USF Libraries
USF Digital Collections

Predicting obesity from four eating behaviors

MISSING IMAGE

Material Information

Title:
Predicting obesity from four eating behaviors
Physical Description:
Book
Language:
English
Creator:
Yanover, Tovah
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Hunger
Eating beyond satiety
Snacking
Night eating
Eating expectancies
Cue reactivity
Dissertations, Academic -- Psychology -- Masters -- USF
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: Obesity is a growing problem in the United States. Research into the causes, treatment, and prevention of obesity is vital. One past study examined four eating behaviors in relation to obesity: eating beyond satiety, snacking, night eating, and feeling hungry within three hours of eating. Only eating beyond satiety was associated with obesity. The present study examined these same eating behaviors while correcting some of the flaws of the previous study. Using a cross-sectional design, university undergraduates reported on the frequency of the above-named eating behaviors. Current weight and height were collected. Multiple regression analyses determined that eating beyond satiety and hunger predicted body mass index (BMI). Race/ethnicity moderated the relationship between hunger and BMI. These findings have important implications for obesity treatment as well as suggesting important avenues for future research.
Thesis:
Thesis (M.A.)--University of South Florida, 2005.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Tovah Yanover.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 63 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001913659
oclc - 174141649
usfldc doi - E14-SFE0001389
usfldc handle - e14.1389
System ID:
SFS0025709:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001913659
003 fts
005 20071015093314.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 071015s2005 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001389
040
FHM
c FHM
035
(OCoLC)174141649
049
FHMM
090
BF121 (ONLINE)
1 100
Yanover, Tovah.
0 245
Predicting obesity from four eating behaviors
h [electronic resource] /
by Tovah Yanover.
260
[Tampa, Fla] :
b University of South Florida,
2005.
3 520
ABSTRACT: Obesity is a growing problem in the United States. Research into the causes, treatment, and prevention of obesity is vital. One past study examined four eating behaviors in relation to obesity: eating beyond satiety, snacking, night eating, and feeling hungry within three hours of eating. Only eating beyond satiety was associated with obesity. The present study examined these same eating behaviors while correcting some of the flaws of the previous study. Using a cross-sectional design, university undergraduates reported on the frequency of the above-named eating behaviors. Current weight and height were collected. Multiple regression analyses determined that eating beyond satiety and hunger predicted body mass index (BMI). Race/ethnicity moderated the relationship between hunger and BMI. These findings have important implications for obesity treatment as well as suggesting important avenues for future research.
502
Thesis (M.A.)--University of South Florida, 2005.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 63 pages.
Includes vita.
590
Adviser: William P. Sacco, Ph.D.
653
Hunger.
Eating beyond satiety.
Snacking.
Night eating.
Eating expectancies.
Cue reactivity.
690
Dissertations, Academic
z USF
x Psychology
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1389



PAGE 1

Predicting Obesity from Four Eating Behaviors by Tovah Yanover A thesis submitted in partial fulfillment Of the requirements for the degree of Master of Arts Department of Psychology College of Arts and Sciences University of South Florida Major Professor: William P. Sacco, Ph.D. Thomas H. Brandon, Ph.D. Kristen Salomon, Ph.D. J. Kevin Thompson, Ph.D. Date of Approval: November 8, 2005 Keywords: hunger, eating beyond satiety, snacking, night eating, eating expectancies, cue reactivity Copyright 2005, Tovah Yanover

PAGE 2

Dedication I dedicate this thesis to my family: to my mother and father who have supported me tirelessly and always encouraged me to value learning and education, and to my husband who is a constant source of encouragemen t and comfort. I would also like to thank my lab mates Christine Vaughan, M.Sc. and Kristi Wells, MPH, who have helped me adjust to life as a graduate student and who have shared the wealth of their experience with me.

PAGE 3

Acknowledgements I would like to th ank Dr. William Sacco for his time, effort, and patience in supervising this project as well as for the opportunity to explore an area of interest to me. I would also like to thank my committee, Dr. Thomas Brandon, Dr. Kristen Salomon, and Dr. J. Kevin Tho mpson for giving their time and for offering their valuable insights. I would like to thank my fellow graduate students, Kristi White, Cathy Bykowski, Guy Cafri, M.A., Slyvia Herbozo, M.A., Megan Roehrig, M.A., Susan Himes, M.A., and Michael Boroughs Stewa rt for taking the time to help me prepare for my defense and for offering constructive feedback. Thank you to Dr. Michael Brannick for his patient assistance with statistical analyses. Finally, thank you to my husband who worked so hard to create the web p ages used for the data collection for this study.

