Citation
The effects of preliminary food intake and dietary restraint during periods of anticipated deprivation

Material Information

Title:
The effects of preliminary food intake and dietary restraint during periods of anticipated deprivation
Creator:
Spauster, Edward T., 1953-
Copyright Date:
1984
Language:
English

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Edward T. Spauster. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
11988656 ( OCLC )
ACR2563 ( LTUF )
30595538 ( ALEPH )

Downloads

This item has the following downloads:

effectsofprelimi00spau.pdf

effectsofprelimi00spau_0025.txt

effectsofprelimi00spau_0113.txt

effectsofprelimi00spau_0031.txt

effectsofprelimi00spau_0012.txt

effectsofprelimi00spau_0027.txt

effectsofprelimi00spau_0099.txt

effectsofprelimi00spau_0049.txt

effectsofprelimi00spau_0085.txt

effectsofprelimi00spau_0082.txt

effectsofprelimi00spau_0036.txt

effectsofprelimi00spau_0083.txt

effectsofprelimi00spau_0044.txt

effectsofprelimi00spau_0015.txt

effectsofprelimi00spau_0021.txt

effectsofprelimi00spau_0079.txt

E4K6V71R8_PJC5W3_xml.txt

effectsofprelimi00spau_0050.txt

effectsofprelimi00spau_0018.txt

ES6ZD9FGC_4TB2B9_xml.txt

effectsofprelimi00spau_0103.txt

effectsofprelimi00spau_0057.txt

effectsofprelimi00spau_0005.txt

effectsofprelimi00spau_0052.txt

effectsofprelimi00spau_pdf.txt

effectsofprelimi00spau_0055.txt

effectsofprelimi00spau_0062.txt

effectsofprelimi00spau_0022.txt

effectsofprelimi00spau_0097.txt

effectsofprelimi00spau_0047.txt

effectsofprelimi00spau_0034.txt

effectsofprelimi00spau_0071.txt

effectsofprelimi00spau_0039.txt

effectsofprelimi00spau_0043.txt

effectsofprelimi00spau_0089.txt

effectsofprelimi00spau_0041.txt

effectsofprelimi00spau_0086.txt

effectsofprelimi00spau_0010.txt

effectsofprelimi00spau_0065.txt

effectsofprelimi00spau_0014.txt

effectsofprelimi00spau_0006.txt

effectsofprelimi00spau_0054.txt

effectsofprelimi00spau_0035.txt

effectsofprelimi00spau_0109.txt

effectsofprelimi00spau_0120.txt

effectsofprelimi00spau_0077.txt

effectsofprelimi00spau_0045.txt

effectsofprelimi00spau_0111.txt

effectsofprelimi00spau_0070.txt

effectsofprelimi00spau_0037.txt

effectsofprelimi00spau_0117.txt

effectsofprelimi00spau_0033.txt

effectsofprelimi00spau_0000.txt

effectsofprelimi00spau_0076.txt

effectsofprelimi00spau_0100.txt

effectsofprelimi00spau_0063.txt

effectsofprelimi00spau_0069.txt

effectsofprelimi00spau_0017.txt

effectsofprelimi00spau_0040.txt

effectsofprelimi00spau_0003.txt

effectsofprelimi00spau_0087.txt

effectsofprelimi00spau_0075.txt

effectsofprelimi00spau_0067.txt

effectsofprelimi00spau_0028.txt

effectsofprelimi00spau_0048.txt

effectsofprelimi00spau_0091.txt

effectsofprelimi00spau_0016.txt

effectsofprelimi00spau_0090.txt

effectsofprelimi00spau_0001.txt

effectsofprelimi00spau_0056.txt

effectsofprelimi00spau_0096.txt

effectsofprelimi00spau_0104.txt

effectsofprelimi00spau_0112.txt

effectsofprelimi00spau_0108.txt

effectsofprelimi00spau_0030.txt

effectsofprelimi00spau_0064.txt

effectsofprelimi00spau_0008.txt

effectsofprelimi00spau_0038.txt

effectsofprelimi00spau_0088.txt

effectsofprelimi00spau_0107.txt

effectsofprelimi00spau_0092.txt

effectsofprelimi00spau_0098.txt

effectsofprelimi00spau_0020.txt

effectsofprelimi00spau_0093.txt

effectsofprelimi00spau_0007.txt

effectsofprelimi00spau_0094.txt

effectsofprelimi00spau_0009.txt

effectsofprelimi00spau_0024.txt

effectsofprelimi00spau_0072.txt

effectsofprelimi00spau_0032.txt

effectsofprelimi00spau_0068.txt

effectsofprelimi00spau_0116.txt

effectsofprelimi00spau_0119.txt

effectsofprelimi00spau_0051.txt

effectsofprelimi00spau_0060.txt

effectsofprelimi00spau_0110.txt

effectsofprelimi00spau_0066.txt

effectsofprelimi00spau_0081.txt

effectsofprelimi00spau_0095.txt

effectsofprelimi00spau_0115.txt

effectsofprelimi00spau_0118.txt

effectsofprelimi00spau_0004.txt

effectsofprelimi00spau_0074.txt

effectsofprelimi00spau_0105.txt

effectsofprelimi00spau_0078.txt

effectsofprelimi00spau_0061.txt

effectsofprelimi00spau_0002.txt

effectsofprelimi00spau_0023.txt

effectsofprelimi00spau_0013.txt

effectsofprelimi00spau_0073.txt

effectsofprelimi00spau_0059.txt

effectsofprelimi00spau_0084.txt

effectsofprelimi00spau_0101.txt

effectsofprelimi00spau_0106.txt

effectsofprelimi00spau_0114.txt

effectsofprelimi00spau_0026.txt

effectsofprelimi00spau_0019.txt

effectsofprelimi00spau_0102.txt

effectsofprelimi00spau_0042.txt

effectsofprelimi00spau_0046.txt

effectsofprelimi00spau_0058.txt

effectsofprelimi00spau_0053.txt

effectsofprelimi00spau_0011.txt

effectsofprelimi00spau_0029.txt

effectsofprelimi00spau_0080.txt


Full Text
THE EFFECT OF PRELIMINARY FOOD INTAKE AND DIETARY RESTRAINT
DURING PERIODS OF ANTICIPATED DEPRIVATION By
EDWARD T. SPAUSTER

August, 1984
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF~ FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THlE DEGREE OF DOCTORR OF PHILOSOPHY
UNIVERSITY OF FLORIDA

1984




ACKNOWLEDGEMENTS

In acknowledging those people who have assisted me in the
completion of this study, I wish to thank the members of my doctoral committee. Dr. Dorothy Nevill has been a constant support and a trusted advisor since my admission to the university. Dr. Joseph Wittmer has provided leadership and counsel throughout the production of this dissertation. Drs. Ellen Amatea and William Froming have both shown support and guidance as instructors and committee members, and Dr. James Archer has given his time and energy on short notice.
Dorene Tranes, Lynn Wigglesworth and Diane Camiter were the three
undergraduate research assistants who gave long hours to gather the data for this study. Sarah Drew assisted in the coordination of their efforts. In addition, many people helped with the mechanics of producing a finished document and arranging a doctoral defense. Among them, Ces Bibby, Ann Cusick, Nancy Lubell, Joyce Perrotta, Len Travaglione, Curtis Walling and Flora Zaken-Greenberg stand out as particularly important.
For her unconditional faith in the quality of my endeavors, I would like to thank my wife, Elizabeth Coburn. She is the single greatest support in my work and in my life.




TABLE OF CONTENTS

PAGE
ACKNOWLEDGEMENTS~~~~ ~ ~ ~ ~ i ii
ABSTrRACT .. ....................... . . v
CHAPTER
ONkE INTRODLJJCTION .. .............. ........1
Statement of the Problem .. ............. ...4
Need for the Study .. ............. ......5
Purpose of the Study .. .............. ....7
Significance ofthe Study. .................8
definitions of Termns. .. .................10
Organization ofthe Study .. ...............12
TWO REVIEW OF RELATED LITERATURE .. ..............13
Binge Eating and besity. .. ...............13
Binge Eating in Anorexia Nervosa .. ............19
Binge Eating in "Normal" Weight Individuals. .......22 Dietary Restraint and Counter-regulation .. ........27
Treatment ofBinge Eating. .................38
Summ ry . . . . . . 44
THREE METH'I-OLLGY .. ......................46
Experimental Hypotheses .. ................46
The Research Design .. ..................47
Subjects . .. .. .. .. .. .. .. .. .. .. ..49
Experimental Procedures .. ................50
I nstrumnentation .. ....................55
Data Analysis .. .....................58
Rationalefor Methodologjy.. .............. 59
Limitations of the Study. .. ...............60




RESULTS. ............... ..... .. .. .
Food Consumpt ion....... ..... . .... .. .. .. ..
Taste and Buying ntention..... ... .. .. .. .. ..
Summiary . . . . . . . . . . . .
SUMMARY, DISCUSSION, CONCLUSIONS, IMPLICATIONS, RECOMIMENDATIONS ....................

Summiary.... .. .. ..
Discussion.. .. ....
Conclusions.... .. ..
Implications... .. .. Reccriirndat ions ....

APPENDICES
A RESTRAINT QUESTIONNAIRE ........
B MARKETING QUESTIONNAIRE ........
C TASTE PREFERENCE SHEET ........
D SUBJECT QUESTIONNAIRE .........
E GENERAL INSTRUCTIONS FOR EXPERIMENTERS.
F INFORMED CONSENT FORM.... .. .. ....
BIORAPHICAL SKETCH. ........... .. .. .

FOUR FIVE

PAGE
62
63
64
76
77
77 79 89
90 91




Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
TE EFFECTS OF PRELIMINARY FOOD INTAKE AND DIETARY RESTRAINT
DURING PERIODS OF ANTICIPATED DEPRIVATION By
Edward T. Spauster
August, 1984
Chairperson: Dr. Paul J. Wittmer Co-Chairperson: Dr. Dorothy D. Nevill Major Department: Counselor Education
This study examined the effect of level of dietary restraint and the presence or absence of a food preload on subjects expecting a four-hour period of food deprivation. A preload is a pre-determined
amount of food administered to subjects prior to a free eating period (i.e., a period of time during which they may eat as much as they wish). Thirty restrained female subjects (high preoccupation with food and body size) and thirty unrestrained subjects (low preoccupation with food and body size) were randomly assigned to one of two preload conditions (preloaded; non-preloaded). Preloaded subjects were asked to consume a standard amount of highly caloric food prior to a period of time during which they could eat as much as desired. Non-preloaded subjects were not administered this standard amount of food. Under the guise of a study examining the effects of hunger on taste, all subjects were asked to fast for the four hours subsequent to the morning experimental




session. After this request was made all subjects were given 15 minutes to privately consume the food presented (M&M's, peanuts and animal crackers) and provide ratings of taste and buying intention.
The dependent variables measured were amount of food consumed, taste ratings and buying intentions. Analysis of variance found no differences between the four groups in the amount of food consumed. Significant Restraint X Preload interactions were found for taste ratings (p < .01) and buying intention (p < .05). This indicated that restrained preloaded subjects had higher ratings of taste and buying intention than restrained non-preloaded subjects but unrestrained preloaded subjects had lower ratings than those unrestrained subjects who were not preloaded.
The prediction that restrained subjects expecting a period of food deprivation would overconsume in the absence of a restraint breaking preload was not supported. The results for the cognitive variables, taste ratings and buying intention, are discussed in light of this finding of no overconsumption. Other implications of the study and reccnmendations for future research are discussed.




CHAPTER ONE
INTRODUCION
Binges speak the voice of survival. They are protective
mechanisms. Binges are signals that something is terribly
wrong, that you are not giving yourself what you need--either
physically (with food) or emotionally (with intimacy, work
relationships). They are your last stand against deprivation.
(Roth, 1982, p. 16)
In an investigation of the eating patterns of the obese, Stunkard (1959) identified binge eating as a distinct eating disturbance in a
percentage of his subjects. Described as an orgiastic consumption of large amounts of food in a relatively short period of time, binge eating has been the subject of many investigations into obesity. TWO hypotheses, the psychosomatic and the Schacterian, have received much attention in these studies. The psychosomatic hypothesis (Kaplan & Kaplan, 1957) holds that non-nutritive eating is a learned coping response associated with anxiety reduction. Having learned this response in childhood, the overweight adult resorts to this behavior when faced with anxiety. The Schacterian hypothesis (Schacter, 1967) holds that the obese are more sensitive to external cues, such as the appearance and taste of food, and less sensitive to internal ones, such as hunger and blood sugar level.
Both hypotheses have been supported and refuted in the literature (Herman & Polivy, 1975; Rodin, 1975; Schacter, 1967; Schacter & Rodin, 1974; Slochower, 1976). Efforts to incorporate both hypotheses and identify clear differences between obese and normal weight individuals
1




have met with mixed results (McKenna, 1972; Tim & Rucker, 1975) and studies of caloric intake differences between the two groups indicate that the obese may actually eat less than normal weight individuals (Johnson, Burke & Mayer, 1956; Stefanik, Heald & Mayer, 1959). In more recent years, binge eating has been identified as a problem behavior for scme normal and underweight individuals, showing its more extreme forms in anorexia nervosa and bulimia (DSM-III, 1980). In light of these findings, it becomes purposeful to consider binge eating behavior across the spectrum of weight classifications and not as a behavior exclusive to the obese.
Herman and Polivy (1975) have postulated that restraint (dietary consciousness) plays a role in overeating. In laboratory studies, subjects categorized as "restrained" (great preoccupation with body weight and food) have been found to consume more food in a taste test
simulation after they had ingested a food preload perceived to be high in calories (Herman & Mack, 1975; Hibscher & Herman, 1977; Spencer & Fremouw, 1979). A preload is a pre-determined amount of food administered to subjects prior to a free eating period (i.e., a period of time during which they may eat as much as they wish). It appears that these dietary conscious individuals once they have indulged in forbidden food continue to eat rather than return to their strict standards of food intake. Subjects categorized as "unrestrained" (low preoccupation with body size and food) do not exhibit this phencimenon and decrease intake after a preload of food. This overconsumption (or, counter-regulation) by restrained subjects has been likened to binge eating (Hibscher & Herman, 1977).




3
Wardle and Bienart (1981) in a theoretical review of binge eating suggest that these studies confirm two ideas that have emerged from clinical literature. First, there is an association between binges and dietary restriction; and second, a binge can be triggered by the ingestion of a quantity ot highly caloric or "forbidden" food. They further state, "it is clear that dietary restriction is implicated in a causal role in relation to binges. However, its action is primarily as a background against which specific events can tip the balance into uncontrolled consumption" (p. 107).
Lowe (1982), in seeking to account for the pattern of overeating in restrained subjects, proposed that a situational factor, antici--ated deprivation, influences a subsequent binge. Lowe's study demonstrated that after an ingestion of a food preload, restrained subjects who anticipated a four hour fast ate significantly more than their counterparts who were not expected to fast. Lowe suggests that restrained individuals, constantly aware of their eating habits, realize they will eventually return to dieting, a form of deprivation. Thus, when restraint is broken by a preload, they continue to eat before having to return to a "tomorrow" of deprivation.
In an area often left to postulation based on clinical observation, the concepts of dietary restraint and anticipated deprivation have been significant contributions to the study of eating behavior. Research has identified a population of dietary restrainers who, when given a preload of forbidden food, abandon their rules about dieting and binge. The influence of anticipated deprivation after a preload further enhances this behavior. However, the role of anticipated deprivation before the




rules have been broken remains a question for future research. If anticipated deprivation is found to play a significant part in the choice to begin an episode of overconsumption, the further understanding of this variable will provide insight into binge eating and the dynamics of weight loss.
Statement of the Problem
If researchers accept Wardle and Bienart's (1982) assertion that
restraint serves primarily as a background for binge eating, it becomes the task of scientific inquiry to investigate what situational factors impact upon the overconsumption of food in the individual with high restraint. Lowe (1982) in identifying anticipated deprivation as a variable that can be experimentally induced has begun this effort. However, Lowe's study conformed to prior restraint investigations in utilizing a preload of food to first break down the restraint of her subjects. Her results, therefore, reflect an understanding of anticipated deprivation only as it affects the already thwarted dieter. Investigators are left to speculate if anticipated deprivation is, in itself, a sufficient cause for the temporary abdication of dietary restraint.
Fron our knowledge of the problems of the dieter, it is reasonable to assume that every binge or deviation front a diet does not result from the consumption of "forbidden" food accidentally or at the request of another. The factors which precipitate the restrained individual to choose to initially overeat remain, for the most part, uninvestigated. This study investigated the role of anticipated deprivation in the




counter-regulation of food intake for restrained and unrestrained individuals with and without a preload of food and attempted to answer the following questions:
1. Is anticipation of a period of toad deprivation a
sufficient factor to cause the individual with a
preoccupation with food and body size to overeat?
2. Does the ingestion of a preload that consists of food
normally considered as highly caloric and "forbidden"
differentially affect the later consumption and attitudes toward food of restrained and unrestrained individuals anticipating a period of food deprivation?
Need for the Study
In recent years, the quest to be thin has taken on the qualities of a national obsession. Repeatedly, consumers are bombarded with advertisements for diet aids, fast weight-loss programs and reducing devices that promise a slender body, exciting social life and miraculous health benefits. This movement toward thinness has had a pronounced effect on women and our image of them. Garner, Garfinkel, Schwartz and Thompson (1980) report that American society has changed its conception of the ideal woman to one who is taller, lighter and smaller proportioned. They point to a growing number of diet articles in American magazines as further evidence of the public's increased weight consciousness. This movement is further supported by industry's mass-marketing of health and beauty aids, professional sports and fashion for the fitness-minded.




6
on the surface, this growing preoccupation with health and beauty appears in order for a population often viewed as overweight and overworked. It is not, however, without its consequences. Disturbances in eating have taken on epidemic proportions. Obesity presently affects 23%-68% of wnen in the United States (United States Department of Health, Education and Welfare, 1967). Anorexia nervosa develops in one of every 250 females between 12 and 18 years old and claims the lives of 15%-21% of them (DSM-III, 1980). It is estimated that as many as 18% of college women meet the diagnostic criteria (DSM-III, 1980) of bulimia (Halmi, Falk & Schwartz, 1981).
Binge eating has been identified as a behavioral response to food present in many American women regardless of weight classification or presence of a psychiatric disorder. Stunkard (1959) first coined the term from his studies of the eating patterns of the obese. Wilson (1976) discussed the failure of standard behavioral management programs in treating the obese binge eater. Several studies have examined binge-purge behavior in patients with anorexia nervosa (Beaumont, George & Smart, 1976; Bruch, 1973; Dally & Gcmez, 1979; Halmi, 1974). In 1980, the American Psychiatric Association established bulimia as a distinct eating disorder with uncontrolled binge eating as the primary characteristic (DSM-III, 1980). Orbach (1978) describes binge eating as a problem with many successful, middle class and well-educated women. Recent college surveys point to a high prevalence of binge eating among college students with as many as 35% of respondents in one survey labelling themselves "binge-eaters" (Halmi et al., 1981) and 78% of respondents in another reporting episodes of overconsumption (Ondercin, 1979).




Wardle and Bienart (1981) state, "the division (within clinical literature) on the basis of weight has obscured same camxn ground in the eating disturbances experienced by these patients (obese, normal weight and anorexic" (p. 97). The concept of dietary restraint has proven to be a bridge across weight classifications. The preload of "forbidden" food has been repeatedly demonstrated as a trigger for overconsumption by the dietary conscious (Herman & Mack, 1975; Herman &
Polivy, 1975; Polivy, 1976; Polivy & Herman, 1976, Hibscher & Herman, 1977; Spencer & Fremouw, 1979). The situational factor, anticipated deprivation, has been shown to exert an even stronger effect on preloaded subjects. Its effect on individuals who have not broken their strict dietary rules was the focus of this study.
Purpose of the Study
This study expanded upon the prior work in dietary restraint and
anticipated deprivation. It examined a laboratory phenomenon, counterregulation, that has been likened to binge eating (Polivy, 1976; Spencer & Fremouw, 1979; Wardle & Bienart, 1981). It is a behavior where restrained individuals will consume more in a free eating situation after eating a preload of food believed to be high in calories than if they had not eaten anything previously. The effects of two independent variables, level of restraint and preload, were investigated in subjects expecting a period of food deprivation, as described by Lowe (1982). Level of restraint, an internal variable, was assessed as restrained or unrestrained by use of the Restraint Questionnaire (Polivy, Howard & Herman, 1976). Subjects were assigned to either a preload or a no preload condition. All experienced an anticipation of deprivation.




8
The purpose was to assess the effects of the presence or absence of a preload on the consumption and attractiveness of food by restrained and unrestrained subjects who expected a period of food deprivation. Actual food consumption served as a behavioral measure, consistent with the measures used in previous restraint designs. Attractiveness of food was assessed by a questionnaire administered to the subjects. Each dependent measure permitted the examination of any existing main and interactive effects of levels of restraint and presence or absence of preload.
Significance of the Study
Wilson (1976), in commenting on his treatment of six binge eaters, believes these clients differ sharply from the obese client upon whom the typical behavioral treatment program is based. He reports that two of his clients had little difficulty between binges with the standard behavior modification program of Stuart and Davis (1972) but failed to control the intensity or frequency of their binges. He agrees with Mahoney (1975) that cognitive variables exert a strong influence on eating behavior.
An understanding of the factors that affect the relinquishment of restraint can be assisted by an examination of the role of anticipated deprivation. An identification of a component of the cognitive set that predisposes the binge eater to consume unwanted quantities of forbidden food will be useful in the construction of more effective behavioral treatment. Confirmation that anticipated deprivation exerts a strong influence on binge eating will question the wisdom of long-term, weight loss programs and overly restrictive diets.




Within the arena of the restraint research itself, it becomes important to assess if the ingestion of an initial preload is a necessary factor in binge eating. All studies to date have been based on this assumption. None have attempted to introduce an experimental variable to restrained subjects without a preload. If anticipated deprivation is found to be powerful enough to prompt the individual to overeat, it will invite the testing of other factors believed to influence binge eating.
At present, reports on developing treatment programs for bulimia (Kubistant, 1982; Leclaire & Berkowitz, 1982; Boskind-Lodahl & White, 1978) indicate a substantial reliance on the clinical observations of the counselors providing therapy. Further understanding of the roles of restraint and anticipated deprivation can lend additional support to efforts to find adequate treatment methods for bulimia.
The belief that the restrictive nature of dieting confounds the
efforts of the female dieter to gain control of her eating disturbance is described in detail by Orbach (1978a). In treating compulsive eating disturbances in women she first prescribes a curtailment of all efforts
to lose weight. Unfortunately, she reports no empirical evidence to support what may be an appropriate treatment strategy. The experimental investigation of restraint and anticipated deprivation will offer
insight into this area. It will also respond to the call by Wardle and Bienart (1981) for an alliance between the clinicians and the experimentalists in our attempts to increase our knowledge.




definitions of Terms
Anorexia Nervosa
DSM-III (1980, p. 69) outlines the following diagnostic criteria:
A. An intense fear of becciming obese, which does not
diminish as weight loss progresses.
B. Disturbance of body image, e.g., claiming to "feel fat"
even when emaciated.
C. %bight loss of at least 25% of original body weight or,
if under 18 years of age, weight loss fran original body
weight plus projected weight gain fran expected growth
charts may be combined to make the 25%.
D. Refusal to maintain body weight over a minimal normal
weight for age and height.
E. No known physical illness that would account for the
weight loss.
Anticipated Deprivation
The expectation of a period of time during which individuals will
be restricting their intake of food by either fasting or dieting.
Binge
The consumption of large quantities of food, unrelated to hunger,
in a short period of time.
Binge Eater
An individual who exhibits a pattern of binge eating episodes.
This pattern does not necessarily follow a particular periodicity.
Bulunia
DSM-IlI (1980, p. 70) outlines the following diagnostic criteria:
A. Recurrent episodes of binge eating (rapid consumption of a
large amount of food in a discrete period, usually less than
two hours).
B. At least three of the following:
1) consumption of high calorie, easily ingested food
during a binge.
2) inconspicuous eating during a binge.
3) termination of such eating episodes by abdominal pain,
sleep, social interruption, or self-induced vciniting.
4) repeated attempts to lose weight by severely restrictive




diets, self-induced vomiting, or use of cathartics or
diuretics.
5) frequent weight fluctuations greater than ten pounds due to
alternating binges and fasts.
C. Awareness that the eating pattern is abnormal and fear of not being
able to stop voluntarily.
D. pressed imood and self-deprecating thoughts following eating binges. E. The bulimic episodes are not due to Anorexia Nervosa or any known
physical disorder.
Counter-regulation
A laboratory phenomenon, similar to binge eating, where subjects categorized by the Restraint Questionnaire (Rolivy et al. 1976) as restrained (high preoccupation with body weight and food) consume more after eating a preload of food believed to be high in calories than if they had not eaten anything previously. obesity
A weight classification categorizing those individuals who weigh more than 15% of the acceptable weight range for their sex, age, height and body frame size.
Pre load
An amount of food administered to subjects for their consumption
prior to a free eating period during which the subjects will be allowed to eat as much as they wish. Preloads are often used to guarantee the subject is full before the experimental manipulation is administered. They may also be used to break down a subject's resistance to eating food that is high in calories.
Restraint
A level of dietary consciousness as measured by the Restraint
Questionnaire (Polivy et al., 1976). It assesses the extent to which individuals exhibit attitudinal and behavioral concern about dieting.




organization of the Study
The remainder of this study is organized into four chapters. The second chapter is a review of the related literature. Included in this chapter are considerations of 1) binge eating in obesity, anorexia nervosa, and normal weight ranges, 2) the concept of dietary restraint in binge eating, and 3) current treatment strategies for binge eating.
The third chapter presents the research methodology. The fourth chapter contains the results of the study. A surinary of the study, which includes a discussion of its implications and suggestions for future research, cnprises the final chapter.




CHAPTER TWO
REVIEW OF RELATED LITERATURE
The literature review in this chapter consists of six sections and focuses on the theories and research which form the basis for this study. The first three sections examine, separately, the role of binge eating in obesity, anorexia nervosa and "normal" weight groups. Research efforts in eating behavior have often focused on a specific weight group, e.g., obese individuals, and compared its members to a control group of normal weight subjects. As this reviewer will discuss, this division by weight classification has not proved to be an adequate method of describing differences in eating behaviors. However, in light of the quantity of studies that have differentiated by weight, this division serves as a useful organizational tool.
The fourth section presents the findings in the area of counterregulation and the role of dietary restraint in this behavior. These studies form the basis for the experimental methodology that will be utilized in this study. The fifth section discusses the current strategies and concerns in the treatment of binge eating. The last section is a sLvimary of the reviewed research and the implications for the study.
Binge Eating and obesity
The earliest behavioral reports of binge eating are to be found in studies of the obese. Stunkard's (1959) analysis of the eating habits 13




of the obese identified three specific patterns: night eating, binge eating and eating without satiation. Although there has been no continuation of this specific line of research (Wardle & Bienart, 1981), the term binge eating has become widely used.
Cbesity research has consistently sought to establish clear
differences between normal and overweight individuals with the goal of identifying the causative factors in eating disturbances. However, many of these attempts have been unsuccessful. Studies of the caloric intake of obese and normal weight groups (Johnson et al., 1956; Stefanik et al., 1959) have demonstrated that the obese may actually eat fewer calories per day than normal weight individuals. Hawkins (1979) has shown that obese college students consume fewer meals each day than normal weight students. Incidence of mental illness among the obese has been found to be the same or slightly less than the rate among normals (Holland, Masling & Copley, 1970; Leon, 1982; Moore, Stunkard & Srole, 1962) and no single personality abnormality has been found to describe the obese (Moore et al., 1962).
Binge eating, believed to be a behavior exclusive to the obese,
became an element of two prominent obesity theories, the psychosomatic and Schacterian hypotheses. Both theories explained obesity as a function of overconsumption, unrelated to hunger, that resulted in weight gain. The action of eating without hunger became the basis of many studies which, in time, would both support and refute the two hypotheses.
The psychoscaatic hypothesis (Kaplan & Kaplan, 1957) holds that non-nutritive eating is a learned coping response associated with anxiety reduction. Learned in childhood, the response is used by adults




faced with increased anxiety concerning aversive emotional states. Bruch (1961) further elaborated on the theory with her belief that the
obese have not learned to discriminate between the physiological symptcrns which accompany food deprivation (hunger) and the symptcms of arousal characteristic of emotions such as anger, anxiety and fear.
Although the psychosomatic concept of obesity has gained popular acceptance, there is limited research to support it. Leon and Chamberlain (1973a, 1973b) found that those unsuccessful in maintaining
weight loss reported greater consumption in response to feelings of loneliness, boredom and anger than those who maintained their weight loss. In a study using a diffuse free-floating arousal state, 5lochower (1976) found that the obese ate more than did normal weight subjects when they could not identify the cause of arousal and, in this condition, experienced affect reduction as a negative reinforcer for their eating behavior. Beyond these studies, there has been little demonstration of a unique obese eating pattern that is activated as a
response to stress. Further doubt of the psychosomatic hypothesis has been generated by Schacter's externality theory and the research that
supports it.
In a study of gastric motility, Stunkard and Koch (1964) used a gastric balloon to simulate stomach states of full and empty and the contractions associated with these states. They found that in sharp contrast to normal weight subjects, the obese showed little correspondence between the state of their stomachs and feelings of hunger. In a follow-up study, Griggs and Stunkard (1964) demonstrated that the obese could be trained to recognize gastric contractions, thus eliminating the possibility of physiological deficits. They also found




when physiological states were altered by inducing fear or preloading
the subjects with food, normal weight subjects reduced their intake but the obese did not. It was concluded from these two studies that internal states are irrelevant to the obese as triggers for hunger. Schacter (1967, 1971) further extends this conclusion and proposes that external cues are the significant determinant in the eating behavior of the obese. He posits that the obese, in comparison to normal weight individuals, are more sensitive to external cues such as the appearance and taste of food and less sensitive to internal cues, such as hunger and blood sugar level.
A number of early studies generated by Schacter's hypothesis, also known as the externality theory, support his contentions. Schacter, Goldman and Gordon (1968) showed that normal weight subjects ate significantly less when faced with anxiety and hypothesized that anxiety inhibits gastric contractions, releases sugar and inhibits appetite. The anxious obese showed a slight increase in eating further supporting the belief that they fail to respond to external states. In studies of the strength of external cues Ross (1974) found the appearance of food had a greater influence on the eating of the obese and Rodin (1975) found the time~ of day to be a more significant eating cue than hunger
for the obese. In an extension of the hypothesis, Schacter and Rodin (1974) proposed that the obese are also more susceptible than normal weight individuals to a wide range of external emotion-arousing stimuli and not solely limited in their externality to food cues.
Although the initial findings were supportive of the theory of externality, further research raised many questions about Schacter's hypothesis. Attempts to test one hypothesis against the other produced




mixed results. Meyer and Pudel (1972) found differences in stress-induced intake regardless of subject weight. Tom and Rucker (1975) found the obese more sensitive to external food-related cues but also sensitive to internal cues related to mood. The deprived obese subjects in their study showed low mood and decreased intake.
In a study combining appealing food and anxiety, McKenna's (1972) results partially support both hypotheses. He found a significant interaction between anxiety and weight level (the obese ate more when anxious; others ate less) but no evidence of anxiety reduction after consumption. No interaction was found between weight level and attractiveness of food. The attractive food influenced all groups to eat more, contrary to what Schacter would have predicted. Abramson and Stinson (1977) preloaded, until full, all of their subjects and confronted each with either a boring or interesting task. Their results showed the obese eating significantly more than normals in both conditions and that the boring task caused both obese and normals to consume more than their counterparts who completed the interesting task. These results are inconsistent with Schacterian theory which would predict that normal weight subjects would only respond to their physiological cues of being full fron the preload.
It has become increasingly clear that patterns of overconsumption such as binge eating are not the exclusive domain of the obese. Rodin and Slochower (1976) have shown that externality does not show a linear relationship to degree of excess body fat. Externality has been found in the non-obese (Rodin & Slochower, 1976; Meyer & Pudel, 1977). Pliner and Iappa (1978) using a mirror to increase subjects' awareness of their




