Source Sum of DF Mean - F Squares Square

condition 0.09 1 0.09 0.82

sex

al1 covariates 3.77 2 1.89 16.36**


Table 31

Follow-up "Gaining" Analysis of Covariance


condition 0.04 1 0.04 0.90

sex

condXsex

preGain

age

ail covariates

Error


Table 32

High Risk Knowledge Questionnaire (KQ) Analysis of Covariance

Source Sum of - DF Mean - F Squares Sauare

condition 45.71 1 45.71 19.03**

sex 0.00 I 0.00 0.00


Table 33

High Risk Body Esteem Scale (BES) Analysis of Covariance

Source Sum of - DF Mean F - Squares Square

condition 145.31 1 145.31 13.21'*

sex

condxsex

Error 31 8.89 29 31 8.89


Table 34

High Risk Body Satisfaction Measure (BSM) Analysis of Covariance


condition 0.06 1 0.06 0.04

sex 13.36 1 13.36 9.64**

Preventing Disordered Eating 1 5 1

Table 35

High Rlsk Childrens Eating Attitudes Test (ChEAT) Analysis of Covariance


condition 79.1 8 1 79.18 1.36

sex 65.98 1 65.98 1 .14

preCh EAT 628.29 1 628.29 10.82*

Error 1741.55 30 58.05


Table 36

High Risk Dieting Analysis of Covariance


condition 0.30 1 0.30 2.37

sex 0.09 1 0.09 0.72


Table 37

High Risk Gaining Analysis of Covariance

Source Sum of - DF Mean - F Sauares Sauare

condition 0.04 1 O .O4 2.3 1

sex 0.04 1 0.04 2.08

Error 0.55 32 0.02


Table 38

High Risk Body Mass Index (BMI) Analysis of Covariance


condition 5.65 1 5.65 12.44*

sex 0.61 1 0.61 1.34

E rro r 4.55 1 O 0.45


Figure 1

- - Pre post

Time

Figure 2

BES

8 I - - Pre post

Time


Figure 3

BSM

- . Pre post

Time

Figure 4

ChEAT I

I I I I I 1 I 1 I -----_ OlCompanronI I

I I

post

Time


Figure 5

ChEAT Factor 1

Time

Figure 6

1 1 . - - Pre post f O Il0 wup

Time


Figure 7

BES 30.

25'

2 0 20' O cn

15'

10

Time

- - -

--r-- I

4

I I

Figure 8

Pre post followup

BSM

I ----- I - - - - - I --L,-

--r-- I I I I

---- 4 --------- I I I

--r-- --r-- --r-- I I I I I l I I I

1) - I 1 Cornparison

I 1 I I 1 I --L-- I I ----- 1 -----

--r-- I

Treaune

I 1

20 1 - - - Pre post f O Il0 wup

Time


Figure 9

ChEAT

0J I I I

Pre post followup

Time

Figure 10

ChEAT Factor 1

I 9 I

Pre post f O Il0 wup

Time


Figure 11

KQ High Risk

Time

Figure 12

BES High Risk

I I

Pre post

Time


Figure 13

BSM High Risk

Time

Figure 14

ChEAT High Risk

Time


Figure 15

BMI High Risk

1s I I I

Pre post

Time

IMAGE EVALUATION TEST TARGET (QA-3)

APPLIED IMAGE. Inc 1653 East Main Street - -. - Rochester, NY 14609 USA -- - Phone: 71 6/48263OO -= Fax: 71 6/288-5989

O 1993. AOp l i Image. lnc.. AJl Rlghri fleserved

dissertation submllted in partial fulfillment of the … · 2004-09-01 · exclusive licence...

Documents