PAGE 4

i Table of Contents List of Tables ................................ ................................ ................................ .................... iii Lis t of Figures ................................ ................................ ................................ ................... iv Abstract ................................ ................................ ................................ .............................. v Introduction ................................ ................................ ................................ ....................... 1 S NACKING ................................ ................................ ................................ ........................ 3 E ATING B EYOND S AT IETY ................................ ................................ ............................... 3 N IGHT E ATIN G ................................ ................................ ................................ ................. 4 H UNGER ................................ ................................ ................................ ........................... 5 E XPECTANCIES ................................ ................................ ................................ ................. 5 L EARNING T HEORY ................................ ................................ ................................ .......... 6 P RESENT S TUDY ................................ ................................ ................................ ............... 8 H YPOTHESES ................................ ................................ ................................ .................... 9 Method ................................ ................................ ................................ ............................. 10 P ARTICIPANTS ................................ ................................ ................................ ................ 10 M EASURES ................................ ................................ ................................ ..................... 11 Body Mass Index ................................ ................................ ................................ ...... 11 Paffenbarger Physical Activity Questionnaire ................................ .......................... 11 Eating Disorder Diag nostic Scale ................................ ................................ ............. 12 Eating Beyond Satiety Questionnaire ................................ ................................ ....... 13 Snacking Questionnaire ................................ ................................ ............................ 14 Hunger Questionnaire ................................ ................................ ............................... 14 Night Eating Scale ................................ ................................ ................................ .... 15 Brewer et al. (2003) Measures ................................ ................................ .................. 15

PAGE 5

ii Eating Expectancies Inventory ................................ ................................ .................. 15 Self Reported Cue Reactivity Questionnaire ................................ ............................ 16 P ROCEDURE ................................ ................................ ................................ ................... 17 Results ................................ ................................ ................................ .............................. 18 H YPOTHESIS 1 ................................ ................................ ................................ ................ 18 H YPOTHESIS 2 ................................ ................................ ................................ ................ 22 H YPOTHESIS 3 ................................ ................................ ................................ ................ 25 H YPOTHESIS 4 ................................ ................................ ................................ ................ 26 H YPOTHESIS 5 ................................ ................................ ................................ ................ 27 Discussion ................................ ................................ ................................ ......................... 29 L IMITATIONS ................................ ................................ ................................ .................. 35 F UTURE R ESEARCH ................................ ................................ ................................ ........ 36 References ................................ ................................ ................................ ........................ 38 Appendices ................................ ................................ ................................ ....................... 44 A PPEN DIX A: P AFFENBARGER P HYSICAL A CTIVITY Q UESTIONNAIRE ............................ 45 A PPENDIX B: E ATING D ISORDER D IAGNOSTIC S CALE ................................ ................... 48 A PPENDIX C: E ATING B EYOND S ATIETY Q UESTIONNAIRE ................................ ............. 50 A PPENDIX D: S NACKING Q UESTIONNAIRE ................................ ................................ ..... 52 A PPENDIX E: H UNGER Q UESTIONNAIRE ................................ ................................ ......... 53 A PPENDIX F: N IGHT E ATING S CALE ................................ ................................ ............... 54 A PPENDIX G: B REWER ET AL (2003) I TEMS ................................ ................................ .. 55 A PPENDIX H: E ATING E XPECTANCIES I NVENTORY ................................ ........................ 56 A PPENDIX I: S ELF R EPORTED C UE R EACTIVITY Q UESTIONNAIRE ................................ 59 A PPENDIX J: D EMOGRAPHIC I NFORMATION ................................ ................................ ... 62 About the Author ................................ ................................ ................................ End P age

PAGE 6

iii List of Tables Table 1. Demographic characteristics ................................ ................................ ............... 11 Table 2. Means and standard deviations for scale scores ................................ ................. 18 Table 3. Correlations amon g the eating behaviors, BMI, physical activity, eating expectancies and self reported cue reactivity. ................................ .......................... 19 Table 4. Correlations among the eating behaviors (Brewer et al. (2003) measures), BMI, physical activity, eating expectan cies and self reported cue reactivity .................... 19 Table 5. Correlations among the eating behaviors: Present study measures vs. Brewer et al. (2003) measures ................................ ................................ ................................ ... 20 Table 6. Standardized regression coefficients and p values for the regression predicting BMI from all four eating behaviors and from the Brewer et al. (2003) items .......... 21 Table 7. Odds ratios for the eating behaviors predicting normal weight versus overweight ................................ ................................ ................................ ................................ ... 22 Table 8. Standardized regression coefficients for simple regression of BMI on hunger for each race/ethnicity group ................................ ................................ .......................... 24 Table 9. Race/ethnicity differences on the eating behaviors ................................ ............ 26