18
intake found both normal and overweight subjects consumed less when able to observe themselves. However, the external cue of seeing another eat caused all subjects of all weights to increase their consumption (Rodin, 1977; Leon, 1982).
After two decades of attempts to differentiate the obese from the normal weight population in characteristics other than weight and calorie expenditure, few conclusions can be drawn. Bruch (1973) reported binge eating in a proportion of her obese patients but also found it among other weight groups. In a review of obesity studies, Leon (1982) reported that the obese eat faster, use fewer bites and tend to have poor body images, particularly if their obesity began in adolescence. Loro, Levenkron and Fisher (1979) believe that many of the more complicated aspects of obesity (e.g. the overweight individual's cognitions and skills) have received too little attention in the behavioral literature. Recently, two studies (Loro & Orleans, 1981; Gormally, Black, Daston & Rardin, 1982) have moved toward an examination and behavioral analysis of binge eating in obesity.
In a study of the contribution of binge eating to obesity, 280 participants in a behavioral weight loss program at Duke University Medical Center were observed and surveyed by Loro and Orleans (1981). Binge frequencies were reported as follows: 28.6% binged at least twice per week, 22.1% binged once per week and 20% reported no binges. No differences were found between men and women and early onset of obesity was found to be significantly associated with severity of binge eating. Amounts of food eaten during a binge ranged from 1,000 to 10,000 calories and took between 15-60 minutes to consume. Their subjects reported feelings of frustration, disappointment, boredan and rejection




19
before a binge. Based on their observations, Loro and Orleans believe binges are a negative reinforcer and extremely resistant to extinction.
Gormally et al. (1982) sought to assess the severity of binge eating amng the obese. They found that degree of obesity did not correlate with severity of binge eating and that 22% of their sample reported having no problem with binge eating. Through the use of the Binge Eating and the Cognitive Scales, they found that high dietary standards accompanied by a low belief in self-efficacy to adhere to a diet correlated highly with binge eating. Binges can then serve to reinforce the belief that stricter dietary controls are needed.
In conclusion, binge eating has been found to be a behavior
experienced by many but not all of the obese. Efforts to attribute binge eating solely to the obese and consider it an exclusive causative factor of the disorder have been unsuccessful. Reports of size, duration and frequency of binges by the obese are similar to those described in studies of under and normal weight individuals (Wardle & Bienart, 1981). Binge eating, as predicted by Bruch (1973), is a conmn~ disorder occurring across all weight classifications. Its investigation requires an examination of the cognitive and behavioral ccxrronalities that exist within both the obese and non-obese.
Binge Eating in Anorexia Nervosa
Reports of eating binges can be found throughout the literature on anorexia nervosa, usually in association with vaniting or sane form of purging (e.g., laxatives, diuretics). Although early reports of bingepurging in anorexia nervosa indicated the behavior was only an




occasional feature (Crisp, 1965; Russell, 1979), later findings have identified it in a significant proportion of cases. Halmi (1974) reported a 10 percent prevalence rate and Bruch (1973) found binging to be a characteristic of 22 percent of her patients. Beaumont et al. (1976) found binge behavior in 30 percent of their patients and Casper, Eckert, Halmi, Goldberg and Davis (1980) report that 46 percent of their anorexic patients show same form of binge characteristics. Although the increasing proportions may seem to indicate a growing problem of binge eating among the anorexic population, Wardle and Bienart (1981) point out that differences in diagnosis may explain the variability and that only in later studies (Beaumont et al., 1976; Casper et al., 196J) were patients directly questioned about vomiting. Despite the discrepant reports of prevalence, it is clear that a pattern of alternating dieting and binging (and often vomniting) exists in a substantial number of anorexics.
Although anorexia nervosa has been identified as an eating disorder for a number of years, there is confusion around the terms and diagnostic criteria used to describe it. As the investigation of such specific behaviors as binge eating, vomiting and purging began, many terms were used to describe subtypes ot the patient population. A distinction is frequently made between anorexics who reduce their weight by fasting and those who use other methods. The former are often referred to as
"dieters" (Beaumont et al., 1976) and/or "fasters" (Garfinkel, Moldofsky & Garner, 1980). Anorexics who rely on dietary restriction with binge eating and/or vomniting have been called "bulimic" (Casper et al., 1980) and "bulimic anorexic" (Garfinkel & Garner, 1982). The term "bulimia nervosa" (Russell, 1979) has been used to describe the form of




21
anorexia that includes singing. In 1980, DSM-III established bulimia as a distinct classification for individuals who exhibit abnormal consumption patterns but who are not 25 percent below their normal body weight. However, many authors still use the term, bulimic, as a
descriptor for any person who engages in ego-dystonic overconsumption.
Recognized as more than an occasional feature of anorexia nervosa, singing has been investigated by a number of recent studies. Beaumont
et al. (1976) found that vcmiter-purgers have a 43 percent likelihood of singing while dieters have only a likelihood of 28 percent. Vomiter-purgers were also found to be less inhibited, more likely to have had a past weight problem and less obsessional than dieters. The
prognosis for vomiter-purgers was considered worse than for dieters. In a test of body perception, anorexics who overestimated their own size were found to be more likely to binge (Button, Fransella & Slade, 1977). Dally and GCmez (1979), in an impressionistic account of 400 anorexics, state that singing begins in a subgroup of patients who cannot maintain their dietary restriction. This often occurs after nine months of dieting and is characterized by preferences for sweet, highly palatable foods. Bruch (1973) reports a similar pattern in her patients.
The identification of bulimic behavior in a proportion of the
anorexic population has brought a call by professionals to label bulimic anorexia as a distinct subgroup of anorexia nervosa (Russell, 1979; Casper et al., 1980; Garfinkel et al., 1980). Russell (1979) believes
binge eating may be a response of the hypothalamus to sub-optimal body weight. He found, in an investigation of 30 patients with a history of
binge-purge patterns, moderate levels of depression, preoccupation with food and body size, and significant weight fluctuations. Garfinkel and




Garner (1982) believe vomiting begins as a means of controlling the effects of singing but gradually becomes a reinforcer for binge eating. They believe that bulimic anorexics display confusion in the identification of their affective states and tend to misinterpret various unpleasant emotions as urges to eat. In support of Beaumont et
al. (1976), Garfinkel and Garner (1982) also found bulimic anorexics to have a worse long term prognosis and to be more outgoing than fasters.
Patients with anorexia nervosa have characteristics which clearly distinguish them fran others with eating disturbances. The lifethreatening nature of the weight loss, the unique population of young females affected and the morbid fear of obesity clearly identify the anorexic (DSM-III, 1980). However, the descriptions of dietary restraint and subsequent singing are similar to reports by obese and normal weight bingers (Wardle & Bienart, 1981), lending support to the belief that weight classification is a categorization that obscures the commronalities among all bingers.
Binge Eating in "Normal" Weight Individuals
It is possible that binge eating in normal weight individuals had been overlooked for so many years because these persons do not show any physical manifestations (e.g., obesity) of their behavior. Or perhaps, it was because we agree with Wardle and Bienart (1981), that cravings for food, dietary restriction and eating binges are not of themselves abnormal and this has caused us to ignore such variables as frequency and quantity that can indicate an eating disturbance. Kornhaber (1970) was one of the first practitioners to identify binge eating as a problem




among his patients and termed it, "the stuffing syndromee" He listed hyperphagia, depression and withdrawal as its main symptoms. Bruch's (1973) description of "thin fat people" was another early report of
binge eating in normal weight persons. Characterized by a preoccupation with food and weight and a pattern of fasting and binging often followed by self-induced voniting, "thin fat people" can appear to be anorexic in style but rarely suffer life-threatening weight loss. Bruch reports many of these patients have a history of obesity and she suspects that
the eating disturbance may be common among Americans.
In a survey of 279 college women (Ondercin, 1979), 78 percent
reported binge eating episodes and 18 percent identified themselves as compulsive eaters. This latter group of women showed a higher degree of dieting, more prior treatment for overweight and greater dissatisfaction with their present weight. Those high in compulsive eating reported increased eating in response to unpleasant emotional states and more frequent thoughts about food during the day. In a study to validate a self-report measure for binge eating, Hawkins and Clement (1980) surveyed 255 women and 110 men who were normal and overweight undergraduates. Their findings showed 79 percent of the women and 49 percent of the men engaged in binge eating with no differences between normal and overweight subjects. A significant partial correlation (r = .58, p < .001) was found between dietary concern and singing after weight was taken into consideration. Both surveys confirm Bruch's belief that there are widespread eating disturbances among the population of normal weight individuals.
Several authors have attempted to define and label the pattern of
disordered eating evidenced by normal weight binge eaters. Rau and




Green (1975) believe compulsive eating to be a neurological disorder, similar to epilepsy, that can be best described as ego-dystonic consumption. The term, "dietary chaos syndrome" was coined by Palmer (1979) to describe a disturbance characterized by a disordered eating pattern, preoccupation with food, eating and weight, and fluctuating body weight. He believed vomiting was used by sane to break the link between eating and weight gain.
Orbach (1978b) believes compulsive eating in women has critical social dimensions related to sex-role stereotypes and discrimination against women. She believes women who binge eat maintain overly strict dietary standards. Eventually, they lose control and binge. Fear of thinness is stated as a major causal factor for the behavior of the binger. Three types of non-physical hunger, prophylactic eating, mouth hunger and nervous eating, are described by Orbach (1978) as responsible for much excessive eating. Prophylactic eating occurs when food is available in the present but its future availability is unknown. Fearful of possible deprivation, the individual eats. Mouth hunger occurs when a non-hungry person sees or imagines the pleasure of tasting food and develops an urge to taste it. The consumption of food as a reaction to stress typifies nervous eating.
Boskind-Lodahl (1976) also offers a feminist perspective of eating disturbances and posits that binge-purgers have an exaggerated concept of femininity that produces unrealistic personal standards. The term, bulimarexia, was coined by Boskind-Lodahl and Sirlin (1977) to describe a syndrome characterized by secretive binges followed by fasting, vomiting or purging and a sense of shame and despair. In a pilot treatment of bulimarexia, Boskind-Lodahl and White (1978) identified




25
dietary restraint, societal pressure, lack of assertiveness and negative body image as factors contributing to the disturbance. Brenner (1980) and Kubistant (1982) have also adopted the term in their descriptions of college wcmen plagued by uncontrollable urges to eat in response to emotional stress.
With the publication of DSM-III (1980) by the American Psychiatric Association, bulimia was officially established as an eating disorder distinct from anorexia nervosa. The new classification incorporated the main features of the disturbances previously termed, "the stuffing syndrome" (Kornhaber, 1970), "dietary chaos syndrome" (Palmer, 1979) and "bulimarexia" (Boskind-Lodahl & Sirlin, 1977). Even with its established diagnostic criteria, bulimia remains a confusing term for many practitioners. It is freely used to describe both a symptom and a syndrome and, prior to its classification in DSM-III, there was little understanding of its role in normal weight individuals (Mitchell & Pyle, 1982).
The appearance of bulimia as an eating disorder category comes at a time when the problems of binge eating and such post-binge behaviors as vomiting and purging have become more evident. In a survey of 355 college students, Halmi et al. (1981) found 13 percent with the major symptoms of bulimia. This percentage represented 19 percent of the female sample and 5 percent of the males. Among the women, 35 percent labelled themselves "binge eaters" and 68 percent reported episodes of
binge eating.
Efforts to understand bulimia and its major characteristics such as binge eating, vomiting, fasting and purging have resulted in a number of recent descriptive studies. Pyle, Mitchell and Eckert (1981) report




that in an investigation of 34 cases they found their average patient had the disorder for four years before seeking treatment. They report most bulimics went on binges in response to stress, and found the behavior to interfere with their daily activities and to inhibit their interest in sexual activity. Pyle et al. (1981) believe the incidence of bulimia is under-reported and the severity of the disorder is minimized even when reported. In a study of the frequency and duration of binge eating in bulimics, Mitchell, Pyle and Eckert (1981) had 40 patients monitor their binge behavior for one week. They found that the average binge totalled 3,415 calories, occurred in the evening, was often followed by vomiting and consisted of highly caloric food.
A comprehensive survey of 316 cases of bulimia was conducted by Johnson, Stuckey, Lewis and Schwartz (1982). Of their sample, 96 percent were white, 83 percent had some post-secondary school education, 50 percent were students and 80 percent were unmarried. Average age of onset was 18; modal age of onset was 16. The average patient had the disorder for five years five months at the time of the investigation. Examining weight classifications of the subjects they found: 61 percent were normal weight, 17 percent were overweight and 21 percent were underweight. Half the sample reported a history of overweight and only
5 percent had a history of anorexia nervosa. Direct investigation of binge behaviors found 51 percent binge eating once each day or more and 78 percent binge eating in less than two hours. Examination of post-binge behaviors showed 63 percent using laxatives and 81 percent vomiting. Included in the preceding post-binge figures were the 26 percent ot the total sample who reported using both laxatives and vomiting after a binge. The quantity of food consumed during a binge




27
ranged fran 1,000-55,000 calories with the average being 4,800 calories and costing $8.30.
One important feature of the Johnson et al. (1982) survey was the report by 34 percent of the sample that binging began after the physiological and psychological deprivation of restrictive dieting. This supports Roth's (1982) belief that binge eating is a response to scarcity and that it is often the emotional and cognitive notion of deprivation that activates a "hunger from the heart." Johnson et al. (1982) in discussing the implications of their findings state:
For certain vulnerable individuals, it seems that prolonged
restrictive dieting may exacerbate a tendency toward buliic
behavior. It will be the task of future research to identify
biological and psychological risk factors which predispose
individuals to developing this type of behavior. (p. 14)
Dietary Restraint and Counter-regulation
In an extension of earlier research in obesity, a number of studies have examined the concept of dietary restraint across all levels of weight classifications. As measured by the Restraint Questionnaire (Herman & Polivy, 1975), dietary restraint refers to the extent to which a person demnstrates a concern for dieting by his/her attitudes and behaviors. In efforts to explain contradictory findings in obesity research, this concept of dietary restraint has been of great
assistance. The theoretical background behind it can be best understood after a consideration of two related bodies of research, Schacter's (1967) externality theory and Nisbett's set-point theory.
As discussed previously, Schacter's (1967) externality theory proposes that the obese are more responsive than normal weight individuals




28
to external cues concerning food and less sensitive to internal physiological states such as hunger. Rodin (1977) reports positive results in demonstrating that the time of day and the smell, taste and sight of food are more influential on the consumption patterns of the obese than of the non-obese. The obese have also been found to be more emotional and distractible (Pliner, Meyer & Blankenstein, 1974; Rodin, 1973). Schacter et al. (1968) found that anxiety would significantly decrease eating in normal weight individuals and caused a small, non-significant increase in consumption by the obese. Eating caused no anxiety reduction tor their obese subjects causing Schacter et al. to reject the psychosomatic hypothesis of obesity (Kaplan & Kaplan, 1957).
In 1972, McKenna, believing that the food (crackers) used in the
Schacter et al. (1968) was too neutral, repeated the design substituting chocolate chip cookies tor the crackers. Although just short of significance, the increase in eating shown by the obese was almost as strong as the decrease in consumption by normal weight subjects. McKenna, in partial support of Schacter et al., did not find food to act as an anxiety reducer. This support, although inconclusive, of the psychosamatic hypothesis together with studies questioning the adequacy of the externality theory (Tom & Rucker, 1975; Rodin, 1976; Leon & Roth, 1977) became the basis for such new hypotheses as set-point theory (Nisbett, 1972) and the concept of dietary restraint (Herman & Polivy, 1975).
Two studies (Nisbett 1968; Nisbett & Kanouse, 1969) introduced
Nisbett to the belief that the externality theory's explanation of the differences in the eating habits of obese and normal weight individuals was not sufficient. In a study of taste, deprivation and weight,




Nisbett (1968) found that "the overweight subjects exhibited a remarkable pattern of behavior. . They tended to eat either very large or very mnall amounts of food" (p. 114). He also found that individuals of normal weight who were once obese were as responsive to external cues as those presently obese and more responsive than normal weight, never obese individuals. In a later study of supermarket buying habits, Nisbett and Kanouse (1969) found that overweight subjects bought more and intended to buy more if they had recently eaten than if they had not. To explain these and other findings concerning the obese, Nisbett (1972) proposed set-point theory.
Set-point theory holds that the amount of stored body fat varies from person to person. Weight reduction reduces overall body size but not the number of fat cells in the body. Each person has a physiologically programmed level of fat tissue storage which is correct for them. This level, known as one's set-point, will be higher in sane persons than in others. The central nervous system tries to keep each individual at his/her proper set-point or base-line for fat storage. Overweight people are usually hungry because they are attempting to achieve a socially acceptable body size that is below their needs as determined by their set-point. Therefore, they are biologically underweight and hungry (Nisbett, 1972).
In extending set-point theory to Schacter's (1967) externality theory, Nisbett (1972) contends that externality occurs due to the conflict between the physiological need to maintain a higher weight and the social desirability of a lower weight. The obese person, or any person who tries to keep below set-point, finds him/herself more externally responsive because of unmet biological needs. Nisbett states




that the extremely obese who have stopped dieting achieve their set-point and are no longer overly responsive to external cues. Nisbett's theory remains unconfirmed (Wardle & Bienart, 1981) but it does provide a possible explanation for the contradictory findings in externality research. For example, Meyer and Pudel (1977) demonstrated externality in the latent-obese, those of normal weight who actively control their intake to maintain their size, and Rodin and Slochower (1976) found that externality did not show a linear relationship with adiposity (tat tissue). It also appeared in normal weight individuals, particularly those who controlled their weight.
In the initial investigation of dietary restraint, Herman ad
Polivy (1975) designed a study similar to McKenna's (1972) test of the Schacterian and psychoscmatic hypotheses. In that study, McKenna used the anticipation of mild or severe electric shock to create low or high anxiety and provided appealing and non-appealing food for his subjects to consume. His results showed a significant interaction for weight and anxiety with the obese eating more when nervous but with no evidence that differential anxiety reduction occurred. Schacter's theory was partially refuted in that all weight groups responded to the more appealing food. In keeping with set-point theory, Herman and Polivy hypothesized that dieting normal weight individuals (restrainers) would resemble behaviorally the obese who are concerned about their weight. In addition, the obese who have repudiated social pressure and attained their set-point would resemble the unrestrained normal weight individuals who are not food-deprived.
Herman and Polivy (1975) found no main effects for anxiety or restraint but did find a significant interaction between the two




conditions. Anxiety served as a mild cause of overconsum.nption in restrained eaters and significantly decreased consumption in unrestrained ones. When the overweight subjects were taken out of the analysis, the interaction emerged as stronger, further supporting dietary restraint and not weight status as the important factor in overconsumption. In consideration of Nisbett's (1972) theory, Herman and Rolivy raised the possibility that eating behavior differences may be dependent on the relative extent of underweight, and not overweight. They believed the measures on the Restraint Questionnaire may represent the relative deprivation of individuals, who regardless of actual weight restrict their food intake.
Prior to the research in dietary restraint, there was evidence (Wboley, 1972; Mahoney, 1975) that the individual's cognitive beliefs played a major role in the regulation of intake. Wooley (1972) found that belief about caloric intake was a more important determinant in subsequent overeating than actual calorie intake. Polivy and Herman (1976b) used this information to re-examine prior findings (Polivy & Herman, 1976a) on the effect of alcohol. In the first study, Polivy and Herman administered alcohol and a placebo, both disguised as Vitamin C, to restrained and unrestrained subjects. They found, contrary to expectation, that alcohol did not disinhibit the restrained subjects but actually caused them to eat less than after the placebo. In their second study, Polivy and Herman expanded the experimental design and correctly labeled the alcohol for half the subjects in each condition. This yielded a significant interaction, demonstrating that the restrained subjects required both a physiological change and appropriate cognitive labels before losing his/her eating inhibitions.




A further investigation of dietary restraint (Polivy & Herman,
1976) found that within a population diagnosed as depressed, restrainers gain weight during periods of emotional agitation but the unrestrained lose weight. The authors speculate that emotional stress may disrupt the strict standards that restrainers maintain. Polivy, Herman and Warsh (1978) compared the emotional responsiveness of restrained subjects to unrestrained ones in a study using a situation similar to one used with obese and normal weight individuals by Pliner, Meyer and Blankstein (1974). In addition, caffeine was administered to half the subjects to manipulate the state of internal arousal and a placebo was given to the other subjects. Similar to past studies of obese-normal differences, the results showed that in the placebo condition restrained subjects rated emotional stimuli more extremely than unrestrained subjects. However, in the caffeine condition, restrained subjects became less emotional and unrestrained subjects grew more emotional. The authors offer many possible explanations for this effect. One possibility is that restrained subjects have an elevated level of internal arousal, as is believed to be true with hyperactive children. When this level of arousal is increased (e.g. caffeine) it actually produces a decrease in emotional responding. Polivy et al. speculate whether unrestrained individuals, if aroused, would resemble restrained subjects.
Herman, Polivy, Pliner, Threlkeld and Munic (1978) in a study of distraction and emotional arousal, performed two experiments with restrained and unrestrained subjects. In the first a proofreading task was given once a month for three months. Similar to Rodin's (1973) findings with the obese, restrained subjects were more distractible in




33
the initial session. However, this effect did not persist over the next two sessions and restrainers became less distractible while the unrestrained grew more distractible. Herman et al. believed changes in arousal level may have accounted for the results and designed a second experiment to manipulate acute arousal. In support of past studies (Herman & Polivy, 1975; Polivy et al., 1978), the results showed restrained subjects to be more emotional in both the intensity and duration of their responses. Restrained subjects were found to be more distractible than unrestrained ones in calmi situations. This effect was reversed in periods of anxiety. Herman et al. raise the question of whether the restrained subject is more aroused in general or simply more arousable. They suggest that situational factors may control the behaviors of restrainers but that the externality this implies may be reversible given variations in arousal.
In another early investigation of dietary restraint, Herman and Mack (1975) hypothesized that there may be many normal weight individuals who are below set-point and have "latent" externality. They proposed that a better predictor of eating behavior would be found by measuring the level of dietary consciousness and restriction of their subjects. Using the Restraint Questionnaire (Herman & Polivy, 1975) to divide their subjects into restrained and unrestrained categories, Herman and Mack predicted that restrainers; would demonstrate counterregulatory eating. Counter-regulation is a laboratory phenanenon which has been likened to binge eating (Polivy, 1976; Spencer & Fremouw, 1979). It is a behavior where restrained (diet-conscious) individuals will consume more in a free eating situation after eating a preload of food believed to be high in calories than if they had not eaten anything




previously. Subjects found to be low in restraint characteristics (little preoccupation with food and weight) usually regulate their intake such that they eat less after an initial preload than if they had gone into the free eating period without eating anything.
In the Herman and Mack (1975) study, under the guise of a taste test, subjects were placed in one of three conditions: no preload, one milkshake preload or two milkshake preload. This was followed by a period during which the subjects could eat as little or as much as they wished. As expected, restrained subjects consumed more after a preload than without a preload. Unrestrained subjects showed an opposite effect, eating less after a preload to balance their intake. Herman and Mack found no significant differences between the obese and non-obese in restraint scores although overweight and restraint did correlate positively. They concluded that the differences in eating behavior previously attributed to the obese were more likely explained in terms of dietary restraint. Counter to what Herman and Mack expected, the two milkshake preload caused no more later consumption than did the one milkshake preload. They concluded that once the individual's restraint had been broken by the first drink the second was inconsequential to their counter-regulatory eating.
To gather greater support for the choice of dietary restraint as a main determinant of eating behavior, rather than weight classification, Hibscher and Herman (1977) repeated the design of Herman and Mack (1975) and tested three distinct weight groups: underweight, normal weight and obese. Each group was divided into restrained and unrestrained subjects. The results showed consumption following a preload varied as a function of restraint and not weight classification. In addition,




35
Hibscher and Herman examined levels of tree fatty acids (FFA's) in their subjects. Nisbett (1972) posited that elevated levels of FFA's in the bloodstream, generally accepted as characteristic of the obese, were actually a consequence of being below set-point (relative deprivation) and could be better predicted by a measure of this factor. In support of Nisbett's theory, Hibscher and Herman found elevated levels of FFA's to be associated with dieting across all weight groups. However, with stress correlated with FFA's in past studies, they speculate that the stress of dieting could account for the association.
Polivy (1976) demonstrated that counter-regulation is a cognitive effect in a study which varied the caloric content of the preloJd and the information given the subject about the preload's content. A statistical trend (p < .07) emerged indicating that restrained subjects ate more after ingesting a preload perceived to be highly caloric, even when it was not. In a study modifying the procedures of Polivy (1976) and Hibscher and Herman (1977), Spencer and Fremouw (1979) provided
substantial evidence for the importance of counter-regulation as a cognitive variable in the short-term regulation of food intake. Spencer and Fremouw preloaded female subjects who had been divided into three weight classifications. Half of each group were told the preload was high in calories; the other half were told the preload was low in calories and made from a new calorie-free sweetener. In actuality, all preloads were identical. Spencer and Fremouw's results showed that restrained subjects who were told the preload was highly caloric ate twice as much during the free-eating period than restrained subjects told the preload was low calorie. This interaction (restraint x




instructions) was significant. Weight level was not found to be a significant factor in any of the results. Unrestrained subjects were found to be unaffected by the instructions.
Lowe (1982) attempted to account for the counter-regulation
demonstrated by restrained subjects. She suggested that chronic dieting causes restrainers to feel that they must constantly deprive themselves of food. overeating after a preload occurs as a logical response to the anticipation that the diet will shortly return. Lowe called this situational factor, anticipated deprivation. She hypothesized that restrainers experience anticipated deprivation when their restrictions
have been loosened by eating "forbidden" food and they subsequently overeat. Non-restrainers perceive food as always available to them and do not experience this effect.
In Lo~we's (1982) study, restrained and unrestrained subjects were preloaded at the beginning of the experimental session. Half of each group were told that after the session they had to fast for four hours and return to ccznplete the experiment. other subjects were also told to cane back in four hours but to eat as they normally would throughout the
day. After these instructions were administered a free-eating period began under the guise of gathering baseline data. In actuality, the follow-up session never occurred and the dependent measure was the amount eaten during the free-eating period. Subjects were run in either a taste test condition or a learning task condition. This was to examine whether the breakdown of restraint would occur in a laboratory setting using a non-food related pretext.
Lowe's results showed anticipated deprivation as a significant main effect in the taste condition and as falling just short of significance




37
in the learning task. In the taste test situation a trend emerged for restrained subjects to consume more but this variable did not reach significance. Restraint was found to be a significant main effect in the learning task (p < .05). In neither condition was there an interaction between deprivation and restraint.
Lowe concluded that anticipated deprivation was a factor in cyclical overeating. The anticipation of a four hour period of restriction was enough to produce a significant amount of overeating on the part of the restrained eaters. Based on her results in the learning task experiment, Lowe believed the importance of restraint was reaffirmed in that it was a main effect in an experiment that was not done within a context of food or taste. Overall, Lowe concluded that both internal factors (e.g., restraint) and external factors (anticipated deprivation) must be taken into account when studying eating behavior. Lowe's design did not determine the relative importance of the preload in the breakdown of restraint in her subjects. She suggested future research to examine the influence of anticipated deprivation without the restraint breaking preload, believing that if anticipated deprivation alone could loosen restraint, the implications for dieters would be extensive.
It is clear frcn examining the work in dietary restraint and
counter-regulation that there are parallels to reports of binge eating by individuals of all weight classifications. Wardle and Bienart (1981) view dietary restraint as having a "causal role in relation to binges. However, its action is primarily as a background against which specific events can tip the balance into uncontrolled consumption" (p. 107). Research has shown preloads as a factor in counter-regulation. Lowe's




work demonstrates a role for anticipated deprivation in enhancing the effects of a preload. Polivy (1976) hypothesized that restraint was a cognitive factor based on a cammon-sense notion that individuals choose to diet. Similarly, outside the laboratory, dieters choose to taste food "forbidden" by their restrictive standards. The preload has been considered a mechanism that cognitively triggers the dieter to believe all control has been lost (Wardle & Bienart, 1981). Anticipated deprivation, as hypothesized by Lowe (1982), has yet to be tested as a sufficient cause for the breakdown of dietary restraint.
Treatment of Binge Eating
Although identified in 1959 by Stunkard, binge eating did not
become a tocus for treatment until recently. Prior to this time most therapists either agreed with Bruch's (1973) emphasis on psychotherapy which did not focus on the single behavior of binge eating or followed a standard behavioral management program as exemplified by the Stuart and D~avis (1972) model. Binge eating was primarily seen as a single eating behavior amidst many that contributed to weight gain.
Coutcome studies of weight loss programs and observations of dieters raised many questions about the success of behavioral methods. Hall and Hall (1974) questioned the long term maintenance of the lower weights and new behaviors achieved in behavioral weight loss programs. Wilson (1976) specifically pinpointed binge eating as a major canmplication in weight loss programs. He believes binge eating may be associated with
complex psychological and behavioral difficulties. Based on observations of his own clients, Wilson reported binge eaters are often




underassertive, harbor negative feelings toward themselves and are deficient in basic coping skills. Stunkard (1976) also pointed out binge eating as a block to weight loss and believes psychotheraphy needs to focus on the body image disturbances and actual binge behaviors of patients. Brightwell and Sloan (1977) in a review of long term studies in behavioral weight loss programs reported only 17 had a follow-up of 26 weeks or more and concluded in his review that behavioral programs
might only help a certain subgroup of patients.
There are many reports of weight loss treatment for the obese and same make specific reference to binge eating. Wollersheim (1970) found in a carnparison of different approaches that his "focal" therapy based on learning principles was effective in reducing eating in isolation and emotional/uncontrolled eating. Stalonas, Johnson and Christ (1978) examined ten aspects of a behavioral weight loss program and found only one variable, the control of impulse eating, to be correlated with weight loss. Overeaters Anonymous (R., 1979) has adapted a format similar to Alcoholics Anonymous and specifically focuses on canpulsive eating and the addictive nature of singing. Mahoney (1975) advocates the use of cognitive and behavioral methods with the obese. He believes their "cognitive claustrophobia," or dietary restriction, causes their cravings for forbidden food and he sees a need to use cognitive techniques to manipulate thoughts around food and dieting.
In a study of 280 participants in an intensive behavioral weight
loss program, Loro and Orleans (1981) conducted a behavioral analysis of binge eating in obese patients. Based on their observations, they
recormiended a number of behavioral treatment approaches and one quasi-behavioral strategy for the control of the antecedents and




40
consequences of binge eating. They suggested such methods as assertive training, anxiety management, relaxation and problem-solving training to help bingers cope with problems of anxiety, stress and poor social skills. They recommended thought-stopping to help with obsessions about food and cognitive restructuring to change irrational ideas about weight reduction. lb help control consequences, Loro and Orleans suggest aversion and punishment procedures, response cost strategies and the control of social reinforcement through the help of family and friends.
Programmed binging is reccamTended for clients who binge eat out of rebellion against authority figures, of which the therapist is one. It involves the use of paradoxical intention and is done under close therapeutic direction. Clients are instructed to plan and implement a binge, usually in a specified fashion. The benefits of this practice are many. Clients have the opportunity to analyze their binge behavior and, with the help of the therapist, restructure their thoughts around it. For those clients who have felt totally out of control, the programmned binge brings them more power in the situation and enhances feelings of self control. Programrmed binging can also serve to help the client view a single binge as a slip and not a total relapse.
With anorexics exhibiting binge purge behavior, Russell (1979) reccmmends hospitalization as a means to interrupt the cycle. This inpatient care is highly dependent on the ability of the nurses to work with the patient as attempts are made to achieve a pattern of regular meals and an acceptance of a higher weight. Monti, McCrady and Barlow (1977) report success using positive reinforcement, informational feedback and contingency contracting to reduce binge purging and increase body weight in a bulimic anorexic. Believing bulimic behavior




to be analogous to an obsessive-compulsive disorder, Rosen and Leitenberg (1982) treated a bulimic anorexic female with the behavioral techniques of exposure and response prevention. The patient was instructed that vomiting was an escape from binging which served to maintain binge behavior and that the relief from the physical discomfort of being too "stuffed" maintained vomiting. In treatment sessions, the patient was exposed to binging and the discomfort around it, but not allowed to vamit. Gradually the binge behavior, now without an escape mechanism, decreased as did anxiety toward food.
The growing public awareness of bulimia as an eating disorder has prompted new treatment strategies. Boskind-Lodahl and White (1978) reported a pilot group treatment approach termed "experientialbehavioral" that utilizes a feminist perspective. The format included assertiveness training, role-plays, sensory-awareness and guided fantasy exercises as well as such behavioral tasks as monitoring and contracting. Boskind-Lodahl and White reported dieting and restraint as leading contributors to the feelings of guilt and shame experienced by group members. Results from this pilot treatment showed success in 10 of 12 but follow-up results demonstrated a regression toward pretreatment levels. After a refinement of their program, Boskind-Lodahl and White's (1978) six month follow-up showed all fourteen subjects were able to eliminate vomiting, three stopped singing and seven decreased the number and durations of their binge episodes.
Fairburn (1981) has reported success with a cognitive-behavioral treatment with bulimics that he believes could also be helpful with overeaters. This approach seeks to immediately disrupt the binge-purge cycle through strict behavioral monitoring, frequent therapy sessions




and the inclusion of family and friends in treatment. Once reduced vomiting is achieved, Fairburn uses problem solving, cognitive restructuring and gradual exposure to forbidden food as ways to modify abnormal attitudes toward food, eating and body size. Leclair and Berkowitz (1983) also believe treatment success depends on an early reduction of binge-purge behavior. This in turn will help reduce depression. To Fairburn's program they add nutritional counseling, sensory awareness and examinations into the links between singing and emotional states. Kubistant (1982) uses similar strategies in his work with bulimics, often utilizing individual therapy and support groups concurrently. He advocates the use of programmed binges and eating practice sessions where clients ork on tasting food and on consuming it more slowly.
There are few treatment programs designed to specifically target binge eating. Green and Rau (1974) consider ccrnpulsive eating a neurological disorder and treated it with anti-convulsant medication. In a replication, Wermuth, Davis, Hollister and Stunkard (1977) failed to find any long term effectiveness of this treatment among their twenty patients. Orbach (1978a) has written extensively of a group treatment approach for compulsive eating which can be started among any group of women. She advocates the identification of the antecedents and consequences of eating, fantasy exercises, the development of hunger recognition and acceptance of one's body. Restrictive diets are seen as the polar opposite of binge eating and Orbach recommends termination of all diets while resolving the issues around compulsive eating. Roth (1982), influenced by Orbach's work, believes that physical and emotional deprivation causes singing and the only release from deprivation will




43
stop ccxnpulsive eating. She has developed a support group program, the Breaking Free Wo.rkshops, that seeks to examine the underlying personal needs that cause binge eating.
The findings in dietary restraint have begun to have an impact on treatment strategies for binge eaters. Wardle and Bienart (1981) favor a release fran overly restrictive diets and the irrational beliefs about intake which accomrpany them. They suggest that the dieter who binges when placed on any restricted regimen may be better served by an exercise program such as the one reported by Gwinup (1975). In his study of the singular effect of exercise on obesity, subjects lost an average of 22 lbs, in a one year period. Spencer and Fremouw (1979) in discussing their results in dietary restraint believe that cognitive techniques may be necessary to assist behavioral programs. Lowe (1982) also supports cognitive approaches and suggests the abandonment of inflexible dietary rules and the inclusion of menus that allow preferred foods.
Wardle and Bienart (1981) in their review of binge eating conclude
"If, as has been suggested, dietary restraint plays a major role in providing a setting within which binges are probable, treatment must aim to reduce restraint" (p. 107). Movement toward this goal requires a better understanding of the factors (e.g., anticipated deprivation) which create and influence restraint. Research efforts have just recently informed us that, contrary to the beliefs of most dieters, dieting causes binges. Successful treatment of binge eating in the future will occur only after an understanding of the many ccznplex factors involved.




This chapter has examined the theories and research findings that comprise the literature on binge eating. It is evident that binge eating and other related behaviors such as vomiting, purging and fasting have become common and problematic for many individuals in our society, particularly wxnen. Throughout the many reports of eating behavior frarn both experimental and descriptive studies certain themes emrge. This study is based on these themes and has the purpose of contributing new knowledge to the literature in binge eating.
In efforts to discover the causative factors in obesity, binge eating became a focus of investigation. It was believed that binge eating was a behavior that existed only within the behavioral repertoire of the obese and attempts to uncover reasons for binge eating and general overconsumption resulted in two prominent hypotheses, the psychosomatic and the Schacterian theory of externality (Schacter, 1967). The many conflicting findings that have arisen from attempts to prove and/or refute these hypotheses gave way to a belief that degree of overweight may not be the factor that best discriminates between binge eaters and normal eaters.
The concept of dietary restraint, as measured by the Restraint
Questionnaire (Herman & Polivy, 1975), has been shown in this review to be a viable and experimentally proven alternative. Studies described within this chapter have demonstrated that many tenets of both the externality and psychosomatic hypotheses are more strongly upheld by the concepts of dietary restraint and counter-regulation than by the original theories themselves. In addition to the concept of dietary




restraint, which is known to be a cognitive factor, a situational variable termed anticipated deprivation has been shown to contribute to overconsnption by binge eaters. It appears that not only do individuals with high restraint overeat after feeling they've "cheated" just a little but they further increase consumption when they anticipate having to return to being self-restrictive.
The issue of deprivation has appeared in several areas of the
literature review. Johnson et al. (1982) reported that 34 percent of the bulimics they surveyed began binge eating after a period of deprivation. Orbach (1977a) lists prophylactic eating, eating because future availability of food is unknown, as one of three types oft non-hunger related eating. Roth (1982) believes the individual will stop binge eating when his/her feelings of physical and psychological deprivation have been resolved. Nisbett (1972) claims that it is the physiological deprivation caused by constant dietary restriction that makes the individual more externally responsive to food cues and therefore, moure likely to binge eat. Bruch (1973) and Dally and Gcmiez (1979) have reported a similar pattern in anorexics.
The investigation of deprivation in the context of the dietary
restraint model of Herman and Rolivy (1975) is an appropriate direction for the examination of binge eating. Lowe (1982) has already demonstrated how the anticipation of deprivation affects those restrained individuals who have already broken their restraint through the consumption of "forbidden" junk food. This study will attempt to determine if the anticipation of deprivation is a powerful enough factor to cause those high restraint individuals to overeat without first externally disrupting their dietary restraint.




CHAPTER THREE
METHODOLOGY
The primary issue examined in this study was the effect of a
preload of food on the eating behavior ot restrained and unrestrained subjects who anticipate a period of fasting. Of secondary interest was the effect of the preload on the attractiveness of food (taste and buying intention) as reported by the subjects. In this chapter, the experimental methodology utilized in the study is described. The topics presented are the experimental hypotheses, the research design, subjects, the experimental procedures, instrumentation, data analyses,
rationale for methodology and limitations of the study.
Experimental Hypotheses
The following hypotheses, stated in null form, were tested within the methodology of this study:
1. There will be no differences between subjects in the four
groups (restrained preloaded, unrestrained preloaded, restrained non-preloaded, unrestrained non-preloaded)
in the amount of food consumed during the free eating period.
2. There will be no differences between subjects in the four
groups (restrained preloaded, unrestrained preloaded, restrained non-preloaded, unrestrained non-preloaded)
in their ratings of the taste of the food consumed.
46




3. There will be no differences between subjects in the
four groups (restrained preloaded, unrestrained preloaded,
restrained non-preloaded, unrestrained non-preloaded) in their ratings of buying intention for the food consumed.
The Research Design
The study utilized a 2 x 2 factorial design to test the above hypotheses. The factors in the design were level of restraint and presence of preload. Level of restraint (restrained or unrestrained) were assessed by a median split of scores from the Restraint Questionnaire (Polivy, Howard & Herman, 1976, Appendix A). Assignment into a preload condition (preloaded or non-preloaded) was done by random assignment fron the two respective restraint levels. All subjects experienced an anticipation of food deprivation by being requested to fast for four hours after the initial session.
The Restraint Questionnaire was administered to all female students enrolled in Introductory Psychology during the first week of the Fall 1983 semester. It was part of a battery of pre-tests being used in different studies within the Psychology Department. A total of 482 female students completed the Restraint Questionnaire. A median split was calculated fran the results o this testing of all potential subjects. Restrained eaters were defined as those scoring 16 or more on the Restraint Questionnaire and unrestrained eaters were those who scored less than 16. The use of a median split to separate restrained and unrestrained eaters was consistent with prior methods and yielded similar results (Herman & Mack, 1975; Herman & Polivy, 1975; Spencer &




F'remouw, 1979). Fran the pool of 482 students, 90 participated in the study. There were a total of 60 subjects in the four equal cells of the factorial design. In addition, 15 restrained and 15 unrestrained subjects %ere pre-tested using the Taste Preference Sheet (see Appendix C).
Since assignment into one of the two restraint levels was not random it was necessary to gather data on any potential differences between restrained and unrestrained subjects in their ratings of taste and buying intention. Duie to the nature of one of the independent variables (presence of preload), the subjects in the cells of the factorial design could not be pre-tested with a measure requiring consumption of food. Tb provide the necessary data, 15 restrained and 15 unrestrained subjects were tested with the Taste Preference Sheet only. This occurred prior to the testing of the subjects who comprised the four cells of the tactorial design.
Actual food consumption during the free eating period, a behavioral measure, served as the first dependent variable. Taste and buying intention, the other dependent variables, were assessed by the Taste Preterence Sheet, a thirty-item questionnaire of paired descriptors developed f or this study. Each dependent measure permitted the examination of possible main and interactive effects of the independent variables, level of restraint and presence of preload.
The design ot the study is graphically depicted as follows: Taste Preterence Pre-test
Restrained Unrestrained
Pre-test




Treatment
Restrained Unrestrained
Preload
No--Preload
Subjects
Subjects in this investigation were 90 female undergraduates enrolled in an Introductory Psychology course at the University of Florida. Students in this course were required to earn five experimental credits during the course of the semester through participation in studies approved by the Psychology Department. Students participating in this study received two credits toward their course requirement.
During the first week of the Fall semester, students in all
sections of Introductory Psychology were given a number of pretests from the various studies planning to use these students as subjects during
the semester. Included in this battery was the ten-item Restraint Questionnaire (Polivy et al., 1976, Appendix A). From this pretest, restraint scores for all female students enrolled in Introductory Psychology were compiled by the experimenter. A median split of scores divided the female students into unrestrained and restrained categories.
Thirty subjects (15 restrained, 15 unrestrained) participated in the Taste Preference Sheet pre-testing. Sixty subjects (15 per cell) comprised the four experimental cells of the factorial design. Any




female enrolled in Introductory Psychology who completed the pre-test battery during the first week of classes was eligible to participate in the study. Subjects elected to participate in the study through writing their instructor-assigned experimental number on a sign-up sheet that listed the times and locations of the experiment.
Before each subject arrived, she was assigned to a restraint level (restrained or unrestrained) based on her Restraint Questionnaire score in relation to the median split of scores gathered during Psychology Department pre-testing. In the Taste Preference pre-test, 15 restrained and 15 unrestrained subjects were tested. In the factorial design, 60 subjects were tested to meet the criterion of 15 per cell. Subjects in each restraint level were randomly assigned to either the preloaded or non-preloaded condition.
Experimental Procedures
Pre-testing with the Restraint Questionnaire
During the first week of classes, all students enrolled in
Introductory Psychology were administered a battery of instruments. These instruments were pre-tests from the various studies in the Psychology Department that would seek subjects from the pool of Introductory Psychology students. The battery was administered and collected during class time by a team of Psychology Department research assistants who visited each section of the course sometime during the first week of the semester. Students received one experimental credit for completing the battery. Upon administration of the battery to all




the enrolled students, the ccinpleted instrunents were returned to the investigators of the respective studies involved. Pre-testing with the Taste Preference Sheet
Subjects signed up for a study entitled "Sensation and Perception" and self-selected date and time tran the choices posted. Sessions were offered between 10:00 a.m. arnd 11:15 a.m. in order to provide pre-lunch ratings. Each subject was run individually and had no contact with other subjects participating in the study.
The session was described as an investigation of the taste perceptions of undergraduate females. All subjects were given a Taste Preference Sheet (see Appendix C) and asked to rate three types of food (M&M's, peanuts and animal crackers). The food was presented in three large bowls, each containing enough food to give the appearance that a substantial portion could be eaten before it would be visibly noticed. Subjects were left alone for fifteen minutes and a timer was set to allow the subject to anticipate the return of the experimenter. The size of the bowls and the timer served to provide a sense of privacy for the subject. The subject was asked to taste the three types of food and fill out the ratings sheet.
At the end of the 15-minute period the experimenter returned to
collect the ratings sheet and explained the purpose of the experiment to the subject.
Treatment Procedures
Subjects signed up for a study entitled "Sensation and Perception" and self-selected date and time from the choices posted. Each subject was asked to sign up for two sessions, the first lasting one hour and the second lasting 15 minutes. These sessions were separated by four




hours on the same day and this four hour period coincided with what is
normally considered lunchtime For example, a session beginning at 10:00 a.m. and had a follow-up session at 3:00 p.m., thereby running through the normal lunch period. Each subject was run individually and had no contact with other subjects participating in the study.
The experiment was described as an investigation of the effects of hunger on the perception of taste. All subjects were given a Marketing Questionnaire which asked for preferred shopping locations and familiarity with certain brand name products. While completing this questionnaire, those subjects assigned to the preload condition were required to consume ten M&M candies, ten peanuts and three animal crackers. This consumption was explained by the experimenter as a necessary preliminary to the later taste ratings. Subjects in the non-preloaded condition were not asked to consume any food while they
completed the survey.
After this initial survey period, each subject was again told that the purpose of the experiment was to assess the effect of hunger on taste and that they would make taste ratings in both sessions of the experiment. It was explained that in order tor the experimenter to measure the effect of hunger each subject must fast from~ the time the first session ended until she returned to their second session. At this time subjects were given the Taste Preference Sheet and asked to provide initial taste ratings of three types of food. The food was presented in three large bowls, each containing enough food to give the appearance that a substantial portion could be eaten before it would be visibly noticed. Subjects were instructed to eat as much or as little of the food as they wished during the 15-minute rating period and that




the experime~nter would return at the end of the 15-minute period to collect the rating sheets and to give the subject instructions for the afternoon session. Before leaving the roan the experimenter set a kitchen timer and turned it toward the subject to allow her to anticipate the experimenter's return.
The use of the large bowls of food and the setting of the timer were to provide the subject with as much perceived privacy as possible within the constraints of a laboratory experience. The large bowls were used to give the subject freedom to consume the food without concern for giving the appearance that she ate a large quantity in a short period of time. The timer allowed the subject to anticipate the experimenter's return and reduced worry about being interrupted in the middle of eating.
Upon returning at the end of the 15-minute period, the experimenter collected the ratings sheet and informed the subject that she was not required to return to the afternoon session. The actual purpose of the study was explained by the experimenter and any questions were answered. The subject was given a Subject Questionnaire (see Appendix D) to check if she intended to fast, if she had any prior knowledge of the true purpose of the study and if she had eaten iirnediately prior to attending the session. After collecting the Subject Questionnaire, the experimenter requested that the subject not discuss the study with any potential subject.
After the subject departed, the bowls of food wre weighed by the experimenter to assess how much was eaten during the 15-minute free eating period. The total amount consumed was used as the behavioral measure for one of the dependent variables of the study.




54
It was not necessary to eliminate any of the subjects who canmpieted the experiment. All subjects who participated and were asked to fast responded that they would have attempted to do so. No subject had eaten a meal within one hour of the start of the experiment nor did any report prior knowledge of the true purposes of the study. All required preloads were consumed by those subjects asked to do so.
Experimenters and Experimenter Training
The experimenters for this study were three female undergraduate research assistants who received credit for their work through the Psychology apartment course--PSY 3912, Independent Research. Each experimenter worked with 30 subjects, five subjects frcan each of the two pretest cells and the four experimental cells. Trhe experimenters were coordinated by a student in the graduate program in Counseling Psychology. This coordinator maintained records of the restraint scores and randcmly assigned subjects to one of the preload conditions. Prior to the arrival at a subject the experimenter was told to which condition of preload that subject had been assigned.
T7he experimenters were trained in the running of the experiment by the principal investigator during the first two weeks of the Fall 1983 semester. A general instruction sheet (see Appendix E) was used and included statements that were read to subjects during each phase of the experiment. Experimenters were instructed to limit their presentation to these statements whenever possible. Rocle-plays were utilized to practice the format of the study and uncover any potential problem areas.




Instrumentation
The Restraint Questionnaire
The Restraint Questionnaire was originally developed by Herman and Polivy (1975) to measure the extent an individual's attitudes and behaviors reflect a concern or preoccupation with dieting. In its original form it consisted of eleven items; the first six items examined diet and weight history and the last five examined concern with food and eating. In their initial use of the Restraint Questionnaire, Herman and Polivy found the scale to have substantial internal consistency (coefficient = .75). They report that the two subscales had internal consistency coefficients of .68 and .62, respectively.
In a later use of the scale, Hibscher and Herman (1977) found the test-retest reliability of the scale to be high (r = .92). In their study, Hibscher and Herman also used a physiological measure of deprivation, free fatty acids (FFA's), as a dependent measure. An analysis of variance, performed to determine whether level of restraint or obesity was a better predictor of FFA's, confirmed that restraint was the significant factor (F 1,80 = 8.54, p < .01). This report lends evidence to the construct validity of the scale.
wo studies (Polivy et al., 1976; Polivy, Herman, Younger &
Erskine, 1977) investigated the psychometric properties of the Restraint Questionnaire. The result was a refined instrument of ten items. This revised version has an internal consistency coefficient of .7 and a test-retest reliability coefficient of .9. The new questionnaire utilizes a forced-choice format and takes approximately 3-5 minutes to




56
complete. Subject scores may range from 0, showing low dietary concern, to 35, showing high dietary concern. Marketing Questionnaire
The Marketing Questionnaire (see Appendix B) is an instrument
developed for this study which requests information about the consumer habits of the subject. It was not intended to provide any information relevant to the purposes of this study. Its utility in this study was as a task for subjects to ccrnplete during the first phase of the study when one half of the subjects were eating a preload. Behavioral Measure
The amount of food (M&M's, peanuts and animal crackers) consumed during the free eating period served as the behavioral measure for the study. This was consistent with prior studies of counter-regulation. Herman and Mack (1975), Hibscher and Herman (1977), Spencer and Fremouw (1979), and [owe (1982) used the amount of highly caloric, easily ingested food as their primary measure. Lowe's study of anticipated deprivation, which forms the basis for this study, used M&M's, peanuts and animal crackers.
Taste Preferences Sheet
The Taste Preferences Sheet (see Appendix C) is an instrument
created for this study. It consists of thirty questions about the taste
and opinion of food. Subjects answered the same ten items for each of the three foods in the experiment. Only two of the ten items were actually used for the purposes of this study. These items were
3. How would you rate the taste of the food overall?
bad 1 2 3 4 5 6 7 8 9 10 great




6. How likely are you to buy this food or a similar food
item within the next two weeks?
very very
unlikely 1 2 3 4 5 6 7 8 9 10 likely
The remaining items were included to support the subject's belief that the experiment was testing the effect of hunger on taste.
The use of a preference scale is a ccnmon tool in consumer studies
(Kassarjian & Nakanishi, 1967). Tom and Rucker (1975) used a similar type of measurement in their study of the eftects of external cues on the consumption and buying intentions of the obese. In a study of selected opinion measurement techniques, Kassarjian and Nakanishi (1967) found no significant differences between the results derived by seven different methodologies (likert rating, two open choice options and four limited choice options).
Subject Questionnaire
The Subject Questionnaire (see Appendix D) is an instrument developed for this study to verify that the manipulations intended in the design of this study were actually present. It seeks information on whether the subject was blind to the true purpose of the study, if she intended to fast as requested, how long it had seen since her last meal and if she ate the preload if instructed to do so.
The Subject Questionnaire was used to eliminate subjects who do not conform to the necessary requirements of the study. A subject would have
been dropped fran the study for any one of the following reasons:
1. Subject was aware of the true purpose of the study.
2. Subject did not intend to fast and therefore did not
anticipate a period of deprivation.




3. Subject ate her last meal within one hour before the
start of the experiment.
4. Subject was assigned to the preload condition and did
not eat all of the preload.
Data Analyses
The data gathered from the pre-test of the Taste Preference Sheet were examined tirst. T-tests were used to investigate if the restrained and unrestrained subjects differed significantly in their ratings of taste and buying intention. When no differences were found, two-way (2 x 2)
analyses of variance were conducted on the two dependent variables, taste ratings and buying intention. Results front earlier research (Herman & Mack, 1975; Hibscher & Herman, 1977; Spencer & Fremouw, 1979) indicated that pre-testing of the dependent variable, food consumption, was not necessary. A two-way (2 x 2) analysis of variance was conducted on these
data.
Kerlinger (1973) points out that the ability to investigate possible interactions is one of the main advantages of the factorial analysis of variance. When an interaction was found between the two independent variables (level of restraint, presence of preload), an examination of simple main effects was conducted as described by Huck, Cormier and Bounds (1974).
Use of the F ratio occurs with the assumption of independent random samples, a normal distribution within the populations and equal variances. These criteria were met by this study. The F test is considered robust to possible inequalities in population variances when sample sizes are equal.




Rationale for Methodology
Wardle and Bienart (1981) point out that the laboratory research in counter-regulation has produced insight into binge eating. Herman and Mack's (1975) study of counter-regulation under the guise of a taste test has been a model successfully followed by a number of researchers (Polivy, 1976; Hibscher & Herman, 1977; Spencer & Fremouw, 1979) in their examinations of eating behaviors. In her study of anticipated deprivation, Lowe (1982) also tested the taste test context used in prior counter-regulation studies against an identical experiment with a non-food context. Lowe's purpose was to examine if the food-related cues implicit in a taste test influenced prior results. Her results found restraint to have a significant main effect on consumption even when the guise of a taste test was not used.
Throughout these studies and in other investigations of dietary
restraint (Herman & Polivy, 1975; Polivy & Herman, 1976b; Herman et al., 1978; Polivy et al., 1978) the Restraint Questionnaire has successfully predicted which subjects would demonstrate behaviors previously associated with the obese. The work of Polivy et al. (1977) has established the instrument as a reliable measure of dietary concern and the studies conducted by Hibscher and Herman (1977) and Spencer and Fremouw (1979) demonstrate that weight classification is not a significant factor in counter-regulation.
In the counter-regulation studies discussed above, subjects in each study were always of the same sex. This has been a standard procedure in studies of counter-regulation since scores on the Restraint Questionnaire have been found to be higher for females than for males (Polivy et al.,




60
1976). If both sexes were randcxuly included in a study, there would be a disproportionate number of males found in the unrestrained category after a median split was done. Therefore, this study limited itself to female subjects. The choice to use only females was made by the author in light of the greater numbers of females with eating disorders (DSM-III, 1980).
The research design in this study closely follows that utilized by Lowe (1982). A similar number of subjects were run and the preload consisted of the same foods in duplicate quantities. Lowe's selection of toods was supported by the findings of McKenna (1972) that demonstrated that neutral foods (e.g., crackers) are not always successful measures of true differences in overconsumption. To allow for ccrparisons to be drawn around the concept of anticipated deprivation, a similar period of fasting was requested ot the subjects in this study. In Lowe's study all subjects were preloaded and level of restraint and anticipated deprivation served as independent variables. In this study all subjects anticipated a period of deprivation and level of restraint and presence of preload were the independent variables.
Limitations of the Study
The use of the laboratory as a setting to test binge eating presents the most serious questions ot limitation in this study. Binge eating has been repeatedly found to be a behavior that is performed in private by the individual. In this study, individual running of subjects and an attempt to assure the subject's privacy (through the use of a timer and large food bowls) were included in the methodology to provide an optimal setting. The use of female undergraduates may limit the generalizability of the




61
results of the study. Although this particular group has been identified in many studies of binge eating, they tend to be a more physically active population and may not be ccmzparable to wcmen past their school years, many ot whcm may lead more sedentary lifestyles. The generalizability of the results of this study to a population of males cannot be assumed.




CHAPTER kWUR
RESULTS
The study was designed to investigate the effects of the presence or absence of a preload on restrained and unrestrained female subjects who anticipated a period of food deprivation. Three experimental hypotheses were formulated for the study. The first was constructed around the actual amount of food consumed by the subjects during a free-eating period. The second hypothesis was concerned with the taste ratings provided by the subjects. Ratings of food buying intention served as the dependent variable for the third hypothesis.
The use of a 2 x 2 factorial design produced four distinct groups of subjects. Restrained preloaded subjects were those who scored high on the Restraint Questionnaire (Polivy et al., 1976), indicating high preoccupation with food and body weight, and were given a preload of food prior to the free eating period. Unrestrained preloaded subjects were low scorers on the questionnaire who received a preload of food before the free eating period. Restrained non-preloaded subjects were high scorers who did not receive a preload and unrestrained non-preloaded subjects were low scorers who did not receive a preload.
The statistical tests used in the study were t-tests and two-way (2 x 2) analyses of variance. An alpha level of .05 was set as the basis for rejecting or failing to reject the null hypotheses. The remainder of this chapter provides a description of the results of the
62




testing of the hypotheses. Numbers appearing in the tables have been rounded off to the second decimal place.
Food Consumption
It was expected that all four groups would differ significantly
from each other on the dependent variable, food consumption. Restrained preloaded subjects were expected to consume the most food followed by restrained non-preloaded subjects. Unrestrained non-preloaded subjects were expected to rank third in consumption and unrestrained preloaded subjects were expected to consume the least. Thus, it was predicted that a main effect for restraint would be found.
These predictions were made with the belief that the condition of anticipated deprivation would cause restrained non-preloaded subjects to consume in a pattern different fran that found in prior studies. Previous research (Herman & Mack, 1975; Spencer & Fremouw, 1979) found that this group ate less than all other groups and, therefore, concluded
that the preload was a necessary factor in counter-regulatory eating. It was predicted in the present study that the anticipation of deprivation would cause all high restrainers to overconsume and that preloaded restrainers would consume the most. Thus, the following null hypothesis was tested:
1. There will be no differences between subjects in the four
groups (restrained preloaded, unrestrained preloaded,
restrained non-preloaded, unrestrained non-preloaded) in the
amount of tood consumed during the free-eating period.




A two-way (2 x 2) analysis of variance was conducted, using
restraint (restrained; unrestrained) as one independent variable and preload (preloaded; non-preloaded) as the other independent variable (see Table 1). The test found no significant main effects nor a significant interaction. With no significant differences found between any two groups and no significant interaction present, it was not possible to reject the null hypothesis.
An examination of the data (see Table 2) indicates that, as
expected, restrained preloaded subjects ate the most. The unrestrained
non-preloaded group ranked second, followed by the restrained non-preloaded group. The unrestrained preloaded subjects, as predicted, ate the least. Although this group ate a full one-third less than the restrained non-preloaded group, the large variations prevent this difference from reaching significance. It is notable that contrary to previous research the restrained preloaded group did not eat the least amount of food. This raises the possibility that the anticipation of deprivation may have had an effect on this group.
Taste and Buying Intention
Non-behavioral measures have not been included in prior studies of dietary restraint. These studies have relied on the amount of food consumed as the sole dependent variable. In this study, a cognitive factor, food attractiveness, was measured by ratings of taste and buying intention. It was predicted that the results of these ratings would parallel the expected results from the analysis of the amount of food consumption. Therefore, it was expected that restrained preloaded




Table 1
Analysis of Variance for Food Consunption

Source SS DF MS F P
Restraint (R) 2.20 1 2.20 1.02 0.31
Preload (P) 0.70 1 0.70 0.33 0.57
Restraint X
Preload 3.04 1 3.04 1.40 0.24
Error 121.30 56 2.17




Table 2
Means and Standard Deviations
of Food Consumption

Group N Mean ounces (rank) SD
Preloaded
Restrained 15 2.07 (1) 1.89
Unrestrained 15 1.23 (4) .96
Non-preloaded
Restrained 15 1.83 (3) 1.59
Unrestrained 15 1.90 (2) 1.28




subjects would rate the taste of the food higher and have greater intention to buy the food than would subjects in the other groups. Restrained non-preloaded subjects were expected to rank second on both of these dependent variables followed by unrestrained non-preloaded subjects. The unrestrained preloaded subjects were expected to provide the lowest ratings. Based on these expectations, it was predicted that a significant main effect for restraint would be found for both the taste and buying intention dependent variables.
In the study, assignment into one of the two restraint categories was based on the subject's score on a questionnaire. Due to this non-random selection, it was necessary to test additional subjects to investigate if any significant differences existed between restrained and unrestrained eaters in their ratings of taste and buying intention. Using the Taste Preference Sheet (see Appendix C) 15 restrained and 15 unrestrained subjects were tested. T-tests revealed no significant differences between the two groups for taste ratings (t(28) = .62, p > .5) nor for buying intention (t(28) = .91, p > .3). Based on this, it was considered reasonable to conduct two-way (2 x 2) analyses of variance on the ratings of taste and buying intention provided by the 60 subjects in the factorial design.
For the dependent variable, food taste, the following null hypothesis was tested:
2. There will be no differences between subjects in the four
groups (restrained preloaded, unrestrained preloaded,
restrained non-preloaded, unrestrained non-preloaded) in
their ratings of the taste of the food consumed.




Table 3
Analysis of Variance
for Taste Ratings

Source SS DF MS F P
Restraint (R) 30.82 1 30.82 2.87 0.10
Preload (P) 8.82 1 8.82 0.82 0.37
Restraint X
Preload 79.35 1 79.35 7.38 0.01
Error 602.00 56 10.75




A two-way (2 x 2) analysis of variance was conducted for taste ratings using restraint (restrained; unrestrained) as one independent variable and preload (preloaded; non-preloaded) as the other independent variable (see Table 3). No significant main effects were found. Contrary to expectations, a significant interaction between restraint and preload existed (F(l,56) = 7.38, p < .01), indicating that the difference between the levels of one factor changes frcm one level to another level (see Figure 1). In this case, restrained subjects scored higher than unrestrained subjects when both groups were preloaded but lower when neither group received a preload.
A simple main effects analysis (Huck, Cormier & Bounds, 1976)
indicated that the unrestrained preloaded group had significantly lower taste ratings than the unrestrained non-preloaded group (t(28) = 2.25, p < .05) and the restrained preloaded group (t(28) = 3.21, p < .01). Based on the statistical analysis, the null hypothesis of no differences between groups was rejected.
Further examination of the cell means revealed that the four
groups' ratings of taste ranked in the identical order to the ranking found for food consumption (see Table 4). Restrained preloaded subjects had the highest ratings of food taste followed by unrestrained non-preloaded subjects. Restrained non-preloaded subjects ranked third. The lowest ratings came from the unrestrained preloaded group, which differed significantly in the investigation of main effects.
For the dependent variable, buying intention, the following null hypothesis was tested:
3. There will be no differences between subjects in the
four groups (restrained preloaded, unrestrained preloaded,




Mean Ratings o Taste on Taste
Preference Sheet

15
10

Restrained Unrestrained

Preloaded Non-pre loaded
Level of Preload Condition
Figure 1
Graph of Restraint X Preload Interaction for Taste Ratings




Table 4
Means and Standard Deviations
of Taste Ratings

Group N Mean Scores (rank) SD
Pre loaded
Restrained 15 22.80 (1) 2.97
Unrestrained 15 19.07 (4) 3.39
Non-preloaded
Restrained 15 21.27 (3) 2.74
Unrestrained 15 22.13 (2) 3.91




Table 5
Analysis of Variance for Buying Intention

Source SS DF MS F P
Restraint (R) 19.27 1 19.27 0.60 0.44
Preload (P) 45.07 1 45.07 1.40 0.24
Restraint X
Preload 180.27 1 180.27 5.60 0.02
Error 1801.33 56 32.17




restrained non-preloaded, unrestrained non-preloaded) in their ratings of buying intention for the food consumed.
A two-way (2 x 2) analysis of variance was conducted for buying intention using the factors, restraint and preload (see Table 5). No significant main effects were found. Contrary to expectations, a significant interaction between restraint and preload existed (F(1,56) = 5.60, p < .05), indicating that the difference between the levels of one factor changes fran one level to another level (see Figure 2). As occurred with taste ratings, restrained subjects scored higher on buying intention than unrestrained subjects when both groups were reloaded and lower when neither group was preloaded.
A simple main effects analysis indicated that the restrained
non-preloaded group had significantly lower ratings of buying intention than the unrestrained non-preloaded group (t(28) = 2.20, p < .05) and the restrained preloaded group (t(28) = 2.29, p < .05). Based on the statistical analysis, the null hypothesis of no differences between groups was rejected.
As in the previous examination of cell means, restrained preloaded subjects ranked highest, reporting the greatest buying intention,
followed by the unrestrained non-preloaded group. However, the order was reversed for the other two groups. Unrestrained preloaded subjects ranked third in buying intention and restrained non-preloaded subjects had the lowest ratings of buying intention (see Table 6).