PAGE 7

iv List of Figures Figure 1. Regression slopes for the simple regression of BMI on hunger for each race/ethnicity group ................................ ................................ ................................ ... 24

PAGE 8

v Predicting Obesity from Four Eati ng Behaviors Tovah Yanover ABSTRACT Obesity is a growing problem in the United States. Research into the causes, treatment, and prevention of obesity is vital. One past study examined four eating behaviors in relation to obesity: eating beyond satiety, sna cking, night eating, and feeling hungry within three hours of eating. Only eating beyond satiety was associated with obesity. The present study examined these same eating behaviors while correcting some of the flaws of the previous study. Using a cross sec tional design, university undergraduates reported on the frequency of the above named eating behaviors. Current weight and height were collected. Multiple regression analyses determined that eating beyond satiety and hunger predicted body mass index (BMI). Race/ethnicity moderated the relationship between hunger and BMI. These findings have important implications for obesity treatment as well as suggesting important avenues for future research.

PAGE 9

1 Introduction Obesity is a growing problem. Recent statistics f rom the Centers for Disease Control and Preventions National Center of Health Statistics (2004a) indicate that in 2000, 64% of Americans over the age of 20 were overweight and 30% were obese. In 1962, these figures were 44.8% and 13.3% respectively. There has been a huge increase in the prevalence of obesity and overweight in recent decades, and these figures continue to rise. Obese individuals are also at higher risk for a multitude of health problems including diabetes mellitus, cardiovascular disease, a nd all cause mortality (National Heart, Lung and Blood Institute, 1998). Research into the causes, treatment, and prevention of obesity is vital in order to reduce the toll this condition takes on peoples lives. Obesity is the result of a positive energy balance wherein individuals take in more calories than they expend (Yanovski & Yanovski, 1999). A great deal of funding for recent research focusing on the energy out or physical activity side of the equation has been made available, with the goal of get ting people to increase their activity. But the energy in or food intake side of the equation deserves just as much attention (Lavizzo Mourey, 2004). Many past studies have examined diet composition to determine whether certain macronutrients contribute to the development of obesity (e.g. Nicklas, Baranowski, Cullen, & Berenson, 2001; Ortega, Redondo, Zamora, Lopez Sobaler, & Andres, 1995). However, results from these studies have failed to identify macronutrient intake patterns that lead to obesity (Togo, Osler, Sorenson, & Heitmann, 2001). Other approaches are required to determine the mechanisms by which individuals take in so much more food than they need, and the determinants thereof. In a recent article, Brewer, Kolotkin, and Baird (2003) investigated whether specific eating behaviors were associated with obesity in premenopausal African American and Caucasian women. They examined four behaviors: eating before bedtime,

PAGE 10

2 feeling hungry within three hours of eating, eating between meals, and eating past t he point of feeling full or the point of satiety. In their study, only eating beyond satiety was associated with obesity. This finding is important. If a specific eating behavior can be shown to be associated with obesity, this is an important first step i n encouraging investigators to look into the causal role of eating behaviors. If eating beyond satiety does indeed play a causal role in the development of obesity, there are major implications for treatment and prevention programs. Teaching people to reco gnize internal signals of satiety and then to stop eating when they feel those signals could help prevent obesity, and help those who are obese to lose weight. There were, however, some problems with the Brewer et al. (2003) study that lead to the need f or new studies to reexamine these variables. Only one question assessed the frequency of each eating behavior, leading to questionable reliability of the measures. In addition, their measure of overweight was artificially dichotomized, forcing them to use nonparametric analyses and leading to a loss of power in the analyses. Furthermore, Brewer et al. (2003) did not control for physical activity. If individuals take in extra calories via one of the eating behaviors under study but then compensate by exercis ing, they could maintain a constant weight. Therefore, physical activity might mask relationships that exist between obesity and eating behaviors. Physical activity was controlled in the present study, to remove its effects from the relationship between ea ting behaviors and obesity. Reexamination of the variables set out in the study by Brewer et al. (2003), while removing these flaws, would strengthen the level of confidence in its findings. If the findings do, in fact, hold up, there are implications for the treatment of obesity. Correlation between eating beyond satiety and obesity does not imply a causal role for the eating behavior. However, establishing a correlation is a prerequisite to establishing causality. If future research can establish a causal role for eating beyond satiety or other eating behaviors, then weight management programs can incorporate strategies that teach individuals to recognize internal cues that signal satiety, and then to stop eating when they sense those cues. Research could also examine environmental or situational factors that are predictive of maladaptive eating behaviors. Weight