Table 6
Means and Standard Deviations
of Buying Intention

Group N Mean Scores (rank) SD
Preloaded
Restrained 15 15.40 (1) 6.32
Unrestrained 15 13.07 (3) 4.83
Non-preloaded
Restrained 15 10.20 (4) 5.28
Unrestrained 15 14.80 (2) 6.12




Mean Ratings of Buying Intention on Taste Preference Sheet

Unrestrained
Restrained

Preloaded Non-preloaded
Level of Preload Condition
Figure 2
Graph of Restraint X Preload Interaction for Buying Intention




The following list sumiarizes the results found in the study:
1. on the dependent variable, food consumption, no main effects were found for restraint or preload and no interaction of these two independent variables occurred.
2. on the dependent variable, taste rating, no main effects were found for restraint or preload.
3. A significant interaction (p < .01) between restraint and preload was found on the dependent variable, taste rating. The interaction indicates that preloaded restrained subjects rated the taste of food higher than non-preloaded restrained subjects and the opposite order occurred with unrestrained subjects. The unrestrained preloaded group was significantly different than the restrained preloaded and the unrestrained non-preloaded groups.
4. on the dependent variable, buying intention, no main effects were found for restraint or preload.
5. A significant interaction (p < .05) between restraint and preload was found on the dependent variable, buying intention. This interaction indicates that preloaded, restrained subjects had higher buying intention than non-preloaded restrained subjects and that the opposite order occurred with unrestrained subjects. The restrained nonpreloaded group was significantly different than the restrained
preloaded and unrestrained non-preloaded groups.




CHAPTER FIVE
SUMMARY, DISCUSSION, CONCLUSIONS,
IMPLICATIONS AND RECOMMENDATIONS
Sunmary
This study examined the effects of restraint level and presence of preload on the amount of food eaten and the ratings of food given by subjects expecting a period of food deprivation. Prior work with dietary restraint (Herman & Mack, 1975; Hibscher & Herman, 1977; Spencer & Fremouw, 1979) demonstrated that subjects with high dietary restraint consune more food in a taste test simulation if they have ingested a food preload perceived to be high in calories. Lowe (1982) preloaded all subjects in her investigation and found that those who anticipated a period of food deprivation ate significantly more than those who did not. She concluded that both internal factors (e.g., restraint level) and situational variables (e.g., anticipated deprivation) play a role in overconsumpt ion.
Lowe's (1982) study did not test if anticipation of deprivation was a sufficient factor to cause restrained eaters to overconsume without first ingesting a preload of restraint-breaking "forbidden" food. This question was of primary interest to this investigation. In this study, restrained and unrestrained subjects, all of whom were expecting a period of food deprivation, were randomly assigned to one of two preload conditions, preloaded or non-preloaded, creating four equal groups of subjects. The amount of food consumed during a free eating period was 77




the dependent variable measured. In addition, the following two cognitive dependent variables were examined; taste ratings and buying intention. Both were measured via the use of a self-report questionnaire (see Appendix C).
The experimental procedure utilized in the study was an extension of the methodology employed by Lowe (1982). Specifically, undergraduate females who had previously completed the Restraint Questionnaire (Polivy et al., 1976) signed up for an experiment titled "Sensation and Perception." Prior to the subject's arrival for the experiment it was noted whether she was a restrained or unrestrained eater (i.e., high or low preoccupation with food and body size) based on her Restraint Questionnaire score. Subjects were randomly assigned to one of two
preload conditions (preloaded; nonpreloaded) producing a 2 x 2 (Restraint X Preload) factorial design.
All subjects were asked to make a variety of food ratings after being informed that they must fast for the four hour period following the morning session of the experiment. They were left alone for 15 minutes during which time they could eat as much food as desired while they made their ratings. Those in the preloaded condition were given a standard amount of food to be eaten prior to this 15-minute free eating period.
Food consumption was measured by weighing the three bowls of food (M&M's, peanuts and animal crackers) before and after the free eating period and ccnputing the differences. Both the ratings of taste and buying intention were taken from items on the Taste Preference Sheet, a questionnaire developed specifically for this study. A preliminary testing of this instrument was conducted with 15 restrained and 15




79
unrestrained subjects to investigate any initial differences that might exist between restrained and unrestrained subjects. No significant differences were found.
For each of the dependent variables, a two-way (2 x 2) analysis of variance was conducted. No main effects were found for any of the three dependent variables. For two of the variables, taste ratings and buying intention, significant Restraint X Preload interactions were found. A test of simple main effects was conducted on each significant interact ion.
Discussion
The results of this experiment failed to support the prediction that there would be significant differences between the groups in the amount of food consumed during the free eating period. It was expected that restrained subjects in both levels of the preload condition (preload; non-preloaded) would consume significantly more than unrestrained subjects randomly assigned to the same levels. Contrary to expectation, no differences in consumption were found between the groups. Therefore, it cannot be concluded from the results of this study that, in the absence of a restraint breaking preload, restrained subjects who expect a period of food deprivation will break their dietary restraint and overconsume.
Past studies in dietary restraint found that the preload of
"forbidden" food was a necessary factor in causing restrained subjects to eat more than unrestrained subjects. Restrained subjects who did not ingest a preload were found to consume less than the unrestrained




subjects. This Restraint X Preload interaction was an indicator of counter-regulatory eating by restrained subjects. It has been hypothesized that restrained eaters maintain strict dietary standards until a "slip" occurs (e.g., ingestion of a highly caloric preload). After this initial "slip" occurs, restrainers abandon their standards and overconsume. If no dietary transgression occurs, restrained eaters remain faithful to their standards and undercons" e. Unrestrained eaters, with little concern about food, are found to naturally regulate their intake (e.g., unrestrained subjects who have ingested a preload eat less during the free eating period than when they have not received a preload).
Lowe (1982) preloaded all her subjects to test the effects of
restraint level and anticipation of deprivation on consumption. She
concluded that both independent variables were factors in the overconsumption exhibited by her subjects. Lowe hypothesized that in her taste test simulation the anticipation of deprivation may have been sufficiently powerful to obscure the effect of restraint, which failed to reach significance.
Based on these past results, it was expected that if anticipated deprivation was a sufficient factor in the breakdown of dietary restraint (i.e., a preload was no longer necessary), then all restrained subjects, preloaded and non-preloaded, would overconsume. Conversely, if anticipated deprivation was not a sufficient factor, then the preload would remain a necessary one, consistent with past findings. In this case, the restrained preloaded subjects would significantly overconsume and the non-preloaded subjects would consume the least. In actuality, the results of this study indicate that neither of these possible expectations occurred.




With no single group mean found to be significantly different,
anticipated deprivation cannot be judged, in the absence of a preload, to be a sufficient factor in overconsumption by restrained eaters. The same reason also prevents this study fromi supporting a reaffirmation of past studies' finding that a food preload is a necessary factor in causing counter-regulatory eating by restrained subjects. Counterregulatory eating would have been demonstrated had the restrained preloaded subjects eaten more than both groups ot unrestrained subjects and had the restrained non-preloaded group eaten less than all other subjects. In this study, restrained subjects ate no more or less than unrestrained subjects.
In seeking an explanation for these results, two avenues of
investigation must be pursued. First, did the experimental procedures successfully create the conditions desired for the study, and thereby, remain consistent with the methodology of past studies? Second, what factors may account for the results found in this study? A ccrnparison of this study with prior research may best facilitate this examination.
A review of the experimental methodology does not reveal any
procedural differences between the present study and past investigations of dietary restraint and anticipated deprivation that would affect the results unpredictably. The materials, time periods, instructions and
type of subject used in the experiment closely paralleled those of prior studies and attempted to replicate the relevant aspects of Lowe's (1982) work. Responses to the Subject Questionnaire (see Appendix D) did not
indicate that subjects were aware of the true purpose of the study. All subjects reported their good intentions to attempt to fast for the four hour period as requested.




In Lowe's (1982) study, all subjects were preloaded to ensure an initial breakdown of dietary restraint. In her taste test simulation she found a main effect for anticipation of deprivation and reported a trend for restraint level. Lowe hypothesized that her situational variable, anticipated deprivation, may have been powerful enough to obscure the usual effects of restraint level on preloaded subjects. After finding a significant main effect for restraint in her learning task simulation, Lowe concluded that overeating is better understood when both restraint and anticipation of deprivation are taken into account. She believed that if the restraint breaking preload could be excluded as a necessary factor the implications of her deprivation theory for restrained eaters would be more far-reaching.
In attempting to answer Lowe's question this study took one level of Lowe's independent variable, anticipated deprivation, and made it a fixed condition of the experiment. Preload, which had a fixed condition of Lowe's study, became an independent variable with two levels, preloaded and non-preloaded. Restraint level was continued as the second independent variable. Except for the introduction of anticipation of deprivation as a fixed condition, the present study's design closely paralleled the original research in dietary restraint and, thereby, created the desired experimental conditions.
Several possibilities exist to explain the results of the study. one is that anticipation of deprivation had no eftect on subjects and the lack of significant ditterences between the groups indicates that restraint level and preload are not factors in causing overconsumption. Given the extensive research in dietary restraint showing a significant Restraint X Preload interaction, this is an unlikely explanation of the results.




A second alternative which may, in part, explain the results is
that the groups did not differ because all subjects expected a period of food deprivation. If this alternative is true, then anticipation of deprivation could, as originally suggested by Lowe, be powerful enough to obscure the effects of restraint level. At the same time, the preload becomes unnecessary because the restrained subjects have already abandoned their strict dietary standards in response to the anticipated deprivation. This would imply that Lowe would have found the same results with subjects who were not given a preload.
This second alternative seems to be a more feasible explanation for the disparate results ot this study. However, the fact that Lowe did find a significant main effect for restraint warrants caution in accepting the above explanation of the results of this study. A method which permits a more thorough investigation of the premises of the two studies is discussed in the Recommendations section of this chapter.
Since previous studies had focused on food consumption as a
dependent variable, the investigation of the attitudes of subjects about the food consumed was an important addition to a study examining dietary restraint and anticipation of deprivation. The present study's examination of the two cognitive dependent variables, taste ratings and buying intention, generated results which parallel the findings of earlier studies of food consumption. No significant main effects Wre tound for the independent variables, restraint and preload, in previous studies nor were any found in the present study. However, both taste ratings and buying intention were found, in the present study, to be significantly affected by the interaction of the restraint and preload independent variables. In both cases, restrained subjects who were




preloaded had higher ratings of taste and buying intention than restrained subjects who were not preloaded, but unrestrained preloaded subjects had lower ratings of both variables than unrestrained subjects who were not preloaded.
The significant results for taste ratings and buying intention
indicate that even though subjects' eating behavior did not differ from each other, their thoughts about the attractiveness of food did. The differential effect of the preload on the cognitions of the restrained and unrestrained subjects seems inconsistent with the earlier hypothesis that the anticipation of deprivation obscured the effects of restraint level and preload. A re-examination of the experimental procedures may assist in resolving this contradiction.
Subjects provided ratings of taste and buying intention during the 15-minute free eating period. If they completed the questionnaire during the first 2-4 minutes of the period their responses might reflect the influence of restraint level and preload. For example, restrained non-preloaded subjects may enter the free eating period with the resolve to control their eating and complete the ratings which this resolve is still intact. However, as time passes and the period of fasting approaches, they abandon their restraint and consume in a similar fashion to preloaded subjects. The ratings of unrestrained subjects could also be influenced by the independent variables if these subjects ccmpleted the questionnaires before they finished consuming the food.
It was expected that the groups' ratings of taste would parallel the expected results of the food consumption variable. Restrained preloaded subjects were expected to provide the highest ratings of taste, followed by the restrained non-preloaded group.




Unrestrained non-preloaded subjects were expected to rank third and unrestrained preloaded subjects were expected to have the lowest ratings of taste.
The above ranking was predicted on the premise that when expecting a period of deprivation, restrained eaters, who usually view highly caloric food as desirable but forbidden, would rate the taste of food higher than unrestrained subjects. It was thought that the restrained preloaded subjects would have higher ratings than restrained nonpreloaded subjects because of the combined influence of the anticipated deprivation and the food preload on the former. Unrestrained preloaded subjects were expected to rank lowest because they would dislike being mandated to eat the preload and subsequently rate the taste lower than the unrestrained subjects who only ate during the free eating period.
The predicted main effect for restraint and the expected ranking of the four groups were not supported by the results of the taste ratings data analysis. Instead, a significant Restraint X Preload interaction was found. This interaction indicates that restrained subjects had higher taste ratings when preloaded than when they were not preloaded but unrestrained subjects had lower taste ratings when preloaded than when they did not receive a preload.
One possible explanation for the differential response exhibited by restrained and unrestrained subjects is that the preload caused different cognitions in the two types of eaters. Restrained subjects who have been forced to break their dietary restraint with the ingestion of the preload may attempt to justify this usually forbidden activity by elevating their taste ratings. Restrained non-preloaded subjects who are completing their ratings while freely sampling the food, may not




86
experience this need to justify their behavior. Unrestrained subjects may have a different reaction to the preload. With low preoccupation with food and body size, unrestrained subjects who are preloaded may rate the taste of food lower as a way to prevent overconsumption or as an indication of their displeasure about being mandated to eat. If this explanation was accepted as true, it would imply that restrained and unrestrained eaters may differ in their cognitions about food when they feel their ability to control their consumption has been diminished.
A test of simple main effects was conducted on the results of the taste ratings of the four groups. The unrestrained preloaded group was found to have significantly lower taste ratings than the restrained preloaded and the unrestrained non-preloaded groups. Two deductions may be drawn frcm this result. First, restrained and unrestrained subjects have different taste ratings when both are preloaded. Second, the taste ratings of unrestrained subjects who are preloaded differ fran those of unrestrained subjects who are not preloaded. Both deductions lend support to the above explanation of the results of the taste ratings.
The predicted main effect for restraint and the expected ranking of the four groups were not supported by the results of the buying intention data analysis. Instead, a significant Restraint X Preload interaction was found. This interaction indicates that restrained subjects who were preloaded had higher ratings of buying intention than
restrained non-preloaded subjects, but unrestrained subjects who were preloaded had lower ratings than unrestrained non-preloaded subjects. Contrary to the group rankings found with the consumption and taste rating variables, unrestrained preloaded subjects ranked third in buying intention and restrained non-preloaded subjects provided the lowest ratings of buying intention.




An important distinction between the cognitive variables taste
ratings and buying intention, is that buying intention represents what subjects believe they will do in the future, while taste ratings deal with a current state. Therefore, subjects are responding about a time frame that occurs long after the experiment and the period of deprivation have passed. However, this distinctive feature of the dependent variable does not provide a sufficient explanation for either the significant interaction or the group rankings.
one possible explanation for the buying intention results is that the restrained subjects exhibited a pattern similar to that which is hypothesized to occur in counter-regulatory eating. Restrained eaters, whose dietary restraint has been broken by the preload, abandon their strict standards and overconsume. Conversely, restrained eaters who do not receive preloads maintain their strict dietary standards and consume less than others. It is possible that in an analogous fashion, restrained preloaded subjects increased their ratings of buying intention while restrained non-preloaded subjects decreased their ratings. When preloaded, restrained subjects anticipating deprivation may feel out of control of their intake and expect a greater likelihood of purchasing forbidden food in the near future. Non-preloaded restrained subjects may react to the challenge of anticipated deprivation with decreased ratings of buying intention. For reasons similar to those proposed to explain the taste ratings results, unrestrained subjects may have responded differently. Disliking the mandatory preload or attempting to prevent a disruption ot their natural balance, the restrained preloaded subjects may have lowered their ratings of buying intention for the food consumed.




A test of simple main effects was conducted on the results of the buying intention ratings of the four groups. The restrained non-preloaded group was found to have significantly lower ratings of buying intention than the restrained preloaded and the unrestrained non-preloaded groups. This indicates that restrained subjects who were preloaded had significantly higher ratings of buying intention than restrained subjects who were not preloaded. It also shows that restrained subjects provided significatnly lower scores than unrestrained subjects when neither group was preloaded. These deductions support the contention of a differential effect of the preload on the buying intention of restrained and unrestrained subjects.
The results of this study shed some light on and generated many questions concerning the factors involved in overconsumption. One purpose of this study was to examine the necessity of a restraint breaking preload in causing overconsumption by restrained eaters who expect a preload of food deprivation. The lack of significant results for eating behavior in this investigation raises two possibilities. The first is that when subjects are expecting a period of food deprivation, the independent variables, restraint and preload, have no influence on subjects' consumption. This hypothesis seemed unlikely in consideration of the numerous studies showing significant results for these variables.
The second possibility is that anticipation of deprivation had an effect that cannot be measured in this study. One hypothesis is that the anticipation of deprivation was so powerful as to cause increased consumption in both restrained and unrestrained eaters. If this were to be supported in future research, this hypothesis would have important implications for the concept of dietary restraint and the treatment of binge eating.




The addition of the non-behavioral measures, taste ratings and
buying intention, revealed significant results and raised many questions concerning a new line of inquiry, the cognitions of subjects who break their dietary restraint. These results were found on the non-behavioral measures even though consumption was not found to differ between the groups. On both dependent variables, a significant Restraint X Preload interaction was present. Restrained subjects who were preloaded were found to have higher ratings of taste and buying intention than non-preloaded restrained subjects but unrestrained preloaded subjects had lower ratings than unrestrained subjects who were not preloaded.
one of the limitations of the concept of dietary restraint is that it established the food preload as a necessary factor in overconsumption by restrained eaters and does not offer an understanding of how restraint is broken outside the laboratory. Lowe's (1982) study introduced a situational variable, anticipated deprivation, and found that it was a factor to be considered in overconsumption. Investigation of the effects of such situational variables however, has exclusively focused on the behavioral measure, food consumption. The significant results found in this study for taste and buying intention indicated that investigation of the cognitions of the eater may also provide insight into factors causing differential beliefs about food and different patterns of consumption.
Conclusions
1. No significant differences in food consumption were found
between any of the tour groups (restrained preloaded, restrained nonpreloaded, unrestrained preloaded, unrestrained non-preloaded).




2. Restrained subjects provided higher ratings of taste when
preloaded than when they were not preloaded. Conversely, unrestrained subjects provided lower taste ratings when preloaded than when they were not preloaded. Unrestrained preloaded subjects had significantly lower taste ratings than subjects in both the restrained preloaded and the unrestrained non-preloaded groups.
3. Restrained subjects provided higher ratings of buying intention
when preloaded than when they were not preloaded. Conversely, unrestrained subjects provided lower ratings of buying intention when preloaded than when they were not preloaded. Restrained non-preloaded subjects had significantly lower ratings of buying intention than both the restrained preloaded and the unrestrained non-preloaded groups.
Implications
In failing to support the contention that the preload is not
necessary to cause overconsumption in restrained eaters who anticipate a period of deprivation, the results of this study limit the number and scope of implications that may be drawn by the investigator. It was expected that the results of the study would make the relationship between binge eating and food deprivation more understandable. Orbach (1978a) has posited that binge eaters overeat when future availability of food is unknown and Roth (1982) believes that binge eating is a response to scarcity. These similar hypotheses, which were based on clinical impressions, cannot be supported by the inconclusive findings of this study.




91
Despite the non-significant results found for food consumption, the significant findings for taste ratings and buying intentions do suggest that cognitions about food differed for scime subjects in the study. This finding implies that the inclusion of other dependent variables (e.g., cognitions about food) in future studies may provide a more
cinplete understanding of the ccinplex problem of dietary restraint.
Investigations into the cognitive components of overconsumption may also produce implications tor treatment. Wilson (1976) has questioned the efficacy of standard behavioral weight control programs for binge eaters. He suggests that these programs fail to intervene on a cognitive level and therefore, the thoughts that trigger a binge are never changed in therapy. The identification of those cognitive variables that influence overconsumption could assist in the development of more appropriate treatment strategies for binge eating.
Recaunendations for Future Research
1. one of the hypotheses generated to explain the results of this study is that the anticipation of deprivation experienced by all subjects obscured the usual effects found for the restraint and preload variables. The investigation of this question could be conducted through a combination of the methodology of this study and Lowe's (1982) work. A three way (2 x 2 x 2) factorial design that utilized anticipation of deprivation, restraint and preload as the independent variables would permit the analysis or the possible main and interactive effects of all three factors.




92
2. The reliance in earlier studies on one dependent variable (food consumption) limits the implications that can be drawn fran reseasrch in dietary restraint. The addition of two cognitive dependent variables in this study yielded significant results about the subjects' taste and buying preferences. In recognition of the common sense notion that individuals must decide to give up their restraint and overeat, it is reasonable to investigate the cognitions that accanmpany the behaviors exhibited by subjects on future studies.
3. If cognitive measures are included in future studies, it is reccmnended that their placement within the experimental procedures be carefully chosen. A questionnaire distributed at the beginning of the free eating period, for example, may receive different responses than one distributed at the end of the period, if subjects choose to break their dietary restraint during the free eating period.
4. The review of the literature conducted for this study found many studies examining the personality variables associated with obese eating styles. Similar investigations into the personality characteristics of restrained and unrestrained eaters have not been reported in the literature. The ccmnparison of the Restraint Questionnaire (Polivy et. al., 1976) with instruments that measure variables hypothesized to be influential in overconsumption (e.g., depression, locus of control, body image) may have valuable implications for treatment of binge eating.
5. Loes methodology allowed a situational variable usually experienced by dieters in daily life to be tested in analog fashion within the confines of a laboratory environment. Similar methods are needed to test other situational variables that may contribute to overeating (e.g., boredan, social anxiety, negative feedback).




APPENDICES




APPENDIX A
RESTRAINT QUESTIONNAIRE
1. How often are you dieting?
a) never b) rarely c) scmetimes d) usually e) always
2. What is the maximum amount of weight (in pounds) you have
ever lost within one month?
a) 0-4 b) 5-9 c) 10-14 d) 15-19 e) 20+
3. What is your maximum weight gain within a week?
a) 0-1 b) 1.1-2 c) 2.1-3 d) 3.1-5 e) 5.1+
4. In a typical week, how much does your weight fluctuate?
a) 0-1 b) 1.1-2 c) 2.1-3 d) 3.1-5 e) 5.1+
5. Would a weight fluctuation of 5 lb. affect the way you
live your life?
a) not at all
b) slightly
c) moderately
d) very much
6. Do you eat sensibly in front of others and splurge alone?
a) never b) rarely c) often d) always
7. Do you give too much time and thought to food?
a) never b) rarely c) often d) always
8. Do you have feelings of guilt after overeating?
a) never b) rarely c) often d) always
9. How conscious are you of what you're eating?
a) not at all
b) slightly
c) moderately
d) extremely
10. How many pounds over your desired weight were you at
your maximum weight?
a) 0-1 b) 1-5 c) 6-10 d) 11-20 e) 21+
Polivy, J., Howard, K. I., and Herman, C. P. (1976). Psychanetric analysis of the restraint scale. Unpublished manuscript. University of Tbronto.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E4K6V71R8_PJC5W3 INGEST_TIME 2017-07-11T22:07:05Z PACKAGE UF00101073_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ES6ZD9FGC_4TB2B9 INGEST_TIME 2017-11-20T17:27:26Z PACKAGE UF00101073_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

THE EFFECTS OF PRELIMINARY FOOD INTAKE AND DIETARY RESTRAINT DURING PERIODS OF ANTICIPATED DEPRIVATION By EDWARD T. SPAUSTER August, 1984 A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1984

PAGE 2

ACKNOWLEDGEMENTS In acknowledging those people who have assisted me in the completion of this study, I wish to thank the members of my doctoral committee. Dr. Dorothy Nevill has been a constant support and a trusted advisor since my admission to the university. Dr. Joseph Wittmer has provided leadership and counsel throughout the production of this dissertation. Drs. Ellen Amatea and William Froming have both shown support and guidance as instructors and committee members, and Dr. James Archer has given his time and energy on short notice. Dorene Tranes, Lynn Wigglesworth and Diane Comiter were the three undergraduate research assistants who gave long hours to gather the data for this study. Sarah Drew assisted in the coordination of their efforts. In addition, many people helped with the mechanics of producing a finished document and arranging a doctoral defense. Among them, Ces Bibby, Ann Cusick, Nancy Lubell, Joyce Bsrrotta, Len Travaglione, Curtis Walling and Flora Zaken-Greenberg stand out as particularly important. For her unconditional faith in the quality of my endeavors, I would like to thank my wife, Elizabeth Coburn. She is the single greatest support in my work and in my life.

PAGE 3

TABLE OF CONTENTS PAGE ACKNOWLEDGEMENTS ii ABSTRACT V CHAPTER ONE INTRODUCTION 1 Statement of the Problem 4 Need for the Study 5 Purpose of the Study 7 Significance of the Study 8 Definitions of Terms 10 Organization of the Study 12 TWO REVIEW OF RELATED LITERATURE 13 Binge Eating and Obesity 13 Binge Eating in Anorexia Nervosa 19 Binge Eating in "Normal" Weight Individuals 22 Dietary Restraint and Counter-regulation 27 Treatment of Binge Eating 38 Summary 44 THREE METHODOLOGY 46 Experimental Hypotheses 46 The Research Design 47 Subjects 49 Experimental Procedures 50 Instrumentation 55 Data Analysis 58 Rationale for Methodology 59 Limitations of the Study 60

PAGE 4

PAGE POUR RESULTS 62 Food Consumption 63 Taste and Buying Intention 64 Summary 76 FIVE SUMMARY, DISCUSSION, CONCLUSIONS, IMPLICATIONS, RECOMMENDATIONS 77 Summary 77 Discussion 79 Conclusions 89 Implications 90 Recommendations 91 APPENDICES A RESTRAINT QUESTIONNAIRE 94 B MARKETING QUESTIONNAIRE 95 C TASTE PREFERENCE SHEET 97 D SUBJECT QUESTIONNAIRE 100 E GENERAL INSTRUCTIONS FOR EXPERIMENTERS 101 F INFORMED CONSENT FORM 104 REFERENCES 105 BIOGRAPHICAL SKETCH 113

PAGE 5

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE EFFECTS OF PRELIMINARY FOOD INTAKE AND DIETARY RESTRAINT DURING PERIODS OF ANTICIPATED DEPRIVATION By Edward T. Spauster August, 1984 Chairperson: Dr. Paul J. Wittmer Co-Chairperson: Dr. Dorothy D. Nevill Major Department: Counselor Education This study examined the effect of level of dietary restraint and the presence or absence of a food preload on subjects expecting a four-hour period of food deprivation. A preload is a pre-determined amount of food administered to subjects prior to a free eating period (i.e., a period of time during which they may eat as much as they wish). Thirty restrained female subjects (high preoccupation with food and body size) and thirty unrestrained subjects (low preoccupation with food and body size) were randomly assigned to one of two preload conditions (preloaded; non-preloaded) Preloaded subjects were asked to consume a standard amount of highly caloric food prior to a period of time during which they could eat as much as desired. Non-preloaded subjects were not administered this standard amount of food. Under the guise of a study examining the effects of hunger on taste, all subjects were asked to fast for the four hours subsequent to the morning experimental

PAGE 6

session. After this request was made all subjects were given 15 minutes to privately consume the food presented (M&M's, peanuts and animal crackers) and provide ratings of taste and buying intention. The dependent variables measured were amount of food consumed, taste ratings and buying intentions. Analysis of variance found no differences between the four groups in the amount of food consumed. Significant Restraint X Preload interactions were found for taste ratings (p < .01) and buying intention (p < .05). This indicated that restrained preloaded subjects had higher ratings of taste and buying intention than restrained non-preloaded subjects but unrestrained preloaded subjects had lower ratings than those unrestrained subjects who were not preloaded. The prediction that restrained subjects expecting a period of food deprivation would overconsume in the absence of a restraint breaking preload was not supported. The results for the cognitive variables, taste ratings and buying intention, are discussed in light of this finding of no overconsumption. Other implications of the study and recommendations for future research are discussed.

PAGE 7

CHAPTER ONE INTRODUCTION Binges speak the voice of survival. They are protective mechanisms. Binges are signals that something is terribly wrong, that you are not giving yourself what you need — either physically (with food) or emotionally (with intimacy, work relationships). They are your last stand against deprivation. (Roth, 1982, p. 16) In an investigation of the eating patterns of the obese, Stunkard (1959) identified binge eating as a distinct eating disturbance in a percentage of his subjects. Described as an orgiastic consumption of large amounts of food in a relatively short period of time, binge eating has been the subject of many investigations into obesity. Two hypotheses, the psychosomatic and the Schacterian, have received much attention in these studies. The psychosomatic hypothesis (Kaplan & Kaplan, 1957) holds that non-nutritive eating is a learned coping response associated with anxiety reduction. Having learned this response in childhood, the overweight adult resorts to this behavior when faced with anxiety. The Schacterian hypothesis (Schacter, 1967) holds that the obese are more sensitive to external cues, such as the appearance and taste of food, and less sensitive to internal ones, such as hunger and blood sugar level. Both hypotheses have been supported and refuted in the literature (Herman & Polivy, 1975; Rodin, 1975; Schacter, 1967; Schacter & Rodin, 1974; Slochower, 1976). Efforts to incorporate both hypotheses and identify clear differences between obese and normal weight individuals 1

PAGE 8

2 have met with mixed results (McKenna, 1972; Ttm & Rucker, 1975) and studies of caloric intake differences between the two groups indicate that the obese may actually eat less than normal weight individuals (Johnson, Burke & Mayer, 1956; Stefanik, Heald & Mayer, 1959). In more recent years, binge eating has been identified as a problem behavior for some normal and underweight individuals, showing its more extreme forms in anorexia nervosa and bulimia (DSM-III, 1980). In light of these findings, it becomes purposeful to consider binge eating behavior across the spectrum of weight classifications and not as a behavior exclusive to the obese. Herman and Polivy (1975) have postulated that restraint (dietary consciousness) plays a role in overeating. In laboratory studies, subjects categorized as "restrained" (great preoccupation with body weight and food) have been found to consume more food in a taste test simulation after they had ingested a food preload perceived to be high in calories (Herman & Mack, 1975; Hibscher & Herman, 1977; Spencer & Fremouw, 1979). A preload is a pre-determined amount of food administered to subjects prior to a free eating period (i.e., a period of time during which they may eat as much as they wish) It appears that these dietary conscious individuals once they have indulged in forbidden food continue to eat rather than return to their strict standards of food intake. Subjects categorized as "unrestrained" (low preoccupation with body size and food) do not exhibit this phenomenon and decrease intake after a preload of food. This overconsumption (or, counterregulation) by restrained subjects has been likened to binge eating (Hibscher & Herman, 1977).

PAGE 9

3 Wardle and Bienart (1981) in a theoretical review of binge eating suggest that these studies confirm two ideas that have emerged from clinical literature. First, there is an association between binges and dietary restriction; and second, a binge can be triggered by the ingestion of a quantity of highly caloric or "forbidden" food. They further state, "it is clear that dietary restriction is implicated in a causal role in relation to binges. However, its action is primarily as a background against which specific events can tip the balance into uncontrolled consumption" (p. 107). Lowe (1982), in seeking to account for the pattern of overeating in restrained subjects, proposed that a situational factor, anticipated deprivation, influences a subsequent binge. Lowe's study demonstrated that after an ingestion of a food preload, restrained subjects who anticipated a four hour fast ate significantly more than their counterparts who were not expected to fast. Lowe suggests that restrained individuals, constantly aware of their eating habits, realize they will eventually return to dieting, a form of deprivation. Thus, when restraint is broken by a preload, they continue to eat before naving to return to a "tomorrow" of deprivation. In an area often left to postulation based on clinical observation, the concepts of dietary restraint and anticipated deprivation have been significant contributions to the study of eating behavior. Research has identified a population of dietary restrainers who, when given a preload of forbidden food, abandon their rules about dietirig and binge. The influence of anticipated deprivation after a preload further enhances this behavior. However, the role of anticipated deprivation before the

PAGE 10

4 rules have been broken remains a question for future research. If anticipated deprivation is found to play a significant part in the choice to begin an episode of overconsumption, the further understanding of this variable will provide insight into binge eating and the dynamics of weight loss. Statement of the Problem If researchers accept Wardle and Bienart's (1982) assertion that restraint serves primarily as a background for binge eating, it becomes the task of scientific inquiry to investigate what situational factors impact upon the overconsumption of food in the individual with high N restraint. Lowe (1982) in identifying anticipated deprivation as a variable that can be experimentally induced has begun this effort. However, Lowe's study conformed to prior restraint investigations in utilizing a preload of food to first break down the restraint of her subjects. Her results, therefore, reflect an understanding of anticipated deprivation only as it affects the already thwarted dieter. Investigators are left to speculate if anticipated deprivation is, in itself, a sufficient cause for the temporary abdication of dietary restraint. From our knowledge of the problems of the dieter, it is reasonable to assume that every binge or deviation from a diet does not result from the consumption of "forbidden" food accidentally or at the request of another. The factors which precipitate the restrained individual to choose to initially overeat remain, for the most part, uninvestigated. This study investigated the role of anticipated deprivation in the

PAGE 11

5 counter-regulation of food intake for restrained and unrestrained individuals with and without a preload of food and attempted to answer the following questions: 1. Is anticipation of a period of food deprivation a sufficient factor to cause the individual with a preoccupation with food and body size to overeat? 2. Does the ingestion of a preload that consists of food normally considered as highly caloric and "forbidden" differentially affect the later consumption and attitudes toward food of restrained and unrestrained individuals anticipating a period of food deprivation? Need for the Study In recent years, the quest to be thin has taken on the qualities of a national obsession. Repeatedly, consumers are bombarded with advertisements for diet aids, fast weightloss programs and reducing devices that promise a slender body, exciting social life and miraculous health benefits. This movement toward thinness has had a pronounced effect on women and our image of them. Garner, Garfinkel, Schwartz and Thompson (1980) report that American society has changed its conception of the ideal wcman to one who is taller, lighter and smaller proportioned. They point to a growing number of diet articles in American magazines as further evidence of the public's increased weight consciousness. This movement is further supported by industry's mass-marketing of health and beauty aids, professional sports and fashion for the fitness-minded.