PAGE 11

3 management programs could also teach individuals how to avoid these situations or deal with them more adaptively. In the present study, the four e ating behaviors investigated by Brewer et al. (2003) were once again examined to determine their relative contributions to obesity. These behaviors include snacking, eating beyond satiety, night eating, and hunger. Although hunger is not an eating behavior it is included in order to re examine all four of the variables proposed by Brewer et al. (2003). In the current study, explanatory mechanisms were also explored. Both expectancy theory and learning theory could potentially explain why some individuals e ngage in maladaptive eating behaviors while others do not. Snacking Concern has been raised about snacking because snack foods are commonly thought to be high fat foods like chocolate or potato chips. Estimates of the contribution of snacking to the overal l caloric intake of Americans range from 16% to 29% in various age groups (Summerbell, Moody, Shanks, Stock, & Geissler, 1995). Basdevant, Craplet, and Guy Grand (1993) observed that the total daily energy intake of snackers was higher than that of non sna ckers. Findings relating snacking to obesity have been mixed (Brewer et al., 2003). For example, Berteus Forslund, Lindroos, Sjostrom, and Lissner (2002) found that obese women consumed more meals per day than a group of reference women. On the other hand, Fabry and Tepperman (1970) concluded that eating frequent meals is not necessarily pathological and may have positive benefits on weight. It is hypothesized that snacking will predict obesity. Eating Beyond Satiety Many now agree that environmental influe nces contribute to the problem of obesity (Hill, 1998). Large portions of tasty, high energy density foods are readily available (DeAngelis, 2004). And people tend to eat until their plate is empty. In a study by Tuomisto, Tuomisto, Hetherington, and Lappa lainen (1998), 39% of obese participants reported that they stopped eating because they had had enough, but 10% reported that they stopped eating because no food was left. The phenomenon of eating

PAGE 12

4 beyond satiety has been linked to disinhibition (Brewer et al., 2003), described as the tendency to eat because of the availability of palatable foods, because others are eating, or because of emotional distress (Lowe & Maycock, 1988). Disinhibition has been associated with body mass index (BMI), an indicator of o verweight, in several studies (e.g. Boschi, Iorio, Margiotta, DOrsi, & Falconi, 2001; Hays et al., 2002) Eating past the feeling of satiety was the only eating behavior associated with obesity in the study by Brewer et al. (2003). It is hypothesized that this relationship will once again be found in the present study. Furthermore, Brewer et al. (2003) found that eating beyond satiety was a stronger predictor of obesity in African American women than in Caucasian women. It is expected that, in the present s tudy, race/ethnicity will once again moderate the relationship between eating behaviors and obesity. Eating beyond satiety falls on a continuum, which at very high levels is likely represented by Binge Eating Disorder (BED). BED is defined as regular eatin g binges wherein the individual eats a large amount of food and perceives a loss of control over eating. These binges are accompanied by distress but not compensatory behaviors such as vomiting, fasting, or excessive exercise (Stice et al., 2000). Binges a re often accompanied by feelings of being overfull or uncomfortably full. They are characterized by the consumption of large amounts of food. Eating beyond satiety is characterized by these same symptoms. BED is commonly found in overweight and obese indiv iduals. Thus, it is possible that the association between eating beyond satiety and obesity is accounted for by BED. The present study therefore examined whether eating beyond satiety adds to the prediction of obesity, over and above the effect of BED. Nig ht Eating Night eating may be informally referred to as bedtime snacking. Popular lore would have us believe that eating before bed will have a negative effect on ones weight. Empirical data does not support this position. Kant, Balfard Barbash, and Schat zkin (1995) conducted a diary study of eating patterns and had women report what they ate during several periods of the day. Despite the fact that one third of the women consumed

PAGE 13

5 more than 50% of their calories after 8 p.m., they found that night eating ha d no relationship to BMI. Brewer et al. (2003) also failed to find that night eating was a significant predictor of obesity. However, reliability of the measures in these studies could not be established because the number of items was too small. The prese nt study used a scale designed to assess the construct of night eating in a more reliable manner. It is hypothesized that night eating will predict obesity in the present study 1 Hunger Research has been inconsistent in determining whether or not feeling h ungry within three hours of eating is associated with obesity. Boschi et al. (2001) surveyed normal weight and overweight women at the same Italian outpatient weight clinic and found that overweight women had higher hunger scores on Stunkard and Messicks (1985) Three Factor Eating Questionnaire. On the other hand, Hays et al. (2002) found that hunger did not predict weight gain or current BMI. Feeling hungry is not, however, an eating behavior. Rather, it is a sensation that may lead to eating. Feeling hun gry before a meal is adaptive and normative. However, feeling constantly hungry and unable to refrain from eating because of that hunger could be maladaptive. Hunger is expected to be correlated with snacking and night eating. Hunger is expected to predic t obesity in the present study. Expectancies Expectancies have been defined as long term memory structures that have an impact on cognitive processes governing behavior, both current and future (Jones, Corbin, & Fromme, 2001). Based on their learning hist ory, individuals come to expect a given behavior will lead to a given outcome. If that outcome is desirable, the probability of the behavior is higher. If the outcome is negative, the probability of the behavior is 1 Stunkard (1955 as cited in Napolitano, Head, Babyak, & Blumenthal, 2001) postulated a disorder known as night eating syndrome (NES) characterized by a lack of appetite in the morning, overeating in the evening, emotional distress, and insomnia. It could be that night eating also falls along a continuum, the extreme represented by NES. The present study did not, however, assess for the presence of NES.

PAGE 14

6 lower. When an individual is given the op portunity to engage in some behavior, expectancies may act as cognitive mediators in the decision making process. Expectancy theory has been extensively investigated in the substance abuse literature, especially in the domain of alcohol. Individuals with m ore positive alcohol expectancies engage in heavier drinking than individuals with less positive alcohol expectancies. Additionally, alcohol expectancies have been found to play a causal role in alcohol consumption in that manipulation of expectancies can reduce drinking (Darkes & Goldman, 1993). Recently, investigators have begun to examine food related expectancies and their relationships to eating disorders such as anorexia and bulimia (Hohlstein, Smith, & Atlas, 1998). Hohlstein et al. (1998) claim that expectancy theory has been successfully applied in numerous other areas of psychology and that expectancy operations appear implicit in a number of current eating disorder constructs (p. 49). They found that bulimic individuals held expectancies that fo od would reduce negative affect and that anorexic individuals tended not to expect positive reinforcement from food. Expectancy theory has not, however, been applied in the context of obesity. This theory has much to offer in the investigation of the relat ionship between eating behaviors and obesity. What an individual comes to expect from food could then predict his or her intake. If an individual expects food to provide relief from negative affect, he or she would be more likely to engage in maladaptive e ating behaviors in stressful situations. If individuals have more positive food or eating related expectancies, they would be expected to consume larger amounts of food and to do so more frequently, thus increasing their chances of being overweight or obe se. The present study investigated whether those individuals who engage in the eating behaviors of interest are those who hold more positive eating or food related expectancies. It is hypothesized that more positive expectancies about eating will be relate d to BMI and eating behaviors. Furthermore, the present study tested the hypothesis that eating behaviors mediate the effects of expectancies on BMI. Learning Theory Jansen (1998) compares food intake to the use of drugs in that both initiate physiological responses. Exposure to the sight, smell or taste of food sends the body into