PAGE 12

6 On the surface, this growing preoccupation with health and beauty appears in order for a population often viewed as overweight and overworked. It is not, however, without its consequences. Disturbances in eating have taken on epidemic proportions. Obesity presently affects 23%-68% of wcmen in the United States (United States Department of Health, Education and Welfare, 1967). Anorexia nervosa develops in one of every 250 females between 12 and 18 years old and claims the lives of 15%-21% of them (DSM-III, 1980). It is estimated that as many as 18% of college women meet the diagnostic criteria (DSM-III, 1980) of bulimia (Halmi, Falk & Schwartz, 1981). Binge eating has been identified as a behavioral response to food present in many American women regardless of weight classification or presence of a psychiatric disorder. Stunkard (1959) first coined the term from his studies of the eating patterns of the obese. Wilson (1976) discussed the failure of standard behavioral management programs in treating the obese binge eater. Several studies have examined binge-purge behavior in patients with anorexia nervosa (Beaumont, George & Smart, 1976; Bruch, 1973; Dally & Gcmez, 1979; Halmi, 1974). In 1980, the American Psychiatric Association established bulimia as a distinct eating disorder with uncontrolled binge eating as the primary characteristic (DSM-III, 1980). Orbach (1978) describes binge eating as a problem with many successful, middle class and well-educated women. Recent college surveys point to a high prevalence of binge eating among college students with as many as 35% of respondents in one survey labelling themselves "binge-eaters" (Halmi et al., 1981) and 78% of respondents in another reporting episodes of overconsumption (Ondercin, 1979).

PAGE 13

7 Wardle and Bienart (1981) state, "the division (within clinical literature) on the basis of weight has obscured some cannon ground in the eating disturbances experienced by these patients (obese, normal weight and anorexic" (p. 97). The concept of dietary restraint has proven to be a bridge across weight classifications. The preload of "forbidden" food has been repeatedly demonstrated as a trigger for overconsumption by the dietary conscious (Herman & Mack, 1975; Herman & Polivy, 1975; Polivy, 1976; Polivy & Herman, 1976, Hibscher & Herman, 1977; Spencer & Fremouw, 1979). The situational factor, anticipated deprivation, has been shown to exert an even stronger effect on preloaded subjects. Its effect on individuals who have not broken their strict dietary rules was the focus of this study. Purpose of the Study This study expanded upon the prior work in dietary restraint and anticipated deprivation. It examined a laboratory phenomenon, counterregulation, that has been likened to binge eating (Polivy, 1976; Spencer & Fremouw, 1979; Wardle & Bienart, 1981). It is a behavior where restrained individuals will consume more in a free eating situation after eating a preload of food believed to be high in calories than if they had not eaten anything previously. The effects of two independent variables, level of restraint and preload, were investigated in subjects expecting a period of food deprivation, as described by Lowe (1982). Level of restraint, an internal variable, was assessed as restrained or unrestrained by use of the Restraint Questionnaire (Polivy, Howard & Herman, 1976). Subjects were assigned to either a preload or a no preload condition. All experienced an anticipation of deprivation.

PAGE 14

8 The purpose was to assess the effects of the presence or absence of a preload on the consumption and attractiveness of food by restrained and unrestrained subjects who expected a period of food deprivation. Actual food consumption served as a behavioral measure, consistent with the measures used in previous restraint designs. Attractiveness of food was assessed by a questionnaire administered to the subjects. Each dependent measure permitted the examination of any existing main and interactive effects of levels of restraint and presence or absence of preload Significance of the Study Wilson (1976), in commenting on his treatment of six binge eaters, believes these clients differ sharply from the obese client upon whom the typical behavioral treatment program is based. He reports that two of his clients had little difficulty between binges with the standard behavior modification program of Stuart and Davis (1972) but failed to control the intensity or frequency of their binges. He agrees with Mahoney (1975) that cognitive variables exert a strong influence on eating behavior. An understanding of the factors that affect the relinquishment of restraint can be assisted by an examination of the role of anticipated deprivation. An identification of a component of the cognitive set that predisposes the binge eater to consume unwanted quantities of forbidden food will be useful in the construction of more effective behavioral treatment. Confirmation that anticipated deprivation exerts a strong influence on binge eating will question the wisdom of long-term weight loss programs and overly restrictive diets.

PAGE 15

9 Within the arena of the restraint research itself, it becomes important to assess if the ingestion of an initial preload is a necessary factor in binge eating. All studies to date have been based on this assumption. None have attempted to introduce an experimental variable to restrained subjects without a preload. If anticipated deprivation is found to be powerful enough to prompt the individual to overeat, it will invite the testing of other factors believed to influence binge eating. At present, reports on developing treatment programs for bulimia (Kubistant, 1982; Leclaire & Berkowitz, 1982; Boskind-Lodahl & White, 1978) indicate a substantial reliance on the clinical observations of the counselors providing therapy. Further understanding of the roles of restraint and anticipated deprivation can lend additional support to efforts to find adequate treatment methods for bulimia. The belief that the restrictive nature of dieting confounds the efforts of the female dieter to gain control of her eating disturbance is described in detail by Orbach (1978a). In treating compulsive eating disturbances in women she first prescribes a curtailment of all efforts to lose weight. Unfortunately, she reports no empirical evidence to support what may be an appropriate treatment strategy. The experimental investigation of restraint and anticipated deprivation will offer insight into this area. It will also respond to the call by Wardle and Bienart (1981) for an alliance between the clinicians and the experimentalists in our attempts to increase our knowledge.

PAGE 16

10 Definitions of Terms Anorexia Nervosa DSM-III (1980, p. 69) outlines the following diagnostic criteria: A. An intense fear of becoming obese, which does not diminish as weight loss progresses. B. Disturbance of body image, e.g., claiming to "feel fat" even when emaciated. C. Weight loss of at least 25% of original body weight or, if under 18 years of age, weight loss from original body weight plus projected weight gain from expected growth charts may be combined to make the 25%. D. Refusal to maintain body weight over a minimal normal weight for age and height. E. No known physical illness that would account for the weight loss. Anticipated Deprivation The expectation of a period of time during which individuals will be restricting their intake of food by either fasting or dieting. Binge The consumption of large quantities of food, unrelated to hunger, in a short period of time. Binge Eater An individual who exhibits a pattern of binge eating episodes. This pattern does not necessarily follow a particular periodicity. Bulumia DSM-III (1980, p. 70) outlines the following diagnostic criteria: A. Recurrent episodes of binge eating (rapid consumption of a large amount of food in a discrete period, usually less than two hours) B. At least three of the following: 1) consumption of high calorie, easily ingested food during a binge. 2) inconspicuous eating during a binge. 3) termination of such eating episodes by abdominal pain, sleep, social interruption, or self-induced vomiting. 4) repeated attempts to lose weight by severely restrictive

PAGE 17

11 diets, self-induced vomiting, or use of cathartics or diuretics. 5) frequent weight fluctuations greater than ten pounds due to alternating binges and fasts. C. Awareness that the eating pattern is abnormal and fear of not being able to stop voluntarily. D. Depressed mood and self-deprecating thoughts following eating binges. E. The bulimic episodes are not due to Anorexia Nervosa or any known physical disorder. Counter-regulation A laboratory phenomenon, similar to binge eating, where subjects categorized by the Restraint Questionnaire (Polivy et al. 1976) as restrained (high preoccupation with body weight and food) consume more after eating a preload of food believed to be high in calories than if they had not eaten anything previously. Obesity A weight classification categorizing those individuals who weigh more than 15% of the acceptable weight range for their sex, age, height and body frame size. Preload An amount of food administered to subjects for their consumption prior to a free eating period during which the subjects will be allowed to eat as much as they wish. Preloads are often used to guarantee the subject is full before the experimental manipulation is administered. They may also be used to break down a subject's resistance to eating food that is high in calories. Restraint A level of dietary consciousness as measured by the Restraint Questionnaire (Polivy et al., 1976). It assesses the extent to which individuals exhibit attitudinal and behavioral concern about dieting.

PAGE 18

12 Organization of the Study The remainder of this study is organized into four chapters. The second chapter is a review of the related literature. Included in this chapter are considerations of 1) binge eating in obesity, anorexia nervosa, and normal weight ranges, 2) the concept of dietary restraint in binge eating, and 3) current treatment strategies for binge eating. The third chapter presents the research methodology. The fourth chapter contains the results of the study. A summary of the study, which includes a discussion of its implications and suggestions for future research, comprises the final chapter.

PAGE 19

CHARIER TWO REVIEW OF RELATED LITERATURE The literature review in this chapter consists of six sections and focuses on the theories and research which form the basis for this study. The first three sections examine, separately, the role of binge eating in obesity, anorexia nervosa and "normal" weight groups. Research efforts in eating behavior have often focused on a specific weight group, e.g., obese individuals, and compared its members to a control group of normal weight subjects. As this reviewer will discuss, this division by weight classification has not proved to be an adequate method of describing differences in eating behaviors. However, in light of the quantity of studies that have differentiated by weight, this division serves as a useful organizational tool. The fourth section presents the findings in the area of counterregulation and the role of dietary restraint in this behavior. These studies form the basis for the experimental methodology that will be utilized in this study. The fifth section discusses the current strategies and concerns in the treatment of binge eating. The last section is a summary of the reviewed research and the implications for the study. Binge Eating and Obesity The earliest behavioral reports of binge eating are to be found in studies of the obese. Stunkard's (1959) analysis of the eating habits 13

PAGE 20

14 of the obese identified three specific patterns: night eating, binge eating and eating without satiation. Although there has been no continuation of this specific line of research (Wardle & Bienart, 1981), the terra binge eating has become widely used. Obesity research has consistently sought to establish clear differences between normal and overweight individuals with the goal of identifying the causative factors in eating disturbances. However, many of these attempts have been unsuccessful. Studies of the caloric intake of obese and normal weight groups (Johnson et al., 1956; Stefanik et al., 1959) have demonstrated that the obese may actually eat fewer calories per day than normal weight individuals. Hawkins (1979) has shown that obese college students consume fewer meals each day than normal weight students. Incidence of mental illness among the obese has been found to be the same or slightly less than the rate among normals (Holland, Masling & Copley, 1970; Leon, 1982; Moore, Stunkard & Srole, 1962) and no single personality abnormality has been found to describe the obese (Moore et al., 1962). Binge eating, believed to be a behavior exclusive to the obese, became an element of two prominent obesity theories, the psychosomatic and Schacterian hypotheses. Both theories explained obesity as a function of overconsumption, unrelated to hunger, that resulted in weight gain. The action of eating without hunger became the basis of many studies which, in time, would both support and refute the two hypotheses The psychosomatic hypothesis (Kaplan & Kaplan, 1957) holds that non-nutritive eating is a learned coping response associated with anxiety reduction. Learned in childhood, the response is used by adults

PAGE 21

15 faced with increased anxiety concerning aversive emotional states. Bruch (1961) further elaborated on the theory with her belief that the obese have not learned to discriminate between the physiological symptoms which accompany food deprivation (hunger) and the symptoms of arousal characteristic of emotions such as anger, anxiety and fear. Although the psychosomatic concept of obesity has gained popular acceptance, there is limited research to support it. L^on and Chamberlain (1973a, 1973b) found that those unsuccessful in maintaining weight loss reported greater consumption in response to feelings of loneliness, boredom and anger than those who maintained their weight loss. In a study using a diffuse free-floating arousal state, Jlochower (1976) found that the obese ate more than did normal weight subjects when they could not identify the cause of arousal and, in this condition, experienced affect reduction as a negative reinforcer for their eating behavior. Beyond these studies, there has been little demonstration of a unique obese eating pattern that is activated as a response to stress. Further doubt of the psychosomatic hypothesis has been generated by Schacter's externality theory and the research that supports it. In a study of gastric motility, Stunkard and Koch (1964) used a gastric balloon to simulate stomach states of full and empty and the contractions associated with these states. They found that in sharp contrast to normal weight subjects, the obese showed little correspondence between the state of their stomachs and feelings of hunger. In a follow-up study, Griggs and Stunkard (1964) demonstrated that the obese could be trained to recognize gastric contractions, thus eliminating the possibility of physiological deficits. They also found

PAGE 22

16 when physiological states were altered by inducing fear or preloading the subjects with food, normal weight subjects reduced their intake but the obese did not. It was concluded from these two studies that internal states are irrelevant to the obese as triggers for hunger. Schacter (1967, 1971) further extends this conclusion and proposes that external cues are the significant determinant in the eating behavior of the obese. He posits that the obese, in comparison to normal weight individuals, are more sensitive to external cues such as the appearance and taste of food and less sensitive to internal cues, such as hunger and blood sugar level. A number of early studies generated by Schacter' s hypothesis, also known as the externality theory, support his contentions. Schacter, Goldman and Gordon (1968) showed that normal weight subjects ate significantly less when faced with anxiety and hypothesized that anxiety inhibits gastric contractions, releases sugar and inhibits appetite. The anxious obese showed a slight increase in eating further supporting the belief that they fail to respond to external states. In studies of the strength of external cues Ross (1974) found the appearance of food had a greater influence on the eating of the obese and Rodin (1975) found the time of day to be a more significant eating cue than hunger for the obese. In an extension of the hypothesis, Schacter and Rodin (1974) proposed that the obese are also more susceptible than normal weight individuals to a wide range of external emotion-arousing stimuli and not solely limited in their externality to food cues. Although the initial findings were supportive of the theory of externality, further research raised many guestions about Schacter' s hypothesis. Attempts to test one hypothesis against the other produced

PAGE 23

17 mixed results. Meyer and Pudel (1972) found differences in stressinduced intake regardless of subject weight, lorn and Rucker (1975) found the obese more sensitive to external food-related cues but also sensitive to internal cues related to mood. The deprived obese subjects in their study showed low mood and decreased intake. In a study combining appealing food and anxiety, McKenna's (1972) results partially support both hypotheses. He found a significant interaction between anxiety and weight level (the obese ate more when anxious; others ate less) but no evidence of anxiety reduction after consumption. No interaction was found between weight level and attractiveness of food. The attractive food influenced all groups to eat more, contrary to what Schacter would have predicted. Abramson and Stinson (1977) preloaded, until full, all of their subjects and confronted each with either a boring or interesting task. Their results showed the obese eating significantly more than normals in both conditions and that the boring task caused both obese and normals to consume more than their counterparts who completed the interesting task. These results are inconsistent with Schacterian theory which would predict that normal weight subjects would only respond to their physiological cues of being full from the preload. It has become increasingly clear that patterns of overconsumption such as binge eating are not the exclusive domain of the obese. Rodin and Slochower (1976) have shown that externality does not show a linear relationship to degree of excess body fat. Externality has been found in the non-obese (Rodin & Slochower, 1976; Meyer & Pudel, 1977). Pliner and lappa (1978) using a mirror to increase subjects' awareness of their

PAGE 24

18 intake found both normal and overweight subjects consumed less when able to observe themselves. However, the external cue of seeing another eat caused all subjects of all weights to increase their consumption (Rodin, 1977; Leon, 1982). After two decades of attempts to differentiate the obese from the normal weight population in characteristics other than weight and calorie expenditure, few conclusions can be drawn. Bruch (1973) reported binge eating in a proportion of her obese patients but also found it among other weight groups. In a review of obesity studies, Leon (1982) reported that the obese eat faster, use fewer bites and tend to have poor body images, particularly if their obesity began in adolescence. Loro, Levenkron and Fisher (1979) believe that many of the more complicated aspects of obesity (e.g. the overweight individual's cognitions and skills) have received too little attention in the behavioral literature. Recently, two studies (Loro & Orleans, 1981; Gormally, Black, Daston & Rardin, 1982) have moved toward an examination and behavioral analysis of binge eating in obesity. In a study of the contribution of binge eating to obesity, 280 participants in a behavioral weight loss program at LXike University Medical Center were observed and surveyed by Loro and Orleans (1981). Binge frequencies were reported as follows: 28.6% binged at least twice per week, 22.1% binged once per week and 20% reported no binges. No differences were found between men and women and early onset of obesity was found to be significantly associated with severity of binge eating. Amounts of food eaten during a binge ranged from 1,000 to 10,000 calories and took between 15-60 minutes to consume. Their subjects reported feelings of frustration, disappointment, boredom and rejection

PAGE 25

19 before a binge. Based on their observations, Loro and Orleans believe binges are a negative reinforcer and extremely resistant to extinction. Gormally et al. (1982) sought to assess the severity of binge eating among the obese. They found that degree of obesity did not correlate with severity of binge eating and that 22% of their sample reported having no problem with binge eating. Through the use of the Binge Eating and the Cognitive Scales, they found that high dietary standards accompanied by a low belief in self-efficacy to adhere to a diet correlated highly with binge eating. Binges can then serve to reinforce the belief that stricter dietary controls are needed. In conclusion, binge eating has been found to be a behavior experienced by many but not all of the obese. Efforts to attribute binge eating solely to the obese and consider it an exclusive causative factor of the disorder have been unsuccessful. Reports of size, duration and frequency of binges by the obese are similar to those described in studies of under and normal weight individuals (Wardle & Bienart, 1981). Binge eating, as predicted by Bruch (1973), is a common disorder occurring across all weight classifications. Its investigation requires an examination of the cognitive and behavioral commonalities that exist within both the obese and non-obese. binge Eating in Anorexia Nervosa Reports of eating binges can be found throughout the literature on anorexia nervosa, usually in association with vomiting or some form of purging (e.g., laxatives, diuretics). Although early reports of bingepurging in anorexia nervosa indicated the behavior was only an

PAGE 26

20 occasional feature (Crisp, 1965; Russell, 1979), later findings have identified it in a significant proportion of cases. Halmi (1974) reported a 10 percent prevalence rate and Bruch (1973) found Dinging to be a characteristic of 22 percent of her patients. Beaumont et al. (1976) found binge behavior in 30 percent of their patients and Casper, Eckert, Halmi, Goldberg and Davis (1980) report that 46 percent of their anorexic patients show some form of binge characteristics. Although the increasing proportions may seem to indicate a growing problem of binge eating among the anorexic population, Wardle and Bienart (1981) point out that differences in diagnosis may explain the variability and that only in later studies (Beaumont et al., 1976; Casper et al., 1980) were patients directly questioned about vomiting. Despite the discrepant reports of prevalence, it is clear that a pattern of alternating dieting and binging (and often vomiting) exists in a substantial number of anorexics. Although anorexia nervosa has been identified as an eating disorder for a number of years, there is confusion around the terms and diagnostic criteria used to describe it. As the investigation of such specific behaviors as binge eating, vomiting and purging began, many terms were used to describe subtypes of the patient population. A distinction is frequently made between anorexics who reduce their weight by fasting and those who use other methods. The former are often referred to as "dieters" (Beaumont et al., 1976) and/or "rasters" (Garfinkel, Moldofsky & Garner, 1980). Anorexics who rely on dietary restriction with binge eating and/or vomiting have been called "bulimic" (Casper et al., 1980) and "bulimic anorexic" (Garfinkel & Garner, 1982). The term "bulimia nervosa" (Russell, 1979) has been used to describe the form of

PAGE 27

21 anorexia that includes binging. In 1980, DSM-III established bulimia as a distinct classification for individuals who exhibit abnormal consumption patterns but who are not 25 percent below their normal body weight. However, many authors still use the term, bulimic, as a descriptor for any person who engages in ego-dystonic overconsumption. Recognized as more than an occasional feature of anorexia nervosa, binging has been investigated by a number of recent studies. Beaumont et al. (1976) found that vomiter-purgers have a 43 percent likelihood of binging while dieters have only a likelihood of 28 percent. Vomiter-purgers were also found to be less inhibited, more likely to have had a past weight problem and less obsessional than dieters. The prognosis for vomiter-purgers was considered worse than for dieters. In a test of body perception, anorexics who overestimated their own size were found to be more likely to binge (Button, Fransella & Slade, 1977). Dally and Gomez (1979), in an impressionistic account of 400 anorexics, state that binging begins in a subgroup of patients who cannot maintain their dietary restriction. This often occurs after nine months of dieting and is characterized by preferences for sweet, highly palatable foods. Bruch (1973) reports a similar pattern in her patients. The identification of bulimic behavior in a proportion of the anorexic population has brought a call by professionals to label bulimic anorexia as a distinct subgroup of anorexia nervosa (Russell, 1979; Casper et al., 1980; Garfinkel et al., 1980). Russell (1979) believes binge eating may be a response of the hypothalamus to sub-optimal body weight. He found, in an investigation of 30 patients with a history of binge-purge patterns, moderate levels of depression, preoccupation with food and body size, and significant weight fluctuations. Garfinkel and

PAGE 28

22 Garner (1982) believe vomiting begins as a means of controlling the effects of binging but gradually becomes a reinforcer for binge eating. They believe that bulimic anorexics display confusion in the identification of their affective states and tend to misinterpret various unpleasant emotions as urges to eat. In support of Beaumont et al. (1976), Garfinkel and Garner (1982) also found bulimic anorexics to have a worse long term prognosis and to be more outgoing than f asters. Patients with anorexia nervosa have characteristics which clearly distinguish them from others with eating disturbances. The lifethreatening nature of the weight loss, the unique population of young females affected and the morbid fear of obesity clearly identify the anorexic (DSM-III, 1980). However, the descriptions of dietary restraint and subsequent binging are similar to reports by obese and normal weight bingers (Wardle & Bienart, 1981), lending support to the belief that weight classification is a categorization that obscures the commonalities among all bingers. Binge Eating in "Normal" Weight Individuals It is possible that binge eating in normal wsight individuals had been overlooked for so many years because these persons do not show any physical manifestations (e.g., obesity) of their behavior. Or perhaps, it was because we agree with Wardle and Bienart (1981), that cravings for food, dietary restriction and eating binges are not of themselves abnormal and this has caused us to ignore such variables as frequency and quantity that can indicate an eating disturbance. Kornhaber (1970) was one of the first practitioners to identify binge eating as a problem

PAGE 29

23 among his patients and termed it, "the stuffing syndrcme." He listed hyperphagia, depression and withdrawal as its main symptoms. Bruch's (1973) description of "thin fat people" was another early report of binge eating in normal weight persons. Characterized by a preoccupation with food and weight and a pattern of fasting and binging often followed by self-induced vomiting, "thin fat people" can appear to be anorexic in style but rarely suffer lifethreatening weight loss. Bruch reports many of these patients have a history of obesity and she suspects that the eating disturbance may be common among Americans. In a survey of 279 college women (Ondercin, 1979), 78 percent reported binge eating episodes and 18 percent identified themselves as compulsive eaters. This latter group of women showed a higher degree of dieting, more prior treatment for overweight and greater dissatisfaction with their present weight. Those high in compulsive eating reported increased eating in response to unpleasant emotional states and more frequent thoughts about food during the day. In a study to validate a self-report measure for binge eating, Hawkins and Clement (1980) surveyed 255 women and 110 men who were normal and overweight undergraduates. Their findings showed 79 percent of the women and 49 percent of the men engaged in binge eating with no differences between normal and overweight subjects. A significant partial correlation (r = .58, p < .001) was found between dietary concern and binging after weight was taken into consideration. Both surveys confirm Bruch's belief that there are widespread eating disturbances among the population of normal weight individuals. Several authors have attempted to define and label the pattern of disordered eating evidenced by normal weight binge eaters. Rau and

PAGE 30

24 Green (1975) believe compulsive eating to be a neurological disorder, similar to epilepsy, that can be best described as ego-dystonic consumption. The term, "dietary chaos syndrcme" was coined by Palmer (1979) to describe a disturbance characterized by a disordered eating pattern, preoccupation with food, eating and weight, and fluctuating body weight. He believed vcmiting was used by seme to break the link between eating and weight gain. Orbach (1978b) believes compulsive eating in wemen has critical social dimensions related to sex-role stereotypes and discrimination against women. She believes women who binge eat maintain overly strict dietary standards. Eventually, they lose control and binge. Fear of thinness is stated as a major causal factor for the behavior of the binger. Three types of non-physical hunger, prophylactic eating, mouth hunger and nervous eating, are described by Orbach (1978) as responsible for much excessive eating. Prophylactic eating occurs when food is available in the present but its future availability is unknown. Fearful of possible deprivation, the individual eats. Mouth hunger occurs when a non-hungry person sees or imagines the pleasure of tasting food and develops an urge to taste it. The consumption of food as a reaction to stress typifies nervous eating. Boskind-Lodahl (1976) also offers a feminist perspective of eating disturbances and posits that binge-purgers have an exaggerated concept of femininity that produces unrealistic personal standards. The term, bulimarexia, was coined by Boskind-Lodahl and Sirlin (1977) to describe a syndrcme characterized by secretive binges followed by fasting, vomiting or purging and a sense of shame and despair. In a pilot treatment of bulimarexia, Boskind-Lodahl and White (1978) identified

PAGE 31

25 dietary restraint, societal pressure, lack of assertiveness and negative body image as factors contributing to the disturbance. Brenner (1980) and Kubistant (1982) have also adopted the terra in their descriptions of college women plagued by uncontrollable urges to eat in response to emotional stress. With the publication of DSM-III (1980) by the American Psychiatric Association, bulimia was officially established as an eating disorder distinct from anorexia nervosa. The new classification incorporated the main features of the disturbances previously termed, "the stuffing syndrome" (Kornhaber, 1970), "dietary chaos syndrome" (Palmer, 1979) and "bulimarexia" (Boskind-Lodahl & Sirlin, 1977). Even with its established diagnostic criteria, bulimia remains a confusing term for many practitioners. It is freely used to describe both a symptom and a syndrome and, prior to its classification in DSM-III, there was little understanding of its role in normal weight individuals (Mitchell & Pyle, 1982). The appearance of bulimia as an eating disorder category comes at a time when the problems of binge eating and such post-binge behaviors as vomiting and purging have become more evident. In a survey of 355 college students, Halrai et al. (1981) found 13 percent with the major symptoms of bulimia. This percentage represented 19 percent of the female sample and 5 percent of the males. Among the women, 35 percent labelled themselves "binge eaters" and 68 percent reported episodes of binge eating. Efforts to understand bulimia and its major characteristics such as binge eating, vomiting, fasting and purging nave resulted in a number of recent descriptive studies. Pyle, Mitchell and Eckert (1981) report

PAGE 32

26 that in an investigation of 34 cases they found their average patient had the disorder for four years before seeking treatment. They report most bulimics went on binges in response to stress, and found the behavior to interfere with their daily activities and to inhibit their interest in sexual activity. Pyle et al. (1981) believe the incidence of bulimia is under-reported and the severity of the disorder is minimized even when reported. In a study of the frequency and duration of binge eating in bulimics, Mitchell, Pyle and Eckert (1981) had 40 patients monitor their binge behavior for one week. They found that the average binge totalled 3,415 calories, occurred in the evening, was often followed by vomiting and consisted of highly caloric food. A comprehensive survey of 316 cases of bulimia was conducted by Johnson, Stuckey, Lewis and Schwartz (1982). Of their sample, 96 percent were white, 83 percent had some post-secondary school education, 50 percent were students and 80 percent were unmarried. Average age of onset was 18; modal age of onset was 16. The average patient had the disorder for five years five months at the time of the investigation. Examining weight classifications of the subjects they found: 61 percent were normal weight, 17 percent were overweight and 21 percent were underweight. Half the sample reported a history of overweight and only 5 percent had a history of anorexia nervosa. Direct investigation of binge behaviors found 51 percent binge eating once each day or more and 78 percent binge eating in less than two hours. Examination of post-binge behaviors showed 63 percent using laxatives and 81 percent vomiting. Included in the preceding post-binge figures were the 26 percent of the total sample who reported using both laxatives and vomiting after a binge. The quantity of food consumed during a binge

PAGE 33

27 ranged from 1,000-55,000 calories with the average being 4,800 calories and costing $8.30. One important feature of the Johnson et al. (1982) survey was the report by 34 percent of the sample that binging began after the physiological and psychological deprivation of restrictive dieting. This supports Roth's (1982) belief that binge eating is a response to scarcity and that it is often the emotional and cognitive notion of deprivation that activates a "hunger from the heart." Johnson et al. (1982) in discussing the implications of their findings state: For certain vulnerable individuals, it seems that prolonged restrictive dieting may exacerbate a tendency toward bulimic behavior. It will be the task of future research to identify biological and psychological risk factors which predispose individuals to developing this type of behavior, (p. 14) Dietary Restraint and Counterregulation In an extension of earlier research in obesity, a number of studies have examined the concept of dietary restraint across all levels of weight classifications. As measured by the Restraint Questionnaire (Herman & Polivy, 1975), dietary restraint refers to the extent to which a person demonstrates a concern for dieting by his/her attitudes and behaviors. In efforts to explain contradictory findings in obesity research, this concept of dietary restraint has been of great assistance. The theoretical background behind it can be best understood after a consideration of two related bodies of research, Schacter's (1967) externality theory and Nisbett's set-point theory. As discussed previously, Schacter's (1967) externality theory proposes that the obese are more responsive than normal weight individuals

PAGE 34

28 to external cues concerning food and less sensitive to internal physiological states such as hunger. Rodin (1977) reports positive results in demonstrating that the time of day and the smell, taste and sight of food are more influential on the consumption patterns of the obese than of the non-obese. The obese have also been found to be more emotional and distractible (Pliner, Meyer & Blankenstein, 1974; Rodin, 1973). Schacter et al. (1968) found that anxiety would significantly decrease eating in normal weight individuals and caused a small, non-significant increase in consumption by the obese. Eating caused no anxiety reduction for their obese subjects causing Schacter et al. to reject the psychosomatic hypothesis of obesity (Kaplan & Kaplan, 1957). In 1972, McKenna, believing that the food (crackers) used in the Schacter et al. (1968) was too neutral, repeated the design substituting chocolate chip cookies tor the crackers. Although just short of significance, the increase in eating shown by the obese was almost as strong as the decrease in consumption by normal weight subjects. McKenna, in partial support of Schacter et al., did not find food to act as an anxiety reducer. This support, although inconclusive, of the psychosomatic hypothesis together with studies questioning the adequacy of the externality theory (Tom & Rucker, 1975; Rodin, 1976; Leon & Roth, 1977) became the basis for such new hypotheses as set-point theory (Nisbett, 1972) and the concept of dietary restraint (Herman & Polivy, 1975). Two studies (Nisbett 1968; Nisbett & Kanouse, 1969) introduced Nisbett to the belief that the externality theory's explanation of the differences in the eating habits of obese and normal weight individuals was not sufficient. In a study of taste, deprivation and weight,

PAGE 35

29 Nisbett (1968) found that "the overweight subjects exhibited a remarkable pattern of behavior. . They tended to eat either very large or very small amounts of food" (p. 114). He also found that individuals of normal weight who were once obese were as responsive to external cues as those presently obese and more responsive than normal weight, never obese individuals. In a later study of supermarket buying habits, Nisbett and Kanouse (1969) found that overweight subjects bought more and intended to buy more if they had recently eaten than if they had not. TO explain these and other findings concerning the obese, Nisbett (1972) proposed set-point theory. Set-point theory holds that the amount of stored body fat varies from person to person. Weight reduction reduces overall body size but not the number of fat cells in the body. Each person has a physiologically programmed level of fat tissue storage which is correct for them. This level, known as one's set-point, will be higher in some persons than in others. The central nervous system tries to keep each individual at his/her proper set-point or base-line for fat storage. Overweight people are usually hungry because they are attempting to achieve a socially acceptable body size that is below their needs as determined by their set-point. Therefore, they are biologically underweight and hungry (Nisbett, 1972). In extending set-point theory to Schacter's (1967) externality theory, Nisbett (1972) contends that externality occurs due to the conflict between the physiological need to maintain a higher weight and the social desirability of a lower weight. The obese person, or any person who tries to keep below set-point, finds him/herself more externally responsive because of unmet biological needs. Nisbett states

PAGE 36

30 that the extremely obese who have stopped dieting achieve their set-point and are no longer overly responsive to external cues. Nisbett's theory remains unconfirmed (Wardle & Bienart, 1981) but it does provide a possible explanation for the contradictory findings in externality research. For example, Meyer and Pudel (1977) demonstrated externality in the latent-obese, those of normal weight who actively control their intake to maintain their size, and Rodin and Slochower (1976) found that externality did not show a linear relationship with adiposity (fat tissue). It also appeared in normal weight individuals, particularly those who controlled their weight. In the initial investigation of dietary restraint, Herman dad Polivy (1975) designed a study similar to McKenna's (1972) test of the Schacterian and psychosomatic hypotheses. In that study, McKenna used the anticipation of mild or severe electric shock to create low or high anxiety and provided appealing and non-appealing food for his subjects to consume. His results showed a significant interaction for weight and anxiety with the obese eating more when nervous but with no evidence that differential anxiety reduction occurred. Schacter's theory was partially refuted in that all weight groups responded to the more appealing food. In keeping with set-point theory, Herman and Polivy hypothesized that dieting normal weight individuals (restrainers) would resemble behaviorally the obese who are concerned about their weight. In addition, the obese who have repudiated social pressure and attained their set-point would resemble the unrestrained normal weight individuals who are not food-deprived. Herman and Polivy (1975) found no main effects for anxiety or restraint but did find a significant interaction between the two

PAGE 37

31 conditions. Anxiety served as a mild cause of overconsuraption in restrained eaters and significantly decreased consumption in unrestrained ones. When the overweight subjects were taken out of the analysis, the interaction emerged as stronger, further supporting dietary restraint and not weight status as the important factor in overconsumption. In consideration of Nisbett's (1972) theory, Herman and Polivy raised the possibility that eating behavior differences may be dependent on the relative extent of underweight, and not overweight. They believed the measures on the Restraint Questionnaire may represent the relative deprivation of individuals, who regardless of actual weight restrict their food intake. Prior to the research in dietary restraint, there was evidence (Wboley, 1972; Mahoney, 1975) that the individual's cognitive beliefs played a major role in the regulation of intake. Wooley (1972) found that belief about caloric intake was a more important determinant in subsequent overeating than actual calorie intake. Polivy and Herman (1976b) used this information to re-examine prior findings (Polivy & Herman, 1976a) on the effect of alcohol. In the first study, Polivy and Herman administered alcohol and a placebo, both disguised as Vitamin C, to restrained and unrestrained subjects. They found, contrary to expectation, that alcohol did not disinhibit the restrained subjects but actually caused them to eat less than after the placebo. In their second study, Polivy and Herman expanded the experimental design and correctly labeled the alcohol for half the subjects in each condition. This yielded a significant interaction, demonstrating that the restrained subjects required both a physiological change and appropriate cognitive labels before losing his/her eating inhibitions.