PAGE 15

7 preparations for consumption (Mattes, 1997). The bodys cephalic phase responses, those processes that prepare it for food intake, include salivation, the release of gastric juic es and insulin as well as preparations by the pancreas, kidneys, and heart (Mattes, 1997). The whole body gears up in preparation for food intake. The fact that the body can prepare for ingestion upon the mere sight of food suggests a role for conditioning processes. The learning or cue reactivity model proposes that the cephalic phase responses are experienced as craving for food. They increase food intake and increase the likelihood of maladaptive eating patterns (Nederkoorn, Smulders, & Jansen, 2000). Le arning theory applies to obesity as well and has been researched in this context. According to Jansen et al. (2003), learning theory predicts that overeating is due to learned associations between appetitive food cues like smell and taste on the one hand, and food intake on the other. Jansen et al. (2003) found that normal weight children regulate their eating by eating less following a preload of tasty foods, whereas overweight children do not. Epstein, Paluch, and Coleman (1996) found that obese women did not habituate their salivation in response to repeated presentation of the same food, whereas normal weight women did habituate their salivation response. The women had food placed on their tongues, but were not permitted to consume it. The obese womens response to the food remained strong, while that of normal weight women weakened over time. One could argue that the obese women had stronger and more enduring responses to food stimuli in that the salivation response, which prepares them for food intake, does not diminish over time in response to the same food stimulus. Because the responses last longer, it may be more difficult for an obese individual to resist the urge to eat. Their bodies are preparing for food intake and signaling readiness, and they m ay give in when the signals do not dissipate, resulting in more frequent and long lasting eating episodes. Further support for this proposal comes from a study comparing cue exposure treatment to self control treatment of binge eating. Cue exposure treatme nt was more effective in reducing the frequency of binges (Jansen, Broekmate, & Heymans, 1992). In the cue exposure treatment, participants were exposed to cues that signaled a binge. These cues included binge clothing, location, and foods. Participants we re encouraged to touch, smell, and even taste small amounts of the binge foods, but they were not permitted to

PAGE 16

8 engage in a binge. The goal of this treatment was to break the bond between cues that signal a binge and the binge response. The self control tr eatment taught participants to identify high risk situations for a binge and to learn alternative ways to deal with these difficult situations. Role playing and self efficacy exercises were used in the cognitive restructuring portion of the treatment where participants learned and practiced the alternative behaviors. Jansen et al. (1992) found that individuals in the self control treatment group had a 33% relapse rate while those in the cue exposure treatment group were able to abstain totally from bingeing over the one year follow up period. These findings are consistent with the idea that responses to food stimuli can be powerful motivators for food consumption. It is hypothesized, therefore, that stronger and more enduring responses to food cues, in the f orm of self reported cue reactivity, will be positively related to BMI and the eating behaviors of interest in the present study: eating beyond satiety, snacking, night eating, and hunger. Further, this study tested the hypothesis that these eating behavio rs mediate the relationship between self reported cue reactivity to food and BMI. Present Study Previous research (Brewer et al., 2003) examined only African American and Caucasian women. The present study also included Hispanic women. Statistics show that rates of obesity are high in all three of these groups (National Center for Health Statistics, 2004b). Between 1999 and 2000, approximately 62% of Caucasian females aged 20 years and above were overweight and 33% were obese. The figures for African Americ an women are 77% overweight and 49% obese. For Hispanic women of Mexican origin, 69% were overweight and 38% were obese. These figures indicate that obesity and overweight are problematic in all of these groups and that all three of these groups merit incl usion in a study of this nature. Previous research (Brewer et al., 2003) has also focused on older populations; specifically premenopausal women aged 35 to 49 in their sample. With the rapid increase in obesity at all ages, it is important to understand ob esity in younger individuals and to target interventions at those individuals. In the present study, the population of interest

PAGE 17

9 was undergraduate students, a population younger than that studied by Brewer et al. (2003). The prevalence of overweight among i ndividuals between the ages of 12 and 19 was 6.1% between 1971 and 1974 and it rose to 15.5% between 1999 and 2000 (National Center for Health Statistics, 2004c). Overweight and obesity has more than doubled in this age group in the last 30 years making th is group an important one to study. Targeting undergraduates captures individuals who fall in this age group as well as the older age group, individuals 20 years and older among whom 30% are obese and 64% are overweight, (National Center for Health Statist ics, 2004a). The present study aimed to pull all of the above threads together to begin to better understand which eating behaviors predict obesity and whether expectancy theory or learning theory can help explain why certain individuals develop maladapti ve eating patterns while others do not. The present study hopes to initiate and stimulate research in an area that, if fruitful, could help reduce the disease burden and lives lost due to obesity. Hypotheses 1. Of the four eating behaviors assessed in the pr esent study, eating beyond satiety will be the best predictor of BMI when the effects of the others are controlled. 2. Race/ethnicity will moderate the relationship between the eating behaviors and BMI. 3. Positive food expectancies will correlate positively wit h BMI and with the eating behaviors. Further, the eating behaviors will mediate the relationship between expectancies and BMI. 4. Strong, enduring self reported cue reactivity to food stimuli will be associated with BMI and with the eating behaviors. Further, the eating behaviors will mediate the relationship between self reported cue reactivity and BMI. 5. Eating beyond satiety will account for additional variance in BMI over and above the presence of a diagnosis of Binge Eating Disorder.

PAGE 18

10 Method Participants Pa rticipants were recruited through the University of South Floridas Experimentrack system, which allows researchers to recruit students and reward their research participation with extra course credit. Inclusion criteria required participants to be female, able to read English, over the age of 18, and to belong to one of three racial/ethnic groups (African American Non Hispanic, Caucasian Non Hispanic or Hispanic). For the purposes of this study, participants must self identify as belonging to only one raci al/ethnic group. Six hundred and twenty four individuals participated in the present study. Of these, 191 were eliminated either because they were male or because they did not belong to one of the racial/ethnic groups of interest. The remaining 45 individu als were eliminated because they answered one or more of the five attention items incorrectly. These were items scattered throughout the survey to detect careless responding. The final sample consisted of 388 female undergraduates at the University of Sout h Florida. The mean age of the sample was 21.8 years ( SD = 4.66) with a range of 18 to 53 years. The majority of the sample was in their junior (32.0%) or senior (31.2%) year. The sample was predominantly Caucasian Non Hispanic (74.0%) with 14.4% of the pa rticipants self identifying as Hispanic and 11.6% of the sample self identifying as African American Non Hispanic. The largest proportion of the sample reported being single (29.6%) followed closely by being in a serious dating relationship (26.5%). Less t han one quarter of the sample (23.2%) endorsed current dieting efforts. Of these, 36.4% reported being on other diet plans, with the second most popular choice being counting calories (23.2%). The mean BMI of the sample was 24.3 ( SD = 5.29), which falls at the high end of the normal weight range (see Table 1 for racial/ethnic breakdown). Mean weight was 146.7

PAGE 19

11 pounds ( SD = 34.27) and mean height was 65.1 inches ( SD = 2.66). Five individuals (1.3%) had subclincial BED according to the Eating Disorder Diagno stic Scale and fourteen individuals (3.6%) met criteria for BED. Table 1 Demographic characteristics Variable Age BMI Group M SD M SD Total Sample 21.8 4.66 24.3 5.29 Caucasian Non Hispanic 21.8 4.83 23.9 5.06 African Americ an Non Hispanic 22.5 4.37 27.3 a 5.99 Hispanic 21.6 3.96 23.5 5.11 a Mean is higher than the other two, p < .05 Measures Measures can be found in Appendices A J. Five items will be interspersed throughout the measures to detect careless responding. Body Ma ss Index Body Mass Index (BMI) is a common index of overweight calculated by [(weight in pounds) / (height in inches) 2 ] x 703. Current BMI was calculated from height and weight information provided by participants. Paffenbarger Physical Activity Question naire Physical activity was assessed using the Paffenbarger Physical Activity Questionnaire (Paffenbarger, Wing, & Hyde, 1978). The questionnaire consists of eight items asking individuals their typical level of activity. Key items ask participants to repo rt the number of flights of stairs they climb up each day and the number of blocks they walk each day. Participants also report the various types of physical activity in which they participate, the number of times a year they do so, and the duration of eac h

PAGE 20

12 episode. The types of activity are then multiplied by their estimated intensity in METS, a unit that represents the metabolic equivalent of an activity in multiples of the resting rate of oxygen consumption (Courneya & Hellsten, p. 627). The MET values were obtained primarily from Lee, Paffenbarger, and Hsieh (1992). The scale score is calculated by multiplying the activity by the MET score and then by the duration per episode in minutes and the number of episodes per week. The number of kilocalories ex pended on stairs weekly is calculated by multiplying the number of flights of stairs climbed daily by 7 and then by 8. The number of kilocalories expended walking is calculated by multiplying the number of blocks walked daily by 7 and then 4. The sports, s tairs, and blocks scores are then summed. The Paffenbarger thus yields a weekly kilocalorie expenditure for each participant. The Paffenbarger has one month test retest reliability ranging from .63 for the number of blocks walked daily to .78 for the numbe r of flights of stairs climbed daily. It has been shown to have convergent validity in that it correlates .31 with a four week history of physical activity and .52 with a physical fitness workload test. The Paffenbarger scores are related to risk of first heart attack and hypertension, and to HDL cholesterol (Paffenbarger et al., 1978). Alpha was not calculated for the current sample, as one would not expect the items to hold together well. Eating Disorder Diagnostic Scale Binge eating was assessed using th e Eating Disorder Diagnostic Scale (EDDS: Stice et al., 2000). The EDDS is a brief, self report measure of eating pathology that provides separate diagnostic information on Anorexia, Bulimia, and Binge Eating Disorder. Test retest reliability of the EDDS i s .87 and Cronbachs alpha is .91. The EDDS discriminates among the different eating disorder diagnoses (Stice et al., 2000). The EDDS has also been shown to agree with the Eating Disorder Examination, a structured interview of eating pathology (Kappa = 78; Stice et al., 2004). The final four items of the EDDS collect height, weight, and menstrual cycle information. These items were omitted. Height and weight were collected later and menstrual status was not of interest in the present study. The EDDS yiel ds a diagnosis of Binge Eating Disorder