PAGE 38

32 A further investigation of dietary restraint (Polivy & Herman, 1976) found that within a population diagnosed as depressed, restrainers gain weight during periods of emotional agitation but the unrestrained lose weight. The authors speculate that emotional stress may disrupt the strict standards that restrainers maintain. Polivy, Herman and Warsh (1978) compared the emotional responsiveness of restrained subjects to unrestrained ones in a study using a situation similar to one used with obese and normal weight individuals by Pliner, Meyer and Blankstein (1974). In addition, caffeine was administered to half the subjects to manipulate the state of internal arousal and a placebo was given to the other subjects. Similar to past studies of obese-normal differences, the results showed that in the placebo condition restrained subjects rated emotional stimuli more extremely than unrestrained subjects. However, in the caffeine condition, restrained subjects became less emotional and unrestrained subjects grew more emotional. The authors offer many possible explanations for this effect. One possibility is that restrained subjects have an elevated level of internal arousal, as is believed to be true with hyperactive children. When this level of arousal is increased (e.g. caffeine) it actually produces a decrease in emotional responding. Polivy et al. speculate whether unrestrained individuals, if aroused, would resemble restrained subjects. Herman, Polivy, Pliner, Threlkeld and Munic (1978) in a study of distraction and emotional arousal, performed two experiments with restrained and unrestrained subjects. In the first a proofreading task was given once a month for three months. Similar to Rodin's (1973) findings with the obese, restrained subjects were more distractible in

PAGE 39

33 the initial session. However, this effect did not persist over the next two sessions and restrainers became less distractible while the unrestrained grew more distractible. Herman et ai. believed changes in arousal level may have accounted for the results and designed a second experiment to manipulate acute arousal. In support of past studies (Herman & Polivy, 1975; Polivy et al., 1978), the results showed restrained subjects to be more emotional in both the intensity and duration of their responses. Restrained subjects were found to be more distractible than unrestrained ones in calm situations. This effect was reversed in periods of anxiety. Herman et al. raise the question of whether the restrained subject is more aroused in general or simply more arousable. They suggest that situational factors may control the behaviors of restrainers but that the externality this implies may be reversible given variations in arousal. In another early investigation of dietary restraint, Herman and Mack (1975) hypothesized that there may be many normal weight individuals who are below set-point and have "latent" externality. They proposed that a better predictor of eating behavior would be found by measuring the level of dietary consciousness and restriction of their subjects. Using the Restraint Questionnaire (Herman & Polivy, 1975) to divide their subjects into restrained and unrestrained categories, Herman and Mack predicted that restrainers would demonstrate counterregulatory eating. Counter-regulation is a laboratory phenomenon which has been likened to binge eating (Polivy, 1976; Spencer & Fremouw, 1979). It is a behavior where restrained (diet-conscious) individuals will consume more in a free eating situation after eating a preload of food believed to be high in calories than if they had not eaten anything

PAGE 40

34 previously. Subjects found to be low in restraint characteristics (little preoccupation with food and weight) usually regulate their intake such that they eat less after an initial preload than if they had gone into the free eating period without eating anything. In the Herman and Mack (1975) study, under the guise of a taste test, subjects were placed in one of three conditions: no preload, one milkshake preload or two milkshake preload. This was followed by a period during which the subjects could eat as little or as much as they wished. As expected, restrained subjects consumed more after a preload than without a preload. Unrestrained subjects showed an opposite effect, eating less after a preload to balance their intake. Herman and Mack found no significant differences between the obese and non-obese in restraint scores although overweight and restraint did correlate positively. They concluded that the differences in eating behavior previously attributed to the obese were more likely explained in terms of dietary restraint. Counter to what Herman and Mack expected, the two milkshake preload caused no more later consumption than did the one milkshake preload. They concluded that once the individual's restraint had been broken by the first drink the second was inconsequential to their counter-regulatory eating. To gather greater support for the choice of dietary restraint as a main determinant of eating behavior, rather than weight classification, Hibscher and Herman (1977) repeated the design of Herman and Mack (1975) and tested three distinct weight groups: underweight, normal weight and obese. Each group was divided into restrained and unrestrained subjects. The results showed consumption following a preload varied as a function of restraint and not weight classification. In addition,

PAGE 41

35 Hibscher and Herman examined levels of free fatty acids (FFA's) in their subjects. Nisbett (1972) posited that elevated levels of FFA's in the bloodstream, generally accepted as characteristic of the obese, were actually a consequence of being below set-point (relative deprivation) and could be better predicted by a measure of this factor. In support of Nisbett' s theory, Hibscher and Herman found elevated levels of FFA's to be associated with dieting across all weight groups. However, with stress correlated with FFA's in past studies, they speculate that the stress of dieting could account for the association. Polivy (1976) demonstrated that counterregulation is a cognitive effect in a study which varied the caloric content of the preload and the information given the subject about the preload's content. A statistical trend (p < .07) emerged indicating that restrained subjects ate more after ingesting a preload perceived to be highly caloric, even when it was not. In a study modifying the procedures of Polivy (1976) and Hibscher and Herman (1977), Spencer and Fremouw (1979) provided substantial evidence for the importance of counter-regulation as a cognitive variable in the short-term regulation of food intake. Spencer and Fremouw preloaded female subjects who had been divided into three weight classifications. Half of each group were told the preload was high in calories; the other half were told the preload was low in calories and made from a new calorie-free sweetener. In actuality, all preloads were identical. Spencer and Fremouw' s results showed that restrained subjects who were told the preload was highly caloric ate twice as much during the free-eating period than restrained subjects told the preload was low calorie. This interaction (restraint x

PAGE 42

36 instructions) was significant. Weight level was not found to be a significant factor in any of the results. Unrestrained subjects were found to be unaffected by the instructions. Lowe (1982) attempted to account for the counterregulation demonstrated by restrained subjects. She suggested that chronic dieting causes restrainers to feel that they must constantly deprive themselves of food. Overeating after a preload occurs as a logical response to the anticipation that the diet will shortly return. Lowe called this situational factor, anticipated deprivation. She hypothesized that restrainers experience anticipated deprivation when their restrictions have been loosened by eating "forbidden" food and they subsequently overeat. Nonrestrainers perceive food as always available to them and do not experience this effect. In Lowe's (1982) study, restrained and unrestrained subjects were preloaded at the beginning of the experimental session. Half of each group were told that after the session they had to fast for four hours and return to complete the experiment. Other subjects were also told to come back in four hours but to eat as they normally would throughout the day. After these instructions were administered a free-eating period began under the guise of gathering baseline data. In actuality, the follow-up session never occurred and the dependent measure was the amount eaten during the free-eating period. Subjects were run in either a taste test condition or a learning task condition. This was to examine whether the breakdown of restraint would occur in a laboratory setting using a nonfood related pretext. Lowe's results showed anticipated deprivation as a significant main effect in the taste condition and as falling just short of significance

PAGE 43

37 in the learning task. In the taste test situation a trend emerged for restrained subjects to consume more but this variable did not reach significance. Restraint was found to be a significant main effect in the learning task (p < .05). In neither condition was there an interaction between deprivation and restraint. Lowe concluded that anticipated deprivation was a factor in cyclical overeating. The anticipation of a four hour period of restriction was enough to produce a significant amount of overeating on the part of the restrained eaters. Based on her results in the learning task experiment, Lowe believed the importance of restraint was reaffirmed in that it was a main effect in an experiment that was not done within a context of food or taste. Overall, Lowe concluded that both internal factors (e.g., restraint) and external factors (anticipated deprivation) must be taken into account when studying eating behavior. Lowe's design did not determine the relative importance of the preload in the breakdown of restraint in her subjects. She suggested future research to examine the influence of anticipated deprivation without the restraint breaking preload, believing that if anticipated deprivation alone could loosen restraint, the implications for dieters would be extensive. It is clear from examining the work in dietary restraint and counterregulation that there are parallels to reports of binge eating by individuals of all weight classifications. Wardle and Bienart (1981) view dietary restraint as having a "causal role in relation to binges. However, its action is primarily as a background against which specific events can tip the balance into uncontrolled consumption" (p. 107). Research has shown preloads as a factor in counterregulation. Lowe's

PAGE 44

38 work demonstrates a role for anticipated deprivation in enhancing the effects of a preload. Polivy (1976) hypothesized that restraint was a cognitive factor based on a common-sense notion that individuals choose to diet. Similarly, outside the laboratory, dieters choose to taste food "forbidden" by their restrictive standards. The preload has been considered a mechanism that cognitively triggers the dieter to believe all control has been lost (Wardle & Bienart, 1981). Anticipated deprivation, as hypothesized by Lowe (1982), has yet to be tested as a sufficient cause for the breakdown of dietary restraint. Treatment of Binge Eating Although identified in 1959 by Stunkard, binge eating did not become a focus for treatment until recently. Prior to this time most therapists either agreed with Bruch's (1973) emphasis on psychotherapy which did not focus on the single behavior of binge eating or followed a standard behavioral management program as exemplified by the Stuart and Davis (1972) model. Binge eating was primarily seen as a single eating behavior amidst many that contributed to weight gain. Outcome studies of weight loss programs and observations of dieters raised many questions about the success of behavioral methods. Hall and Hall (1974) questioned the long term maintenance of the lower weights and new behaviors achieved in behavioral weight loss programs. Wilson (1976) specifically pinpointed binge eating as a major complication in weight loss programs. He believes binge eating may be associated with complex psychological and behavioral difficulties. Based on observations of his own clients, Wilson reported binge eaters are often

PAGE 45

39 underassertive harbor negative feelings toward themselves and are deficient in basic coping skills. Stunkard (1976) also pointed out binge eating as a block to weight loss and believes psychotheraphy needs to focus on the body image disturbances and actual binge behaviors of patients. Brightwell and Sloan (1977) in a review of long term studies in behavioral weight loss programs reported only 17 had a follow-up of 26 weeks or more and concluded in his review that behavioral programs might only help a certain subgroup of patients. There are many reports of weight loss treatment for the obese and seme make specific reference to binge eating. Wollersheim (1970) found in a comparison of different approaches that his "focal" therapy based on learning principles was effective in reducing eating in isolation and emotional/uncontrolled eating. Stalonas, Johnson and Christ (1978) examined ten aspects of a behavioral weight loss program and found only one variable, the control of impulse eating, to be correlated with weight loss. Overeaters Anonymous (R., 1979) has adapted a format similar to Alcoholics Anonymous and specifically focuses on compulsive eating and the addictive nature of binging. Mahoney (1975) advocates the use of cognitive and behavioral methods with the obese. He believes their "cognitive claustrophobia," or dietary restriction, causes their cravings for forbidden food and he sees a need to use cognitive techniques to manipulate thoughts around food and dieting. In a study of 280 participants in an intensive behavioral weight loss program, Loro and Orleans (1981) conducted a behavioral analysis of binge eating in obese patients. Based on their observations, they recommended a number of behavioral treatment approaches and one quasi-behavioral strategy for the control of the antecedents and

PAGE 46

40 consequences of binge eating. They suggested such methods as assertive training, anxiety management, relaxation and problem-solving training to help bingers cope with problems of anxiety, stress and poor social skills. They recommended thought-stopping to help with obsessions about food and cognitive restructuring to change irrational ideas about weight reduction. To help control consequences, Loro and Orleans suggest aversion and punishment procedures, response cost strategies and the control of social reinforcement through the help of family and friends. Programmed binging is recommended for clients who binge eat out of rebellion against authority figures, of which the therapist is one. It involves the use of paradoxical intention and is done under close therapeutic direction. Clients are instructed to plan and implement a binge, usually in a specified fashion. The benefits of this practice are many. Clients have the opportunity to analyze their binge behavior and, with the help of the therapist, restructure their thoughts around it. For those clients who have felt totally out of control, the programmed binge brings them more power in the situation and enhances feelings of self control. Programmed binging can also serve to help the client view a single binge as a slip and not a total relapse. With anorexics exhibiting binge purge behavior, Russell (1979) recommends hospitalization as a means to interrupt the cycle. This inpatient care is highly dependent on the ability of the nurses to work with the patient as attempts are made to achieve a pattern of regular meals and an acceptance of a higher weight. Monti, McCrady and Barlow (1977) report success using positive reinforcement, informational feedback and contingency contracting to reduce binge purging and increase body weight in a bulimic anorexic. Believing bulimic behavior

PAGE 47

41 to be analogous to an obsessive-compulsive disorder, Rosen and Leitenberg (1982) treated a bulimic anorexic female with the behavioral techniques of exposure and response prevention. The patient was instructed that vomiting was an escape from binging which served to maintain binge behavior and that the relief from the physical discomfort of being too "stuffed" maintained vomiting. In treatment sessions, the patient was exposed to binging and the discomfort around it, but not allowed to vomit. Gradually the binge behavior, now without an escape mechanism, decreased as did anxiety toward food. The growing public awareness of bulimia as an eating disorder has prompted new treatment strategies. Boskind-Lodahl and White (1978) reported a pilot group treatment approach termed "experientialbehavioral" that utilizes a feminist perspective. The format included assertiveness training, role-plays, sensory-awareness and guided fantasy exercises as well as such behavioral tasks as monitoring and contracting. Boskind-Lodahl and White reported dieting and restraint as leading contributors to the feelings of guilt and shame experienced by group members. Results from this pilot treatment showed success in 10 of 12 but follow-up results demonstrated a regression toward pretreatment levels. After a refinement of their program, Boskind-Lodahl and White's (1978) six month follow-up showed all fourteen subjects were able to eliminate vomiting, three stopped binging and seven decreased the number and durations of their binge episodes. Fairburn (1981) has reported success with a cognitive—behavioral treatment with bulimics that he believes could also be helpful with overeaters. This approach seeks to immediately disrupt the binge-purge cycle through strict behavioral monitoring, frequent therapy sessions

PAGE 48

42 and the inclusion of family and friends in treatment. Once reduced vomiting is achieved, Fairburn uses problem solving, cognitive restructuring and gradual exposure to forbidden food as ways to modify abnormal attitudes toward food, eating and body size. Leclair and Berkowitz (1983) also believe treatment success depends on an early reduction of binge-purge behavior. This in turn will help reduce depression. To Fairburn' s program they add nutritional counseling, sensory awareness and examinations into the links between binging and emotional states. Kubistant (1982) uses similar strategies in his work with bulimics, often utilizing individual therapy and support groups concurrently. He advocates the use of programmed binges and eating practice sessions where clients work on tasting food and on consuming it more slowly. There are few treatment programs designed to specifically target binge eating. Green and Rau (1974) consider compulsive eating a neurological disorder and treated it with anti-convulsant medication. In a replication, Wermuth, Davis, Hollister and Stunkard (1977) failed to find any long term effectiveness of this treatment among their twenty patients. Orbach (1978a) has written extensively of a group treatment approach for compulsive eating which can be started among any group of women. She advocates the identification of the antecedents and consequences of eating, fantasy exercises, the development of hunger recognition and acceptance of one's body. Restrictive diets are seen as the polar opposite of binge eating and Orbach recommends termination of all diets while resolving the issues around compulsive eating. Roth (1982), influenced by Orbach' s work, believes that physical and emotional deprivation causes binging and the only release from deprivation will

PAGE 49

43 stop compulsive eating. She has developed a support group program, the Breaking Free Workshops, that seeks to examine the underlying personal needs that cause binge eating. The findings in dietary restraint have begun to have an impact on treatment strategies for binge eaters. Wardle and Bienart (1981) favor a release from overly restrictive diets and the irrational beliefs about intake which accompany them. They suggest that the dieter who binges when placed on any restricted regimen may be better served by an exercise program such as the one reported by Gwinup (1975). In his study of the singular effect of exercise on obesity, subjects lost an average of 22 lbs in a one year period. Spencer and Fremouw (1979) in discussing their results in dietary restraint believe that cognitive techniques may be necessary to assist behavioral programs. Lowe (1982) also supports cognitive approaches and suggests the abandonment of inflexible dietary rules and the inclusion of menus that allow preferred foods. Wardle and Bienart (1981) in their review of binge eating conclude "If, as has been suggested, dietary restraint plays a major role in providing a setting within which binges are probable, treatment must aim to reduce restraint" (p. 107). Movement toward this goal requires a better understanding of the factors (e.g., anticipated deprivation) which create and influence restraint. Research efforts have just recently informed us that, contrary to the beliefs of most dieters, dieting causes binges. Successful treatment of binge eating in the future will occur only after an understanding of the many complex factors involved.

PAGE 50

44 Sunmary This chapter has examined the theories and research findings that comprise the literature on binge eating. It is evident that binge eating and other related behaviors such as vomiting, purging and fasting have become cannon and problematic for many individuals in our society, particularly wcmen. Throughout the many reports of eating behavior from both experimental and descriptive studies certain themes emerge. This study is based on these themes and has the purpose of contributing new knowledge to the literature in binge eating. In efforts to discover the causative factors in obesity, binge eating became a focus of investigation. It was believed that binge eating was a behavior that existed only within the behavioral repertoire of the obese and attempts to uncover reasons for binge eating and general overconsumption resulted in two prominent hypotheses, the psychosomatic and the Schacterian theory of externality (Schacter, 1967). The many conflicting findings that have arisen from attempts to prove and/or refute these hypotheses gave way to a belief that degree of overweight may not be the factor that best discriminates between binge eaters and normal eaters. The concept of dietary restraint, as measured by the Restraint Questionnaire (Herman & Polivy, 1975), has been shown in this review to be a viable and experimentally proven alternative. Studies described within this chapter have demonstrated that many tenets of both the externality and psychosomatic hypotheses are more strongly upheld by the concepts of dietary restraint and counterregulation than by the original theories themselves. In addition to the concept of dietary

PAGE 51

45 restraint, which is known to be a cognitive factor, a situational variable termed anticipated deprivation has been shown to contribute to overconsumption by binge eaters. It appears that not only do individuals with high restraint overeat after feeling they've "cheated" just a little but they further increase consumption when they anticipate having to return to being self-restrictive. Ihe issue of deprivation has appeared in several areas of the literature review. Johnson et al. (1982) reported that 34 percent of the bulimics they surveyed began binge eating after a period of deprivation. Orbach (1977a) lists prophylactic eating, eating because future availability of food is unknown, as one of three types or non-hunger related eating. Roth (1982) believes the individual will stop binge eating when his/her feelings of physical and psychological deprivation have been resolved. Nisbett (1972) claims that it is the physiological deprivation caused by constant dietary restriction that makes the individual more externally responsive to food cues and therefore, more likely to binge eat. Bruch (1973) and Dally and Gomez (1979) have reported a similar pattern in anorexics. The investigation of deprivation in the context of the dietary restraint model of Herman and Polivy (1975) is an appropriate direction for the examination of binge eating. Lowe (1982) has already demonstrated how the anticipation of deprivation affects those restrained individuals who have already broken their restraint through the consumption of "forbidden" junk food. This study will attempt to determine if the anticipation of deprivation is a powerful enough factor to cause those high restraint individuals to overeat without first externally disrupting their dietary restraint.

PAGE 52

CHAPTER THREE METHODOLOGY The primary issue examined in this study was the effect of a preload of food on the eating behavior of restrained and unrestrained subjects who anticipate a period of fasting. Of secondary interest was the effect of the preload on the attractiveness of food (taste and buying intention) as reported by the subjects. In this chapter, the experimental methodology utilized in the study is described. The topics presented are the experimental hypotheses, the research design, subjects, the experimental procedures, instrumentation, data analyses, rationale for methodology and limitations of the study. Experimental Hypotheses The following hypotheses, stated in null form, were tested within the methodology of this study: 1. There will be no differences between subjects in the four groups (restrained preloaded, unrestrained preloaded, restrained non-preloaded, unrestrained non-preloaded) in the amount of food consumed during the free eating period. 2. There will be no differences between subjects in the four groups (restrained preloaded, unrestrained preloaded, restrained non-preloaded, unrestrained non-preloaded) in their ratings of the taste of the food consumed. 46

PAGE 53

47 3. There will be no differences between subjects in the four groups (restrained preloaded, unrestrained preloaded, restrained non-preloaded, unrestrained non-preloaded) in their ratings of buying intention for the food consumed. The Research Design The study utilized a 2 x 2 factorial design to test the above hypotheses. The factors in the design were level of restraint and presence of preload. Level of restraint (restrained or unrestrained) were assessed by a median split of scores from the Restraint Questionnaire (Polivy, Howard & Herman, 1976, Appendix A). Assignment into a preload condition (preloaded or non-preloaded) was done by random assignment from the two respective restraint levels. All subjects experienced an anticipation of food deprivation by being requested to fast for four hours after the initial session. The Restraint Questionnaire was administered to all female students enrolled in Introductory Psychology during the first week of the Fall 1983 semester. It was part of a battery of pre-tests being used in different studies within the Psychology Department. A total of 482 female students completed the Restraint Questionnaire. A median split was calculated from the results of this testing of all potential subjects. Restrained eaters were defined as those scoring 16 or more on the Restraint Questionnaire and unrestrained eaters were those who scored less than 16. The use of a median split to separate restrained and unrestrained eaters was consistent with prior methods and yielded similar results (Herman & Mack, 1975; Herman & Polivy, 1975; Spencer &

PAGE 54

48 Fremouw, 1979). From the pool of 482 students, 90 participated in the study. There were a total of 60 subjects in the four equal cells of the factorial design. In addition, 15 restrained and 15 unrestrained subjects were pretested using the Taste Preference Sheet (see Appendix C) Since assignment into one of the two restraint levels was not random it was necessary to gather data on any potential differences between restrained and unrestrained subjects in their ratings of taste and buying intention. Due to the nature of one of the independent variables (presence of preload), the subjects in the cells of the factorial design could not be pretested with a measure requiring consumption of food. TO provide the necessary data, 15 restrained and 15 unrestrained subjects were tested with the Taste Preference Sheet only. This occurred prior to the testing of the subjects who comprised the four cells of the factorial design. Actual food consumption during the free eating period, a behavioral measure, served as the first dependent variable. Taste and buying intention, the other dependent variables, were assessed by the Taste Preference Sheet, a thirty-item questionnaire of paired descriptors developed for this study. Each dependent measure permitted the examination of possible main and interactive effects of the independent variables, level of restraint and presence of preload. The design of the study is graphically depicted as follows: Taste Preference Pre-test Restrained Unrestrained Pre-test

PAGE 55

49 Treatment Restrained Unrestrained Preload NoPre load Subjects Subjects in this investigation were 90 female undergraduates enrolled in an Introductory Psychology course at the University of Florida. Students in this course were required to earn five experimental credits during the course of the semester through participation in studies approved by the Psychology Department. Students participating in this study received two credits toward their course requirement. During the first week of the Fall semester, students in all sections of Introductory Psychology were given a number of pretests from the various studies planning to use these students as subjects during tne semester. Included in this battery was the ten-item Restraint Questionnaire (Polivy et al., 1976, Appendix A). From this pretest, restraint scores for all female students enrolled in Introductory Psychology were compiled by the experimenter. A median split of scores divided the female students into unrestrained and restrained categories. Thirty subjects (15 restrained, 15 unrestrained) participated in the Taste Preference Sheet pre-testing. Sixty subjects (15 per cell) comprised the four experimental cells of the factorial design. Any

PAGE 56

50 female enrolled in Introductory Psychology who completed the pre-test battery during the first week of classes was eligible to participate in the study. Subjects elected to participate in the study through writing their instructor-assigned experimental number on a sign-up sheet that listed the times and locations of the experiment. Before each subject arrived, she was assigned to a restraint level (restrained or unrestrained) based on her Restraint Questionnaire score in relation to the median split of scores gathered during Psychology Department pre-testing. In the Taste Preference pre-test, 15 restrained and 15 unrestrained subjects were tested. In the factorial design, 60 subjects were tested to meet the criterion of 15 per cell. Subjects in each restraint level were randomly assigned to either the preloaded or non-preloaded condition. Experimental Procedures Pre-testing with the Restraint Questionnaire During the first week of classes, all students enrolled in Introductory Psychology were administered a battery of instruments. These instruments were pretests from the various studies in the Psychology Department that would seek subjects from the pool of Introductory Psychology students. The battery was administered and collected during class time by a team of Psychology Department research assistants who visited each section of the course sometime during the first week of the semester. Students received one experimental credit for completing the battery. Upon administration of the battery to all

PAGE 57

51 the enrolled students, the completed instruments were returned to the investigators of the respective studies involved. Pre-testing with the Taste Preference Sheet Subjects signed up for a study entitled "Sensation and Perception" and self-selected date and time from the choices posted. Sessions were offered between 10:U0 a.m. and 11:15 a.m. in order to provide prelunch ratings. Each subject was run individually and had no contact with other subjects participating in the study. The session was described as an investigation of the taste perceptions of undergraduate females. All subjects were given a Taste Preference Sheet (see Appendix C) and asked to rate three types of food (M&M's, peanuts and animal crackers). The food was presented in three large bowls, each containing enough food to give the appearance that a substantial portion could be eaten before it would be visibly noticed. Subjects were left alone for fifteen minutes and a timer was set to allow the subject to anticipate the return of the experimenter. The size of the bowls and the timer served to provide a sense of privacy for the subject. The subject was asked to taste the three types of food and fill out the ratings sheet. At the end of the 15-minute period the experimenter returned to collect the ratings sheet and explained the purpose of the experiment to the subject. Treatment Procedures Subjects signed up for a study entitled "Sensation and Perception" and self-selected date and time from the choices posted. Each subject was asked to sign up for two sessions, the first lasting one hour and the second lasting 15 minutes. These sessions were separated by four

PAGE 58

52 hours on the same day and this four hour period coincided with what is normally considered lunchtime. For example, a session beginning at 10:00 a.m. and had a follow-up session at 3:00 p.m., thereby running through the normal lunch period. Each subject was run individually and had no contact with other subjects participating in the study. The experiment was described as an investigation of the effects of hunger on the perception of taste. All subjects were given a Marketing Questionnaire which asked for preferred shopping locations and familiarity with certain brand name products. While completing this questionnaire, those subjects assigned to the preload condition were required to consume ten M&M candies, ten peanuts and three animal crackers. This consumption was explained by the experimenter as a necessary preliminary to the later taste ratings. Subjects in the non-preloaded condition were not asked to consume any food while they completed the survey. After this initial survey period, each subject was again told that the purpose of the experiment was to assess the effect of hunger on taste and that they would make taste ratings in both sessions of the experiment. It was explained that in order for the experimenter to measure the effect of hunger each subject must fast from the time the first session ended until she returned to their second session. At this time subjects were given the Taste Preference Sheet and asked to provide initial taste ratings of three types of food. The food was presented in three large bowls, each containing enough food to give the appearance that a substantial portion could be eaten before it would be visibly noticed. Subjects were instructed to eat as much or as little of the food as they wished during the 15-minute rating period and that

PAGE 59

53 the experimenter would return at the end of the 15-minute period to collect the rating sheets and to give the subject instructions for the afternoon session. Before leaving the room the experimenter set a kitchen timer and turned it toward the subject to allow her to anticipate the experimenter's return. The use of the large bowls of food and the setting of the timer were to provide the subject with as much perceived privacy as possible within the constraints of a laboratory experience. The large bowls were used to give the subject freedom to consume the food without concern for giving the appearance that she ate a large quantity in a short period of time. The timer allowed the subject to anticipate the experimenter's return and reduced worry about being interrupted in the middle of eating. Upon returning at the end of the 15-minute period, the experimenter collected the ratings sheet and informed the subject that she was not required to return to the afternoon session. The actual purpose of the study was explained by the experimenter and any questions were answered. The subject was given a Subject Questionnaire (see Appendix D) to check if she intended to fast, if she had any prior knowledge of the true purpose of the study and if she had eaten immediately prior to attending the session. After collecting the Subject Questionnaire, the experimenter requested that the subject not discuss the study with any potential subject. After the subject departed, the bowls of food were weighed by the experimenter to assess how much was eaten during the 15-minute free eating period. The total amount consumed was used as the behavioral measure for one of the dependent variables of the study.

PAGE 60

54 It was not necessary to eliminate any of the subjects who completed the experiment. All subjects who participated and were asked to fast responded that they would have attempted to do so. No subject had eaten a meal within one hour of the start of the experiment nor did any report prior knowledge of the true purposes of the study. All required preloads were consumed by those subjects asked to do so. Experimenters and Experimenter Training The experimenters for this study were three female undergraduate research assistants who received credit for their work through the Psychology Department course — PSY 3912, Independent Research. Each experimenter worked with 30 subjects, five subjects from each of the two pretest cells and the four experimental cells. The experimenters were coordinated by a student in the graduate program in Counseling Psychology. This coordinator maintained records of the restraint scores and randomly assigned subjects to one of the preload conditions. Prior to the arrival of a subject the experimenter was told to which condition of preload that subject had been assigned. The experimenters were trained in the running of the experiment by the principal investigator during the first two weeks of the Fall 1983 semester. A general instruction sheet (see Appendix E) was used and included statements that were read to subjects during each phase of the experiment. Experimenters were instructed to limit their presentation to these statements whenever possible. Role-plays were utilized to practice the format of the study and uncover any potential problem areas

PAGE 61

55 Instruroentat ion The Restraint Questionnaire The Restraint Questionnaire was originally developed by Herman and Pol ivy (1975) to measure the extent an individual's attitudes and behaviors reflect a concern or preoccupation with dieting. In its original form it consisted of eleven items; the first six items examined diet and weight history and the last five examined concern with food and eating. In their initial use of the Restraint Questionnaire, Herman and Pol ivy found the scale to have substantial internal consistency (coefficient = .75). They report that the two subscales had internal consistency coefficients of .68 and .62, respectively. In a later use of the scale, Hibscher and Herman (1977) found the test-retest reliability of the scale to be high (r = .92). In their study, Hibscher and Herman also used a physiological measure of deprivation, free fatty acids (FFA's), as a dependent measure. An analysis of variance, performed to determine whether level of restraint or obesity was a better predictor of FFA's, confirmed that restraint was the. significant factor (F 1,80 = 8.54, p < .01). This report lends evidence to the construct validity of the scale. Two studies (Polivy et al., 1976; Polivy, Herman, Younger & Erskine, 1977) investigated the psychometric properties of the Restraint Questionnaire. The result was a refined instrument of ten items. This revised version has an internal consistency coefficient of .7 and a test-retest reliability coefficient of .9. The new questionnaire utilizes a forced-choice format and takes approximately 3-5 minutes to

PAGE 62

56 complete. Subject scores may range from 0, showing low dietary concern, to 35, showing high dietary concern. Marketing Questionnaire The Marketing Questionnaire (see Appendix B) is an instrument developed for this study which requests information about the consumer habits of the subject. It was not intended to provide any information relevant to the purposes of this study. Its utility in this study was as a task for subjects to complete during the first phase of the study when one half of the subjects were eating a preload. Behavioral Measure The amount of food (M&M's, peanuts and animal crackers) consumed during the free eating period served as the behavioral measure for the study. This was consistent with prior studies of counterregulation. Herman and Mack (1975), Hibscher and Herman (1977), Spencer and Fremouw (1979), and lowe (1982) used the amount of highly caloric, easily ingested food as their primary measure. Lowe's study of anticipated deprivation, which forms the basis for this study, used M&M's, peanuts and animal crackers. Taste Preferences Sheet The Taste Preferences Sheet (see Appendix C) is an instrument created tor this study. It consists of thirty questions about the taste and opinion of food. Subjects answered the same ten items for each of the three foods in the experiment. Only two of the ten items were actually used for the purposes of this study. These items were 3. How would you rate the taste of the food overall? bad 123456789 10 great

PAGE 63

57 6. How likely are you to buy this food or a similar food item within the next two weeks? very very unlikely 123456789 10 likely The remaining items were included to support the subject's belief that the experiment was testing the effect of hunger on taste. The use of a preference scale is a common tool in consumer studies (Kassarjian & Nakanishi, 1967). lorn and Rucker (1975) used a similar type of measurement in their study of the effects of external cues on the consumption and buying intentions of the obese. In a study of selected opinion measurement techniques, Kassarjian and Nakanishi (1967) found no significant differences between the results derived by seven different methodologies (likert rating, two open choice options and four limited choice options) Subject Questionnaire The Subject Questionnaire (see Appendix D) is an instrument developed for this study to verify that the manipulations intended in the design of this study were actually present. It seeks information on whether the subject was blind to the true purpose of the study, if she intended to fast as requested, how long it had seen since her last meal and if she ate the preload if instructed to do so. The Subject Questionnaire was used to eliminate subjects who do not conform to the necessary requirements of the study. A subject would have been dropped frcm the study for any one of the following reasons: 1. Subject was aware of the true purpose of the study. 2. Subject did not intend to fast and therefore did not anticipate a period of deprivation.

PAGE 64

58 3. Subject ate her last meal within one hour before the start of the experiment. 4. Subject was assigned to the preload condition and did not eat all of the preload. Data Analyses The data gathered from the pre-test of the Taste Preference Sheet were examined first. T-tests were used to investigate if the restrained and unrestrained subjects differed significantly in their ratings of taste and buying intention. When no differences were found, two-way (2x2) analyses of variance were conducted on the two dependent variables, taste ratings and buying intention. Results from earlier research (Herman & Mack, 1975; Hibscher & Herman, 1977; Spencer & Fremouw, 1979) indicated that pre-testing of the dependent variable, food consumption, was not necessary. A two-way (2x2) analysis of variance was conducted on these data. Kerlinger (1973) points out that the ability to investigate possible interactions is one of the main advantages of the factorial analysis of variance. When an interaction was found between the two independent variables (level of restraint, presence of preload), an examination of simple main effects was conducted as described by Huck, Cormier and Bounds (1974). Use of the F ratio occurs with the assumption of independent random samples, a normal distribution within the populations and equal variances. These criteria were met by this study. The F test is considered robust to possible inequalities in population variances when sample sizes are equal.

PAGE 65

59 Rationale for Methodology Wardle and Bienart (1981) point out that the laboratory research in counter-regulation has produced insight into binge eating. Herman and Mack's (1975) study of counterregulation under the guise of a taste test has been a model successfully followed by a number of researchers (Polivy, 1976; Hibscher & Herman, 1977; Spencer & Fremouw, 1979) in their examinations of eating behaviors. In her study of anticipated deprivation, Lowe (1982) also tested the taste test context used in prior counter-regulation studies against an identical experiment with a non-food context. Lowe's purpose was to examine if the food-related cues implicit in a taste test influenced prior results. Her results found restraint to have a significant main effect on consumption even when the guise of a taste test was not used. Throughout these studies and in other investigations of dietary restraint (Herman & Polivy, 1975; Polivy & Herman, 1976b; Herman et al., 1978; Polivy et al., 1978) the Restraint Questionnaire has successfully predicted which subjects would demonstrate behaviors previously associated with the obese. The work of Polivy et al. (1977) has established the instrument as a reliable measure of dietary concern and the studies conducted by Hibscher and Herman (1977) and Spencer and Fremouw (1979) demonstrate that weight classification is not a significant factor in counterregulation In the counter-regulation studies discussed above, subjects in each study were always of the same sex. This has been a standard procedure in studies of counter-regulation since scores on the Restraint Questionnaire have been found to be higher for females than for males (Polivy et al.,

PAGE 66

60 1976). If both sexes were randomly included in a study, there would be a disproportionate number of males found in the unrestrained category after a median split was done. Therefore, this study limited itself to female subjects. The choice to use only females was made by the author in light of the greater numbers of females with eating disorders (DSM-III, 1980). The research design in this study closely follows that utilized by Lowe (1982). A similar number of subjects were run and the preload consisted of the same foods in duplicate quantities. Lowe's selection of foods was supported by the findings of McKenna (1972) that demonstrated that neutral foods (e.g., crackers) are not always successful measures of true differences in overconsumption. To allow for comparisons to be drawn around the concept of anticipated deprivation, a similar period of fasting was requested of the subjects in this study. In Lowe's study all subjects were preloaded and level of restraint and anticipated deprivation served as independent variables. In this study all subjects anticipated a period of deprivation and level of restraint and presence of preload were the independent variables. Limitations of the Study The use of the laboratory as a setting to test binge eating presents the most serious questions of limitation in this study. Binge eating has been repeatedly found to be a behavior that is performed in private by the individual. In this study, individual running of subjects and an attempt to assure the subject's privacy (through the use of a timer and large food bowls) were included in the methodology to provide an optimal setting. The use of female undergraduates may limit the general izability of the

PAGE 67

61 results of the study. Although this particular group has been identified in many studies of binge eating, they tend to be a more physically active population and may not be comparable to women past their school years, many of whom may lead more sedentary lifestyles. The generalizability of the results of this study to a population of males cannot be assumed.

PAGE 68

CHAPTER POUR RESULTS The study was designed to investigate the effects of the presence or absence of a preload on restrained and unrestrained female subjects who anticipated a period of food deprivation. Three experimental hypotheses were formulated for the study. The first was constructed around the actual amount of food consumed by the subjects during a free-eating period. The second hypothesis was concerned with the taste ratings provided by the subjects. Ratings of food buying intention served as the dependent variable for the third hypothesis. The use of a 2 x 2 factorial design produced four distinct groups of subjects. Restrained preloaded subjects were those who scored high on the Restraint Questionnaire (Polivy et al., 1976), indicating high preoccupation with food and body weight, and were given a preload of food prior to the free eating period. Unrestrained preloaded subjects were low scorers on the questionnaire who received a preload of food before the free eating period. Restrained non-preloaded subjects were high scorers who did not receive a preload and unrestrained non-preloaded subjects were low scorers who did not receive a preload. The statistical tests used in the study were t-tests and two-way (2x2) analyses of variance. An alpha level of .05 was set as the basis for rejecting or failing to reject the null hypotheses. The remainder of this chapter provides a description of the results of the 62

PAGE 69

63 testing of the hypotheses. Numbers appearing in the tables have been rounded off to the second decimal place. Food Consumption It was expected that all four groups would differ significantly from each other on the dependent variable, food consumption. Restrained preloaded subjects were expected to consume the most food followed by restrained non-preloaded subjects. Unrestrained non-preloaded subjects were expected to rank third in consumption and unrestrained preloaded subjects were expected to consume the least. Thus, it was predicted that a main effect for restraint would be found. These predictions were made with the belief that the condition of anticipated deprivation would cause restrained non-preloaded subjects to consume in a pattern different frcm that found in prior studies. Previous research (Herman & Mack, 1975; Spencer & Fremouw, 1979) found that this group ate less than all other groups and, therefore, concluded that the preload was a necessary factor in counter-regulatory eating. It was predicted in the present study that the anticipation of deprivation would cause all high restrainers to overconsume and that preloaded restrainers would consume the most. Thus, the following null hypothesis was tested: 1 There will be no differences between subjects in the four groups (restrained preloaded, unrestrained preloaded, restrained non-preloaded, unrestrained non-preloaded) in the amount of food consumed during the free-eating period.

PAGE 70

64 A two-way (2x2) analysis of variance was conducted, using restraint (restrained; unrestrained) as one independent variable and preload (preloaded; non-preloaded) as the other independent variable (see Table 1). The test found no significant main effects nor a significant interaction. With no significant differences found between any two groups and no significant interaction present, it was not possible to reject the null hypothesis. An examination of the data (see Table 2) indicates that, as expected, restrained preloaded subjects ate the most. The unrestrained non-preloaded group ranked second, followed by the restrained non-preloaded group. The unrestrained preloaded subjects, as predicted, ate the least. Although this group ate a full one-third less than the restrained non-preloaded group, the large variations prevent this difference from reaching significance. It is notable that contrary to previous research the restrained preloaded group did not eat the least amount of food. This raises the possibility that the anticipation of deprivation may have had an effect on this group. Taste and Buying Intention Nonbehavioral measures have not been included in prior studies of dietary restraint. These studies have relied on the amount of food consumed as the sole dependent variable. In this study, a cognitive factor, food attractiveness, was measured by ratings of taste and buying intention. It was predicted that the results of these ratings would parallel the expected results from the analysis of the amount of food consumption. Therefore, it was expected that restrained preloaded

PAGE 71

65 Table 1 Analysis of Variance for Food Consumption Source

PAGE 72

66 Table 2 Means and Standard Deviations of Food Consumption Group Mean Ounces (rank) SD Preloaded Restrained 15 Unrestrained 15 2.07 (1) 1.23 (4) 1.89 .96 Non-preloaded Restrained 15 Unrestrained 15 1.83 (3) 1.90 (2) 1.59 1.28

PAGE 73

67 subjects would rate the taste of the food higher and have greater intention to buy the food than would subjects in the other groups. Restrained non-preloaded subjects were expected to rank second on both of these dependent variables followed by unrestrained non-preloaded subjects. The unrestrained preloaded subjects were expected to provide the lowest ratings. Based on these expectations, it was predicted that a significant main effect for restraint would be found for both the taste and buying intention dependent variables. In the study, assignment into one of the two restraint categories was based on the subject's score on a questionnaire. EXie to this non-random selection, it was necessary to test additional subjects to investigate if any significant differences existed between restrained and unrestrained eaters in their ratings of taste and buying intention. Using the Taste Preference Sheet (see Appendix C) 15 restrained and 15 unrestrained subjects were tested. T-tests revealed no significant differences between the two groups for taste ratings (_t(28) = .62, p > .5) nor for buying intention (t(28) = .91, p > .3). Based on this, it was considered reasonable to conduct two-way (2x2) analyses of variance on the ratings of taste and buying intention provided by the 60 subjects in the factorial design. For the dependent variable, food taste, the following null hypothesis was tested: 2. There will be no differences between subjects in the four groups (restrained preloaded, unrestrained preloaded, restrained non-preloaded, unrestrained non-preloaded) in their ratings of the taste of the food consumed.

PAGE 74

68 Table 3

PAGE 75

69 A two-way (2x2) analysis of variance was conducted for taste ratings using restraint (restrained; unrestrained) as one independent variable and preload (preloaded; non-preloaded) as the other independent variable (see Table 3). No significant main effects were found. Contrary to expectations, a significant interaction between restraint and preload existed (F(l,56) = 7.38, p < .01), indicating that the difference between the levels of one factor changes from one level to another level (see Figure 1). In this case, restrained subjects scored higher than unrestrained subjects when both groups were preloaded but lower when neither group received a preload. A simple main effects analysis (Huck, Cormier & Bounds, 1976) indicated that the unrestrained preloaded group had significantly lower taste ratings than the unrestrained non-preloaded group (_t(28) = 2.25, p < .05) and the restrained preloaded group (_t(28) = 3.21, p < .01). Based on the statistical analysis, the null hypothesis of no differences between groups was rejected. Further examination of the cell means revealed that the four groups' ratings of taste ranked in the identical order to the ranking found for food consumption (see Table 4). Restrained preloaded subjects had the highest ratings of food taste followed by unrestrained non-preloaded subjects. Restrained non-preloaded subjects ranked third. The lowest ratings came from the unrestrained preloaded group, which differed significantly in the investigation of main effects. For the dependent variable, buying intention, the following null hypothesis was tested: 3. There will be no differences between subjects in the four groups (restrained preloaded, unrestrained preloaded,

PAGE 76

70 Mean Ratings of Taste on Taste Preference Sheet 25 20 15 10 Restrained Unrestrained Preloaded Non-preloaded Level of Preload Condition Figure 1 Graph of Restraint X Preload Interaction for Taste Ratings

PAGE 77

Table 4 Means and Standard Deviations of Taste Ratings 71 Group Mean Scores (rank) SD Preloaded Restrained 15 Unrestrained 15 22.80 (1) 19.07 (4) 2.97 3.39 Non-pre loaded Restrained 15 Unrestrained 15 21.27 (3) 22.13 (2) 2.74 3.91

PAGE 78

72 Table 5 Analysis of Variance for Buying Intention Source

PAGE 79

73 restrained non-preloaded, unrestrained non-preloaded) in their ratings of buying intention for the food consumed. A two-way (2x2) analysis of variance was conducted for buying intention using the factors, restraint and preload (see Table 5). No significant main effects were found. Contrary to expectations, a significant interaction between restraint and preload existed (F(l,56) = 5.60, p < .05), indicating that the difference between the levels of one factor changes from one level to another level (see Figure 2). As occurred with taste ratings, restrained subjects scored higher on buying intention than unrestrained subjects when both groups were preloaded and lower when neither group was preloaded. A simple main effects analysis indicated that the restrained non-preloaded group had significantly lower ratings of buying intention than the unrestrained non-preloaded group (_t(28) = 2.20, p < .05) and the restrained preloaded group (t(28) = 2.29, p < .05). Based on the statistical analysis, the null hypothesis of no differences between groups was rejected. As in the previous examination of cell means, restrained preloaded subjects ranked highest, reporting the greatest buying intention, followed by the unrestrained non-preloaded group. However, the order was reversed for the other two groups. Unrestrained preloaded subjects ranked third in buying intention and restrained non-preloaded subjects had the lowest ratings of buying intention (see Table 6).

PAGE 80

74 Table 6 Means and Standard Deviations of Buying Intention Group Mean Scores (rank) SD Preloaded Restrained 15 Unrestrained 15 15.40 (1) 13.07 (3) 6.32 4.83 Non-preloaded Restrained 15 Unrestrained 15 10.20 (4) 14.80 (2) 5.28 6.12

PAGE 81

75 Mean Ratings of Buying Intention on Taste Preference Sheet 16 12 Unrestrained Restrained Preloaded Non-preloaded Level of Preload Gondition Figure 2 Graph of Restraint X Preload Interaction for Buying Intention

PAGE 82

76 Surmary The following list summarizes the results found in the study: 1. On the dependent variable, food consumption, no main effects were found for restraint or preload and no interaction of these two independent variables occurred. 2. On the dependent variable, taste rating, no main effects were found for restraint or preload. 3. A significant interaction (p < .01) between restraint and preload was found on the dependent variable, taste rating. The interaction indicates that preloaded restrained subjects rated the taste of food higher than non-preloaded restrained subjects and the opposite order occurred with unrestrained subjects. The unrestrained preloaded group was significantly different than the restrained preloaded and the unrestrained non-preloaded groups. 4. On the dependent variable, buying intention, no main effects were found for restraint or preload. 5. A significant interaction (p < .05) between restraint and preload was found on the dependent variable, buying intention. This interaction indicates that preloaded, restrained subjects had higher buying intention than non-preloaded restrained subjects and that the opposite order occurred with unrestrained subjects. The restrained nonpreloaded group was significantly different than the restrained preloaded and unrestrained non-preloaded groups.

PAGE 83

CHAPTER FIVE SUMMARY, DISCUSSION, CONCLUSIONS, IMPLICATIONS AND RECOMMENDATIONS Summary This study examined the effects of restraint level and presence of preload on the amount of food eaten and the ratings of food given by subjects expecting a period of food deprivation. Prior work with dietary restraint (Herman & Mack, 1975; Hibscher & Herman, 1977; Spencer & Fremouw, 1979) demonstrated that subjects with high dietary restraint consume more food in a taste test simulation if they have ingested a food preload perceived to be high in calories. Lowe (1982) preloaded all subjects in her investigation and found that those who anticipated a period of food deprivation ate significantly more than those who did not. She concluded that both internal factors (e.g., restraint level) and situational variables (e.g., anticipated deprivation) play a role in overconsumption Lowe's (1982) study did not test if anticipation of deprivation was a sufficient factor to cause restrained eaters to overconsume without first ingesting a preload of restraint-breaking "forbidden" food. This question was of primary interest to this investigation. In this study, restrained and unrestrained subjects, all of whom were expecting a period of food deprivation, were randomly assigned to one of two preload conditions, preloaded or non-preloaded, creating four equal groups of subjects. The amount of food consumed during a free eating period was 77

PAGE 84

78 the dependent variable measured. In addition, the following two cognitive dependent variables were examined; taste ratings and buying intention. Both were measured via the use of a self-report questionnaire (see Appendix C) The experimental procedure utilized in the study was an extension of the methodology employed by Lowe (1982). Specifically, undergraduate females who had previously completed the Restraint Questionnaire (Polivy et al., 1976) signed up for an experiment titled "Sensation and Perception." Prior to the subject's arrival for the experiment it was noted whether she was a restrained or unrestrained eater (i.e., high or low preoccupation with food and body size) based on her Restraint Questionnaire score. Subjects were randomly assigned to one of two preload conditions (preloaded; nonpreloaded) producing a 2 x 2 (Restraint X Preload) factorial design. All subjects were asked to make a variety of food ratings after being informed that they must fast for the four hour period following the morning session of the experiment. They were left alone for 15 minutes during which time they could eat as much food as desired while they made their ratings. Those in the preloaded condition were given a standard amount of food to be eaten prior to this 15-minute free eating period. Food consumption was measured by weighing the three bowls of food (M&M's, peanuts and animal crackers) before and after the free eating period and computing the differences. Both the ratings of taste and buying intention were taken from items on the Taste Preference Sheet, a questionnaire developed specifically for this study. A preliminary testing of this instrument was conducted with 15 restrained and 15

PAGE 85

79 unrestrained subjects to investigate any initial differences that might exist between restrained and unrestrained subjects. No significant differences were found. For each of the dependent variables, a two-way (2x2) analysis of variance was conducted. No main effects were found for any of the three dependent variables. For two of the variables, taste ratings and buying intention, significant Restraint X Preload interactions were found. A test of simple main effects was conducted on each significant interaction. Discussion The results of this experiment failed to support the prediction that there would be significant differences between the groups in the amount of food consumed during the free eating period. It was expected that restrained subjects in both levels of the preload condition (preload; non-preloaded) would consume significantly more than unrestrained subjects randomly assigned to the same levels. Contrary to expectation, no differences in consumption were found between the groups. Therefore, it cannot be concluded from the results of this study that, in the absence of a restraint breaking preload, restrained subjects who expect a period of food deprivation will break their dietary restraint and overconsume. Past studies in dietary restraint found that the preload of "forbidden" food was a necessary factor in causing restrained subjects to eat more than unrestrained subjects. Restrained subjects who did not ingest a preload were found to consume less than the unrestrained

PAGE 86

80 subjects. This Restraint X Preload interaction was an indicator of counter-regulatory eating by restrained subjects. It has been hypothesized that restrained eaters maintain strict dietary standards until a "slip" occurs (e.g., ingestion of a highly caloric preload). After this initial "slip" occurs, restrainers abandon their standards and overconsume. If no dietary transgression occurs, restrained eaters remain faithful to their standards and underconsume Unrestrained eaters, with little concern about food, are found to naturally regulate their intake (e.g., unrestrained subjects who have ingested a preload eat less during the free eating period than when they have not received a preload) Lowe (1982) preloaded all her subjects to test the effects of restraint level and anticipation of deprivation on consumption. She concluded that both independent variables were factors in the overconsumption exhibited by her subjects. Lowe hypothesized that in her taste test simulation the anticipation of deprivation may have been sufficiently powerful to obscure the effect of restraint, which failed to reach significance. Based on these past results, it was expected that if anticipated deprivation was a sufficient factor in the breakdown of dietary restraint (i.e., a preload was no longer necessary), then all restrained subjects, preloaded and non-preloaded, would overconsume. Conversely, if anticipated deprivation was not a sufficient factor, then the preload would remain a necessary one, consistent with past findings. In this case, the restrained preloaded subjects would significantly overconsume and the non-preloaded subjects would consume the least. In actuality, the results of this study indicate that neither of these possible expectations occurred.

PAGE 87

81 With no single group mean found to be significantly different, anticipated deprivation cannot be judged, in the absence of a preload, to be a sufficient factor in overconsumption by restrained eaters. The same reason also prevents this study from supporting a reaffirmation of past studies' finding that a food preload is a necessary factor in causing counter-regulatory eating by restrained subjects. Counterregulatory eating would have been demonstrated had the restrained preloaded subjects eaten more than both groups of unrestrained subjects and had the restrained non-preloaded group eaten less than all other subjects. In this study, restrained subjects ate no more or less than unrestrained subjects. In seeking an explanation for these results, two avenues of investigation must be pursued. First, did the experimental procedures successfully create the conditions desired for the study, and thereby, remain consistent with the methodology of past studies? Second, what factors may account for the results found in this study? A comparison of this study with prior research may best facilitate this examination. A review of the experimental methodology does not reveal any procedural differences between the present study and past investigations of dietary restraint and anticipated deprivation that would affect the results unpredictably. The materials, time periods, instructions and type of subject used in the experiment closely paralleled those of prior studies and attempted to replicate the relevant aspects of Lowe's (1982) work. Responses to the Subject Questionnaire (see Appendix D) did not indicate that subjects were aware of the true purpose of the study. All subjects reported their good intentions to attempt to fast for the four hour period as requested.

PAGE 88

82 In Lowe's (1982) study, all subjects were preloaded to ensure an initial breakdown of dietary restraint. In her taste test simulation she found a main effect for anticipation of deprivation and reported a trend for restraint level. Lowe hypothesized that her situational variable, anticipated deprivation, may have been powerful enough to obscure the usual effects of restraint level on preloaded subjects. After finding a significant main effect for restraint in her learning task simulation, Lowe concluded that overeating is better understood when both restraint and anticipation of deprivation are taken into account. She believed that if the restraint breaking preload could be excluded as a necessary factor the implications of her deprivation theory for restrained eaters would be more far-reaching. In attempting to answer Lowe's question this study took one level of Lowe's independent variable, anticipated deprivation, and made it a fixed condition of the experiment. Preload, which had a fixed condition of Lowe's study, became an independent variable with two levels, preloaded and non-preloaded. Restraint level was continued as the second independent variable. Except for the introduction of anticipation of deprivation as a fixed condition, the present study's design closely paralleled the original research in dietary restraint and, thereby, created the desired experimental conditions. Several possibilities exist to explain the results of the study. One is that anticipation of deprivation had no effect on subjects and the lack of significant differences between the groups indicates that restraint level and preload are not factors in causing overconsumption. Given the extensive research in dietary restraint showing a significant Restraint X Preload interaction, this is an unlikely explanation of the results.

PAGE 89

83 A second alternative which may, in part, explain the results is that the groups did not differ because all subjects expected a period of food deprivation. If this alternative is true, then anticipation of deprivation could, as originally suggested by Lowe, be powerful enough to obscure the effects of restraint level. At the same time, the preload becomes unnecessary because the restrained subjects have already abandoned their strict dietary standards in response to the anticipated deprivation. This would imply that Lowe would have found the same results with subjects who were not given a preload. This second alternative seems to be a more feasible explanation for the disparate results of this study. However, the fact that Lowe did find a significant main effect for restraint warrants caution in accepting the above explanation of the results of this study. A method which permits a more thorough investigation of the premises of the two studies is discussed in the Recommendations section of this chapter. Since previous studies had focused on food consumption as a dependent variable, the investigation of the attitudes of subjects about the food consumed was an important addition to a study examining dietary restraint and anticipation of deprivation. The present study's examination of the two cognitive dependent variables, taste ratings and buying intention, generated results which parallel the findings of earlier studies of food consumption. No significant main effects were found for the independent variables, restraint and preload, in previous studies nor were any found in the present study. However, both taste ratings and buying intention were found, in the present study, to be significantly affected by the interaction of the restraint and preload independent variables. In both cases, restrained subjects who were

PAGE 90

84 preloaded had higher ratings of taste and buying intention than restrained subjects who were not preloaded, but unrestrained preloaded subjects had lower ratings of both variables than unrestrained subjects who were not preloaded. The significant results for taste ratings and buying intention indicate that even though subjects' eating behavior did not differ from each other, their thoughts about the attractiveness of food did. The differential effect of the preload on the cognitions of the restrained and unrestrained subjects seems inconsistent with the earlier hypothesis that the anticipation of deprivation obscured the effects of restraint level and preload. A re-examination of the experimental procedures may assist in resolving this contradiction. Subjects provided ratings of taste and buying intention during the 15-minute free eating period. If they completed the questionnaire during the first 2-4 minutes of the period their responses might reflect the influence of restraint level and preload. For example, restrained non-preloaded subjects may enter the free eating period with the resolve to control their eating and complete the ratings which this resolve is still intact. However, as time passes and the period of fasting approaches, they abandon their restraint and consume in a similar fashion to preloaded subjects. The ratings of unrestrained subjects could also be influenced by the independent variables if these subjects completed the questionnaires before they finished consuming the food. It was expected that the groups' ratings of taste would parallel the expected results of the food consumption variable. Restrained preloaded subjects were expected to provide the highest ratings of taste, followed by the restrained non-preloaded group.

PAGE 91

85 Unrestrained non-preloaded subjects were expected to rank third and unrestrained preloaded subjects were expected to have the lowest ratings of taste. The above ranking was predicted on the premise that when expecting a period of deprivation, restrained eaters, who usually view highly caloric food as desirable but forbidden, would rate the taste of food higher than unrestrained subjects. It was thought that the restrained preloaded subjects would have higher ratings than restrained nonpreloaded subjects because of the combined influence of the anticipated deprivation and the food preload on the former. Unrestrained preloaded subjects were expected to rank lowest because they would dislike being mandated to eat the preload and subsequently rate the taste lower than the unrestrained subjects who only ate during the free eating period. The predicted main effect for restraint and the expected ranking of the four groups were not supported by the results of the taste ratings data analysis. Instead, a significant Restraint X Preload interaction was found. This interaction indicates that restrained subjects had higher taste ratings when preloaded than when they were not preloaded but unrestrained subjects had lower taste ratings when preloaded than when they did not receive a preload. One possible explanation for the differential response exhibited by restrained and unrestrained subjects is that the preload caused different cognitions in the two types of eaters. Restrained subjects who have been forced to break their dietary restraint with the ingestion of the preload may attempt to justify this usually forbidden activity by elevating their taste ratings. Restrained non-preloaded subjects who are completing their ratings while freely sampling the food, may not

PAGE 92

86 experience this need to justify their behavior. Unrestrained subjects may have a different reaction to the preload. With low preoccupation with food and body size, unrestrained subjects who are preloaded may rate the taste of food lower as a way to prevent overconsumption or as an indication of their displeasure about being mandated to eat. If this explanation was accepted as true, it would imply that restrained and unrestrained eaters may differ in their cognitions about food when they feel their ability to control their consumption has been diminished. A test of simple main effects was conducted on the results of the taste ratings of the four groups. The unrestrained preloaded group was found to have significantly lower taste ratings than the restrained preloaded and the unrestrained non-preloaded groups. Two deductions may be drawn from this result. First, restrained and unrestrained subjects have different taste ratings when both are preloaded. Second, the taste ratings of unrestrained subjects who are preloaded differ from those of unrestrained subjects who are not preloaded. Both deductions lend support to the above explanation of the results of the taste ratings. The predicted main effect for restraint and the expected ranking of the four groups were not supported by the results of the buying intention data analysis. Instead, a significant Restraint X Preload interaction was found. This interaction indicates that restrained subjects who were preloaded had higher ratings of buying intention than restrained non-preloaded subjects, but unrestrained subjects who were preloaded had lower ratings than unrestrained non-preloaded subjects. Contrary to the group rankings found with the consumption and taste rating variables, unrestrained preloaded subjects ranked third in buying intention and restrained non-preloaded subjects provided the lowest ratings of buying intention.

PAGE 93

87 An important distinction between the cognitive variables taste ratings and buying intention, is that buying intention represents what subjects believe they will do in the future, while taste ratings deal with a current state. Therefore, subjects are responding about a time frame that occurs long after the experiment and the period of deprivation have passed. However, this distinctive feature of the dependent variable does not provide a sufficient explanation for either the significant interaction or the group rankings. One possible explanation for the buying intention results is that the restrained subjects exhibited a pattern similar to that which is hypothesized to occur in counter-regulatory eating. Restrained eaters, whose dietary restraint has been broken by the preload, abandon their strict standards and overconsume. Conversely, restrained eaters who do not receive preloads maintain their strict dietary standards and consume less than others. It is possible that in an analogous fashion, restrained preloaded subjects increased their ratings of buying intention while restrained non-preloaded subjects decreased their ratings. When preloaded, restrained subjects anticipating deprivation may feel out of control of their intake and expect a greater likelihood of purchasing forbidden food in the near future. Non-preloaded restrained subjects may react to the challenge of anticipated deprivation with decreased ratings of buying intention. For reasons similar to those proposed to explain the taste ratings results, unrestrained subjects may have responded differently. Disliking the mandatory preload or attempting to prevent a disruption of their natural balance, the restrained preloaded subjects may have lowered their ratings of buying intention for the food consumed.

PAGE 94

88 A test of simple main effects was conducted on the results of the buying intention ratings of the four groups. The restrained non-preloaded group was found to have significantly lower ratings of buying intention than the restrained preloaded and the unrestrained non-preloaded groups. This indicates that restrained subjects who were preloaded had significantly higher ratings of buying intention than restrained subjects who were not preloaded. It also shows that restrained subjects provided signif icatnly lower scores than unrestrained subjects when neither group was preloaded. These deductions support the contention of a differential effect of the preload on the buying intention of restrained and unrestrained subjects. The results of this study shed some light on and generated many questions concerning the factors involved in overconsumption. CTie purpose of this study was to examine the necessity of a restraint breaking preload in causing overconsumption by restrained eaters who expect a preload of food deprivation. The lack of significant results tor eating behavior in this investigation raises two possibilities. The first is that when subjects are expecting a period of food deprivation, the independent variables, restraint and preload, have no influence on subjects' consumption. This hypothesis seemed unlikely in consideration of the numerous studies showing significant results for these variables. The second possibility is that anticipation of deprivation had an effect that cannot be measured in this study. Oie hypothesis is that the anticipation of deprivation was so powerful as to cause increased consumption in both restrained and unrestrained eaters. If this were to be supported in future research, this hypothesis would have important implications for the concept of dietary restraint and the treatment of binge eating.

PAGE 95

89 The addition of the non-behavioral measures, taste ratings and buying intention, revealed significant results and raised many questions concerning a new line of inquiry, the cognitions of subjects who break their dietary restraint. These results were found on the non-behavioral measures even though consumption was not found to differ between the groups. On both dependent variables, a significant Restraint X Preload interaction was present. Restrained subjects who were preloaded were found to have higher ratings of taste and buying intention than non-preloaded restrained subjects but unrestrained preloaded subjects had lower ratings than unrestrained subjects who were not preloaded. One of the limitations of the concept of dietary restraint is that it established the food preload as a necessary factor in overconsumption by restrained eaters and does not offer an understanding of how restraint is broken outside the laboratory. Lowe's (1982) study introduced a situational variable, anticipated deprivation, and found that it was a factor to be considered in overconsumption. Investigation of the effects of sucn situational variables however, has exclusively focused on the behavioral measure, food consumption. The significant results found in this study for taste and buying intention indicated that investigation of the cognitions of the eater may also provide insight into factors causing differential beliefs about food and different patterns of consumption. Conclusions 1. No significant differences in food consumption were found between any of the four groups (restrained preloaded, restrained nonpreloaded, unrestrained preloaded, unrestrained non-preloaded).

PAGE 96

90 2. Restrained subjects provided higher ratings of taste when preloaded than when they were not preloaded. Conversely, unrestrained subjects provided lower taste ratings when preloaded than when they were not preloaded. Unrestrained preloaded subjects had significantly lower taste ratings than subjects in both the restrained preloaded and the unrestrained non-preloaded groups. 3. Restrained subjects provided higher ratings of buying intention when preloaded than when they were not preloaded. Conversely, unrestrained subjects provided lower ratings of buying intention when preloaded than when they were not preloaded. Restrained non-preloaded subjects had significantly lower ratings of buying intention than both the restrained preloaded and the unrestrained non-preloaded groups. Implications In failing to support the contention that the preload is not necessary to cause overconsumption in restrained eaters who anticipate a period of deprivation, the results of this study limit the number and scope of implications that may be drawn by the investigator. It was expected that the results of the study would make the relationship between binge eating and food deprivation more understandable. Orbach (1978a) has posited that binge eaters overeat when future availability of food is unknown and Roth (1982) believes that binge eating is a response to scarcity. These similar hypotheses, which were based on clinical impressions, cannot be supported by the inconclusive findings of this study.

PAGE 97

91 Despite the non-significant results found for food consumption, the significant findings for taste ratings and buying intentions do suggest that cognitions about food differed for seme subjects in the study. This finding implies that the inclusion of other dependent variables (e.g., cognitions about food) in future studies may provide a more complete understanding of the complex problem of dietary restraint. Investigations into the cognitive components of overconsumption may also produce implications for treatment. Wilson (1976) has questioned the efficacy of standard behavioral weight control programs for binge eaters. He suggests that these programs fail to intervene on a cognitive level and therefore, the thoughts that trigger a binge are never changed in therapy. The identification of those cognitive variables that influence overconsumption could assist in the development of more appropriate treatment strategies for binge eating. Recommendations for Future Research 1. One of the hypotheses generated to explain the results of this study is that the anticipation of deprivation experienced by all subjects obscured the usual effects found for the restraint and preload variables. The investigation of this question could be conducted through a combination of the methodology of this study and Lowe's (1982) work. A three way (2x2x2) factorial design that utilized anticipation of deprivation, restraint and preload as the independent variables would permit the analysis of the possible main and interactive effects of all three factors.

PAGE 98

92 2. The reliance in earlier studies on one dependent variable (food consumption) limits the implications that can be drawn from reseasrch in dietary restraint. The addition of two cognitive dependent variables in this study yielded significant results about the subjects' taste and buying preferences. In recognition of the common sense notion that individuals must decide to give up their restraint and overeat, it is reasonable to investigate the cognitions that accompany the behaviors exhibited by subjects on future studies. 3. If cognitive measures are included in future studies, it is recommended that their placement within the experimental procedures be carefully chosen. A questionnaire distributed at the beginning of the free eating period, for example, may receive different responses than one distributed at the end of the period, if subjects choose to break their dietary restraint during the free eating period. 4. The review of the literature conducted for this study found many studies examining the personality variables associated with obese eating styles. Similar investigations into the personality characteristics of restrained and unrestrained eaters have not been reported in the literature. The comparison of the Restraint Questionnaire (Polivy et. al., 1976) with instruments that measure variables hypothesized to be influential in overconsumption (e.g., depression, locus of control, body image) may have valuable implications for treatment of binge eating. 5. Lowe's methodology allowed a situational variable usually experienced by dieters in daily life to be tested in analog fashion within the confines of a laboratory environment. Similar methods are needed to test other situational variables that may contribute to overeating (e.g., boredom, social anxiety, negative feedback).

PAGE 99

APPENDICES

PAGE 100

APPENDIX A RESTRAINT QUESTIONNAIRE 1. How often are you dieting? a) never b) rarely c) sometimes d) usually e) always 2. What is the maximum amount of weight (in pounds) you have ever lost within one month? a) 0-4 b) 5-9 c) 10-14 d) 15-19 e) 20+ 3. What is your maximum weight gain within a week? a) 0-1 b) 1.1-2 c) 2.1-3 d) 3.1-5 e) 5.1+ 4. In a typical week, how much does your weight fluctuate? a) 0-1 b) 1.1-2 c) 2.1-3 d) 3.1-5 e) 5.1+ 5. Would a weight fluctuation of 5 lb. affect the way you live your life? a) not at all b) slightly c) moderately d) very much 6. Do you eat sensibly in front of others and splurge alone? a) never b) rarely c) often d) always 7. Do you give too much time and thought to food? a) never b) rarely c) often d) always 8. Do you have feelings of guilt after overeating? a) never b) rarely c) often d) always 9. How conscious are you of what you're eating? a) not at all b) slightly c) moderately d) extremely 10. How many pounds over your desired weight were you at your maximum weight? a) 0-1 b) 1-5 c) 6-10 d) 11-20 e) 21+ Polivy, J., Howard, K. I., and Herman, C. P. (1976). Psychometric analysis of the restraint scale Unpublished manuscript. University of Toronto 94

PAGE 101

APPENDIX B MARKETING QUESTIONNAIRE Please respond to the questions below by indicating to the right of the item one of the following choices: a) less than once/month b) 1-3 times/month c) once/week d) 2-6 times/week e) once/day or more A. How often do you purchase food at the following grocery stores? 1. Publix 2. Albertson's 3. Winn-Dixie 4. Pic 'n Save 5. Pantry Pride 6. Kash 'n Karry 7. Jewel T 8. (other-please specify) B. How often do you purchase food at each of the following convenience stores? 9. 7-11 10. Spring 11. Majik Market 12. Pablo's Tunnel 13. Jiffy 14. Min-A-Mart 15. Lil' Champ b5

PAGE 102

96 C. How often do you purchase a meal at one of the following restaurants? 16. McDonald's 17. Burger King 18. Wendy's 19. Arby's 20. Rax 21. In 'n Out 22. Krystal D. How often do you eat each of the following as the main component of a meal? 23. Salad 24. Pizza (frozen or delivered) 25. Frozen dinner (other than pizza) E. How otten do you purchase each of these brand name products? 26. Campbell's soups 27. Poptarts 28. Kraft American Cheese 29. Coca-cola, Tab or Sprite 30. Wonder Bread 31. Green Giant canned or frozen vegetables 32. Ragu spaghetti sauce 33. Gustafson's Milk 34. Wise potato chips 35. Oscar Mayer hot dogs (tranks)

PAGE 103

APPENDIX C TASTE PREFERENCE SHEET Please answer the following questions about the M&M's ; 1. How would you rate the "sweetness" of the food: not sweet 123456789 10 very at all sweet 2. How would you rate the saltiness of the food? not salty 12 3456789 10 very salty 3. How would you rate the taste of the food overall? bad 1 2 3 4 5 6 7 8 9 10 great 4. How would you rate the moistness of the food? very dry 123456789 10 very moist 5. How would you rate the bitterness of the food? not bitter 123456789 10 very bitter at all 6. How likely are you to buy this food or a similar food item within the next two weeks? very 123456789 10 very likely unlikely 7. How strong was the flavor of the food? no flavor 12 3 4 5 6 7 at all 10 very strong flavor 8. How likely are you to be thirsty after eating this food? very unlikely 9 10 very likely 9. How would you rate the consistency of the food? soft 123456789 10. How strong is the aftertaste of the food? no after123456789 taste 10 hard 10 very strong Please answer the following questions about the peanuts : 1. How would you rate the "sweetness" of the food? not sweet at all 1 4 5 97 8 10 very sweet

PAGE 104

98 2. How would you rate the saltiness of the food? not salty 123456789 10 very salty 3. How would you rate the taste of the food overall? bad 1 2 3 4 5 6 7 8 9 10 great 4. How would you rate the moistness of the food? very dry 123456789 10 very moist 5. How would you rate the bitterness of the food? not bitter 123456789 10 very bitter at all 6. How likely are you to buy this food or a similar food item within the next two weeks? very 123456789 10 very likely unlikely 7. How strong was the flavor of the food? no flavor 12 3 4 5 6 7 at all 10 very strong flavor 8. How likely are you to be thirsty after eating this food? very unlikely 9 10 very likely 9. How would you rate the consistency of the food? soft 123456789 10. How strong is the aftertaste of the food? no after123456789 taste 10 hard 10 very strong Please answer the following questions about the animal crackers : 10 very sweet 2. How would you rate the saltiness of the food? 1. How would you rate the "sweetness" of the food? not sweet 123456789 at all not salty 8 3. How would you rate the taste of the food overall? bad" 123456789 4. How would you rate the moistness of the food? very dry 123456789 5. How would you rate the bitterness of the food? not bitter 123456789 at all 10 10 10 10 very salty great very moist very bitter

PAGE 105

99 6. How likely are you to buy this food or a similar food item within the next two weeks? very 123456789 10 very likely unlikely 7. How strong was the flavor of the food: no flavor 123456789 10 very strong at all flavor 8. How likely are you to be thirsty after eating this food? very 123456789 10 very likely unlikely 9. How would you rate the consistency of the food? soft 123456789 10 hard 10. How strong is the aftertaste of the food? no after123456789 10 very strong taste

PAGE 106

APPENDIX D SUBJECT QUESTIONNAIRE 1. Were you aware of the true nature of this study prior to your arrival at the experimental session today? a) Yes b) No 2. Did you intend to fast for the four hour period between sessions as requested by the experimenter? a) Yes, definitely b) Yes, if I was able c) No 3. Prior to arriving at the session, when did you eat your last meal? a) less than one hour ago b) 1-2 hours ago c) 3-4 hours ago d) more than 4 hours ago 4. If you were asked to eat while completing the Marketing Questionnaire (questions about where you shop) did you eat the food given you? a) yes b) part of it c) no d) I wasn't asked to eat during this part of the experiment 5. Please provide the following information: Age Weight Height Body Frame (circle one) SMALL MEDIUM LARGE 100

PAGE 107

APPENDIX E GENERAL INSTRUCTIONS FOR EXPERIMENTERS The following statements should be used in the running of the experiment: Taste Preference Sheet Pre-test This study is investigating the taste perceptions of undergraduate women. Before you are three bowls of food. I would like you to taste food from each bowl and complete the ratings sheet provided. You will be given 15 minutes to complete the ratings. After this time period, I will return to collect the questionnaire. (Experimenter sets timer for 15 minutes, points it toward subject and leaves the room.) Treatment Procedures with Preloaded Subjects Part One This is a study of the effects of hunger on taste. During both sessions today you will be asked to rate three common foods. In order to allow us to assess the effects of hunger on your ratings, you must not eat between the end of the morning session and the beginning of the afternoon a total of four hours. Do you understand these instructions? Before we begin the taste ratings, I would like you to complete this Marketing Questionnaire (give to subject) While responding to the questionnaire you are to eat a preliminary amount of food (give subject a plate of 10 M&M's, 10 peanuts and 3 animal crackers). You have 10 minutes to finish this portion of the experiment. 101

PAGE 108

102 Part Two (Collect the Marketing Questionnaire) Here are three bowls, one each of M&M's, peanuts and animal crackers. Please rate the foods on the Taste Preference Sheet (give subject ratings sheet) You may eat as little or as much of the food as you wish. To assess the effect of hunger, we will take these ratings and compare them to the ratings you will give after your four hour fast. I will return in 15 minutes to collect your ratings sheet (set timer, point it toward subject and leave room) Treatment Procedures with No-Preload Subjects Part One This is a study of the effects of hunger on taste. Daring both sessions today you will be asked to rate three common foods. In order to allow us to assess the effect of hunger, you must not eat between the end of the morning session and the beginning of the afternoon session a total of four hours. Do you understand these instructions? Before we begin the taste ratings, I would like you to complete this Marketing Questionnaire (give to subject) You have ten minutes to finish this portion of the experiment. Part Two Here are three bowls, one each of M&M's, peanuts and animal crackers. Please rate the foods on the Taste Preference Sheet (give ratings sheet to subjects). You may eat as little or as much of the foods as you wish. To assess the effect of hunger, we will take these ratings and compare them to the ratings you will give after your four hour fast (set timer, point it toward subject and leave room). I will return in 15 minutes to collect your ratings sheet (set timer, point it toward subject and leave room)

PAGE 109

103 De-briefing of All Treatment Subjects Thank you for your participation in this experiment. Although we have scheduled you for an afternoon session, you will not need to attend it. You are also released from our request to fast for four hours This study is not an investigation of the effects of hunger on taste, it is an examination of the effect of expecting to fast (anticipation of food deprivation) on eating behavior and attitudes toward food. Before leaving, please complete this questionnaire. Truthful answers will help us in our analysis of the data we have gathered in the experiment (give Subject Questionnaire) It is important to the success of this study that the true nature of its procedures remain confidential. Please do not discuss this experiment with anyone who may be a potential subject. (Collect Subject Questionnaire) I will return in 15 minutes to collect your ratings sheet (set timer, point it toward subject and leave room)

PAGE 110

APPENDIX F INFORMED CONSENT FORM The present study will be an examination of the effects of hunger on taste perception. During the course of this study you will be asked to consume chocolate, peanuts and animal crackers. Questionnaires will be distributed for your completion. They will be composed of questions concerning your normal shopping habits and your preferences about the food you have tasted in this experiment. Some subjects will be asked to fast tor the four hour period following the session and to return in the afternoon. There are no benefits expected to result from participation in this study other than credit toward the research requirements of Introductory Psychology. The use of food (specifically chocolate, peanuts and animal crackers) presents a risk to those persons who experience reactions to these foods. Any subject with food allergies, hypoglycemia, or diabetes and anyone who reacts with dizziness and/or headaches when a meal is missed should not participate in this study and will be excused from the study with no penalty. There will be no monetary compensation for participation in this study. Any subject is free to withdraw his/her consent to participate at any time without prejudice. Questions concerning any aspect of this study may be directed to the Principal Investigator named below. I have read and I understand the procedure described above. I agree to participate in the procedure and I have received a copy of this description. I, to the best of my knowledge, do not have food allergies, diabetes or hypoglycemia nor do I get dizzy or headaches when I miss a meal. Subject Date Witness Date Principal Investigator: Edward Spauster Psychology Bldg. Box 105 (904) 392-6636 104

PAGE 111

REFERENCES Abramson, E., and Stinson, S. (1977). Boredom and eating in obese and non-obese individuals. Addictive Behaviors 2, 181-185. Beaumont, P.J., George, G.C., and Smart, D.E. (1976). "Dieters" and "vomiters and purgers" in anorexia nervosa. Psychological Medicine 6, 617-632. Boskind-Lodahl, M. (1976). Cinderella's stepsister: A feminist perspective on anorexia nervosa and bulimia. Signs: Journal of Women in Culture and Society 2, 342-356. Boskind-Lodahl, M. and Sirlin, J. (1977). The gorging-purging syndrome. Psychology Today March, 50-52, 82-85. Boskind-Lodahl, M., and White, W. (1978). The definition and treatment of bulimarexia in college women — a pilot study. Journal of American College Health Association 27 84-87. Brenner, M. (1980). Bulimarexia. Savvy Magazine June, 54-59. Brightwell, D.R., and Sloan, C.L. (1977). Long term results of behavior therapy for obesity. Behavior Therapy 8_, 898-905. Bruch, H. (1961). Transformation of oral impulses in eating disorders: A conceptual approach. Psychiatric Quarterly 35 458-81. Bruch, H. (1973). Eating disorders New York: Basic Books. Button, E.J., Fransella, F., and Slade, P.D. (1977). A reappraisal of body perception disturbance in anorexia nervosa. Psycho log ical Medicine 7, 235-243. Casper, R., Eckert, E., Halmi, K., Goldberg, S., and Davis, J. (1980). Bulimia: Its incidence and clinical importance in patients with anorexia nervosa. Archives of General Psychiatry 37 1030-1035. Crisp, A.H. (1965). Clinical and therapeutic aspects of anorexia nervosa — a study of 30 cases. Journal of Psychosomatic Research 9, 67-68. Crisp, A.H. (1967). The possible significance of some behavioral correlates of weight and carbohydrate intake. Journal of Psychosomatic Research 11 117-131. 105

PAGE 112

106 Dahlkoetter, J., Callahan, E.J., and Linton, J. (1979). Obesity and the unbalanced energy equation, exercise vs. eating habit change. Journal of Consulting and Clinical Psychology 47, 898-905. Dally, P., and Gomez, J. (1979) Anorexia nervosa London: Heinemann Medical. DSM-III, (1980). American Psychiatric Association: Washington, D. C. Fairburn, C. (1981). A cognitive behavioral approach to the treatment of bulimia. Psycholgical Medicine 11 707-711. Garfinkel, P.E., and Garner, D.M. (1982). Anorexia nervosa: A multidimensional perspective New York: Brunner/Mazel. Garfinkel, P.E., Moldofsky, H., and Garner, D.M. (1980). The heterogeneity of anorexia nervosa: Bulimia as a distinct subgroup. Archives of General Psychiatry 37 1036-1040. Garner, D.M., Garfinkel, P.E., Schwartz, D. and Thompson, M. (1980). Cultural expectations of thinness in women. Psychological Reports 47, 483-491. Gormally, J., Black, S., Daston, S., and Rardin, D. (1982). The assessment of binge eating severity among obese persons. Addictive Behaviors 1_, 47-55. Gormally, J., Rardin, D. and Black, S. (1980). Correlates of successful response to a behavioral weight control clinic. Journal of Counseling Psychology 27, 179-191. Green, R.S., and Rau, J.H. (1974). Treatment of compulsive eating disturbances with anticonvulsant medication. American Journal of Psychiatry 131 428-432. Griggs, R., and Stunkard, A.J. (1964). The interpretation of gastric motility: II. Sensitivity and bias in the perception of gastric motility. Archives of General Psychiatry 11, 82-89. Gwinup, G. (1975). Effect of exercise alone on the weight of obese women. Archives of Internal Medicine 135 676-680. Hall, S.M., and Hall, R.G. (1974). Outcome and methodological considerations in behavioral treatment of obesity. Behavior Therapy 5, 352-364. Halmi, K.A. (1974). Anorexia nervosa: Demographic and clinical features in 94 cases. Psychosomatic Medicine 36 18-26. Halmi, K.A., Falk, J.R., and Schwartz, E. (1981). Binge-eating and vomiting: A survey of a college population. Psychological Medicine, 11, 696-706.

PAGE 113

107 Hawkins, R.C. (1979). Meal/snack frequencies of college students: A normative study. Behavioral Psychotherapy 1_, 85-89. Hawkins, R.C, and Clement, P. (1980). Development and construct validation of a selfreport measure of binge eating tendencies. Addictive Behaviors 5_, 219-226. Herman, C.P., and Mack, D. (1975). Restrained and unrestrained eating. Journal of Personality 43 647-669. Herman, C.P., and Polivy, J. (1975). Anxiety, restraint and eating behavior. Journal of Abnormal Psychology 84 662-672. Herman, C.P., Polivy, J., Pliner, P., Threlkeld, J., and Munic, D. (1978). Distractibility in dieters and nondieters: An alternative view of "externality." Journal of Personality and Social Psychology 36_, 536-548. Hibscher, J., and Herman, C.P. (1977). Obesity, dieting and the expression of "obese" characteristics. Journal of Comparative and Physiological Psychology 91 374-380. Holland, J., Masling, J., and Copley, D. (1970). Mental illness in lower class, normal, obese and hyper-obese women. Psychosomatic Medicine 32, 351-357. Huck, S.W., Cormier, W.H., and Bounds, W.G. (1974). Reading statistics and research New York: Harper & Row. Johnson, C.L., Stuckey, M.K., Lewis, L.D., and Schwartz, D.M. (1982). Bulimia: A descriptive survey of 316 cases. International Journal of Eating Disorders 2, 3-15. Johnson, M.L., Burke, B.S., and Mayer, J. (1956). Relative importance of inactivity and overeating in the energy balance equation of obese high school girls. American Journal of Clinical Nutrition 4, 37-44. Kaplan, H.I., and Kaplan, H.S. (1957). The psychosomatic concept of obesity. Journal of Nervous and Mental Disease 125 181-189. Kassarjian, H.H., and Nakanishi, M. (1967). A study of selected opinion measurement techniques. Journal of Marketing Research _4, 148-153. Kerlinger, F.N. (1973). Foundations of behavioral research New York: Holt, Reinhart and Winston. Kornhaber, A. (1970). The stuffing syndrome. Psychosomatics 11 580-584. Kubistant, T. (1982). Bulimarexia. Journal of College Student Personnel, 23, 333-339.

PAGE 114

108 Leclair, N.J., and Berkowitz, B. (1983). Counseling concerns for the individual with bulimia. The Personnel and Guidance Journal / 62 352-355. Leon, G.R. (1982). Personality and behavioral correlates of obesity. In B. Wblman (Ed.), Psychological Aspects of Obesity New York: Van Nostrand Reinhold. Leon, G.R., and Chamberlain, K. (1973a). Emotional arousal, eating patterns and body image as differential factors associated with varying success in maintaining a weight loss. Journal of Consulting and Clinical Psychology 40 474-480. Leon, G.R., and Chamberlain, K. (1973b). A comparison of daily eating habits and emotional states of overweight persons successful or unsuccessful in maintaining a weight loss. Journal of Consulting and Clinical Psychology 41 108-115. Leon, G.R., and Roth, R. (1977). Obesity: Pathological causes, correlations and speculations. Psychological Bulletin 84 117-139. Loro, A.D., Levenkron, J.C., and Fisher, E.B. (1979). Critical clinical issues in the behavioral treatment of obesity. Addictive Behaviors 3, 383-391. Loro, A.D., and Orleans, C.S. (1981). Binge eating in obesity: Preliminary findings and guidelines for behavioral analysis and treatment. Addictive Behaviors 6, 155-166. Lowe, M.G. (1982). The role of anticipated deprivation in overeating. Addictive Behaviors 1_, 103-112. Mahoney, M.J. (1975). The obese eating style: Bites, beliefs and behavior modification. Addictive Behaviors 1_, 47-53. McKenna, R.J. (1972). Some effects of anxiety level and food cues on the eating behavior of obese and normal subjects. Journal of Personality and Social Psychology 22 311-319. Meyer, J.E., and Pudel, V.E. (1972). Experimental studies on food intake in obese and normal weight subjects. Journal of Psychosomatic Research 16 305-308. Meyer, J.E., and Pudel, V.E. (1977). Experimental feeding in man: A behavioral approach to obesity. Psychomatic Medicine 39 153-157. Mitchell, J.E., and Pyle, R.L. (1982). The bulimic syndrome in normal weight individuals: A review. International Journal of Eating Disorders, 2, 61-73.

PAGE 115

109 Mitchell, J.E., Pyle, R.L., and Eckert, E.D. (1981). Frequency and duration of binge-eating episodes in patients with bulimia. American Journal of Psychiatry 138 75-89. Monti, P.M., McCrady, B.S., and Barlow, D.H. (1977). Effect of positive reinforcement, informational feedback and contingency contracting on a bulimic anorexic female. Behavior Therapy 8_, 258-263 Moore, M.D., Stunkard, A.J., and Srole, L. (1962). Obesity, social class and mental illness. Journal of the American Medical Association 181 962-966. Nisbett, R.E. (1968). Taste, deprivation and weight determinants of eating behavior. Journal of Personality and Social Psychology 10 107-116. Nisbett, R.E. (1972). Hunger, obesity and the ventromedial hypothalamus. Psycho lgical Review 79 433-453. Nisbett, R.E., and Kanouse, D.E. (1969). Obesity, food deprivation and supermarket shopping behavior. Journal of Personality and Social Psychology 12, 289-294. Nisbett, R.E., and Storms, M.D. (1975). Cognitive, social and physiological determinants of food intake. In H. London and R. E. Nisbett (Eds.), Cognitive modification of emotional behavior Chicago: Aldine. Ondercin, P. (1979). Compulsive eating in college women. Journal of College Student Personnel 20 153-157. Orbach, S. (1978a). Fat is a feminist issue New York: Paddington Press. Orbach, S. (1978b). Social dimensions in compulsive eating in women. Psychotherapy: Theory, Research and Practice 15 180-189. Palmer, R.L. (1979). The dietary chaos syndrome: A useful new term? British Journal of Medical Psychology 52 187-190. Pliner, P., and lappa, G. (1978). Effects of increasing awareness of food consumption in obese and normal weight subjects. Addictive Behaviors 3, 19-24. Pliner, P., Meyer, P., and Blankstein, K. (1974). Responsiveness of affective stimuli by obese and normal individuals. Journal of Abnormal Psychology 83 74-80. Polivy, P. (1976). Perceptions of calories and regulation of intake in restrained and unrestrained subjects. Addictive Behaviors 1_, 237-243.

PAGE 116

110 Polivy, J., and Herman, C.P. (1976a). Clinical depression and weight change: A complex relation. Journal of Abnormal Psychology 85 338-340. Polivy, J., and Herman, C.P. (1976b). The effects of alcohol on eating behavior: Influence of mood and perceived intoxication. Journal of Abnormal Psychology 85 601-606. Polivy, J., Herman, C.P., and Warsh, S. (1978). Internal and external components of emotionality in restrained and unrestrained eaters. Journal of Abnormal Psychology 87 497-504. Polivy, J., Herman, C.P., Younger, J.C., and Erskine, B. (1977). Effects of a model on eating behavior: The indication of a restrained eating style Unpublished manuscript, University of Toronto. Polivy, J., Howard, K.I., and Herman, C.P. (1976). Psychometric analysis of the restraint scale Unpublished manuscript, University of Tornoto. Pyle, R.L., Mitchell, J.E., and Eckert, E.D. (1981). Bulimia: A report of 34 cases. Journal of Clinical Psychiatry 42 60-64. R., K. (1979). That first bite New York: Pomerica Press. Rau, J., and Green, R. (1975). Compulsive eating: A neuropsychologic approach to certain eating disorders. Comprehensive Psychiatry 16, 223-231. Rodin, J. (1973). Effects of distraction on performances of obese and normal subjects. Journal of Comparative and Physiological Psychology 83 68-75. Rodin, J. (1975). Causes and consequences of time perception differences in overweight and normal weight people. Journal of Personality and Social Psychology 31 298-310. Rodin, J. (1977). Research on eating behavior and obesity: Where does it fit in personality and social psychology? Personality and Social Psychology Bulletin _3, 333-355. Rodin, J., and Slochower, J. (1976). Externality in the non-obese: Effects of environmental responsiveness on weight. Journal of Personality and Social Psychology 33 338-344. Rosen, J., and Leitenberg, H. (1982). Bulimia nervosa: Treatment with exposure and response prevention. Behavior Therapy 13 117-124. Ross, L. (1974). Effects of manipulating the salience of food upon consumption by obese and normal eaters. In S. Schacter and J. Rodin (Eds.), Obese humans and rats Washington, D. C: Erlbaum/Wiley.

PAGE 117

Ill Roth, G. (1982). Feeding the hungry heart New York: Bobbs-Merrill Russell, G.F. (1979). Bulimia nervosa: An cminous variant of anorexia nervosa. Psychological Medicine % 429-448. Schacter, S. (1967). Cognitive effects on bodily functions: Studies of obesity and eating. In D. Glass (Ed.): Neurophysiology and emotion New York: Rockefeller University Press. Schacter, S. (1971). Seme extraordinary facts about obese humans and rats. American Psychologist 26 129-144. Schacter, S., Goldman, R., and Gordon, A. (1968). Effects of fear, food deprivation and obesity on eating. Journal of Personality and Social Psychology 10 91-97. Schacter, S., and Gross, L. (1968). Manipulated time and eating. Journal of Personality and Social Psychology 10 98-106. Schacter, S., and Rodin, J. (1974). Obese humans and rats Washington, D. C: Erlbaum/Wiley. Slochower, J. (1976). Emotional labeling and overeating in obese and normal weight individuals. Psychosomatic Medicine 38 131-139. Spencer, J. A., and Fremouw, W.J. (1979). Binge eating as a function of restraint and weight classification. Journal of Abnormal Psychology 8, 262-267. Stalonas, P.M., Johnson, W.G. and Christ, M. (1978). Behavior modification for obesity: The evaluation of exercise, contingency management and program adherence. Journal of Consulting and Clinical Psychology 46 463-469. Stefanik, P.A., Heald, F.P., and Mayer, J. (1959). Calorie intake in relation to output of obese and non-obese adolescent boys. American Journal of Clinical Nutrition 1_, 55-62. Stuart, R., and Davis, B. (1972). Slim chance in a fat world Chicago, II.: Research Press. Stunkard, A.J. (1959). Eating patterns and obesity. Psychiatric Quarterly 33, 284-295. Stunkard, A.J. (1976). The pain of obesity Palo Alto, Ca.: Bull Press. Stunkard, A., and Koch, C. (1964). The interpretation of gastric motility: I Apparent bias in the reports of hunger by obese persons. Archives of General Psychiatry 11 74-82.

PAGE 118

112 Tom, G., and Rucker, M. (1975). Fat, full and happy: Effects of food deprivation, external cues and obesity on preference ratings, consumption and buying intentions. Journal of Personality and Social Psychology 32 761-766. United States Department of Health, Education and Welfare. (1967). Obesity and health: A sourcebook of current information for professional health personnel Arlington, Va.: United States Public Health Service. Wardle, J., and Bienart, H. (1981). Binge eating: A theoretical review. British Journal of Clinical Psychology 20 97-109. Wermuth, B.M., Davis, K.L., Hollister, L.E., and Stunkard, A.J. (1977). Phenytoin treatment of the binge eating syndrome. American Journal of Psychiatry 134 1249-1253. Wilson, G.T. (1976). Obesity, binge eating and behavior therapy: Some clinical observations. Behavior Therapy 1_, 700-701. Wollersheim, J. P. (1970). Effectiveness of group therapy based upon learning principles in the treatment of overweight women. Journal of Abnormal Psychology 76 462-474. Wooley, S. (1972). Physiological versus cognitive factors in short-term food regulation for the obese and non-obese. Psychosomatic Medicine, 34, 62-68.

PAGE 119

BIOGRAPHICAL SKETCH Edward Thanas Spauster was born on October 27, 1953, in New York City. He is the son of Anne Marie Spauster and the late Robert J. Spauster. Ed graduated from Archbishop Molloy High School in Jamaica, New York, in 1971 and then entered the State University of New York at Stony Brook. Prior to receiving his B. A. in psychology in 1976, Ed served in a variety of student leadership positions including Student Body President. He enrolled at the State University of New York at Albany in 1977 and received an M.S.W. in 1980. Awarded a Graduate Council Fellowship from the University of Florida, he began doctoral studies in counseling psychology. Mr. Spauster is currently living in Quincy, Massachusetts, with his wife, Elizabeth Coburn. He has recently completed his pre-doctoral internship at the Veterans Administration Medical Center in Brockton, Massachusetts. Following graudation in August, 1984, he plans to work in a medical facility as a psychologist. 113

PAGE 120

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. 1 J. Wittmer', Chairper rofessor of Counselor Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. James Archer Assistant Professor of Counselor Education

PAGE 121

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. 'M^ r of Psysholc llliam Frc Associate iteofessor of Psy&hology This dissertation was submitted to the Graduate Faculty of the Department of Counselor Education in the College of Education and to the Graduate School, and was accepted for partial fulfillment of the requirements of the degree of Doctor of Philosophy. August, 1984 Dean for Graduate Studies and Research

PAGE 122

UNIVERSITY OF FLORIDA 3 1262 rfllllfl'l