PAGE 21

13 (BED) if an individual gives a response of yes to items 5, 6, 14, and at least three of 9, 10, 11, 12 and 13. An individual must also give a response greater than two on item 7 and zero on items 15, 16, 17, and 18. The EDDS provides an eating disorder diagnosis of clinical and subclinical bulimia nervosa and binge eating disorder. In the present sample the full EDDS showed good internal consistency (Alpha = .85). Eating Beyond Satiety Questionnaire Eating beyond sat iety was assessed using 10 items. Four items were taken from the Disinhibition scale of the Three Factor Eating Questionnaire (TFEQ; Stunkard & Messick, 1985), either directly or in modified form. The TFEQ is a commonly used measure of eating behaviors tha t has been employed extensively in a wide variety of populations. It has been shown to discriminate between dieters and free eaters (Stunkard & Messick, 1985). The disinhibition scale was not used in its entirety in the present study because it is not a pu re measure of eating beyond satiety. Some items also assess eating in response to mood cues such as anxiety or feeling blue. Therefore, items that assessed the construct of eating beyond satiety were pulled from the disinhibition scale for use in the pre sent study. These are items 2, 3, 7, and 11. The response options were also changed from true or false to a 5 point Likert type scale, which is preferable because it increases the potential variability of the responses. Words that refer to frequency of beh avior such as usually, sometimes, and always were removed from the items because these ideas are now included in the response options. Leaving them in the items themselves leads to redundancy. Additionally, item 9 was modified to read Even though I am full from the original I am always hungry so in order to put the focus on satiety rather than hunger. The words not difficult were removed from item 7. The remaining items were experimenter generated to measure the construct of eating beyond satiet y and were created to be face valid and construct valid questions. An expert panel rated the items as a good measure of eating beyond satiety. Items 7 and 10 were reverse coded. Responses were summed to create a total score for eating beyond satiety (possi ble total = 50). The measure showed good internal consistency in the present sample (Alpha = .85).

PAGE 22

14 Snacking Questionnaire A thorough literature search failed to generate an existing, validated measure of snacking. Snacking was, therefore, assessed with si x experimenter generated items. These items were designed to assess frequency of snacking during the day. An expert panel rated the items as a good measure of snacking. Item 1 was reverse coded and the responses were summed to create a total score for snac king (possible total = 30). The full scale showed reasonable internal consistency in the present sample (Alpha = .63), however an item analysis revealed that the first five items produced a scale with a more acceptable alpha. Therefore, the final scale con sists of the first five items, which showed good internal consistency (Alpha = .77), with a possible scale total of 25. Hunger Questionnaire Hunger was assessed using seven items. Two of these items were modified from the Hunger scale of the TFEQ (Stunkard & Messick, 1985). The hunger scale of the TFEQ was not used in its entirety because several items confound the sensation of hunger with the behavior of eating. Unconfounded items were pulled from the scale for use in the present study. These are items 1 and 5. The response options for these two items were changed from true or false to a 5 point Likert type scale. The word always was removed from both of these items because it became redundant with the new response format. The remaining items were exper imenter generated to measure the construct of hunger. Items 2, 3, and 4 were designed to extend the question asked in the study by Brewer et al. (2003), which asked how often after eating an individual is likely to feel hungry. The likelihood of feeling hu ngry may vary depending on the meal eaten previously, so the present study has three items; one each for breakfast, lunch, and dinner. These three items were reverse coded so that higher scores reflected greater hunger sensations. Item 6 was designed to be a face valid measure of hunger. An expert panel rated the items as a good measure of hunger. Responses were summed to create a total

PAGE 23

15 hunger score (possible total = 35). Internal consistency for the scale was good in the present sample (Alpha = .74). Nigh t Eating Scale Night eating was assessed using seven items, several of which were from the Night Eating Questionnaire (NEQ; Marshall et al., 2004). The NEQ is a 14 item screening instrument designed to assess Night Eating Syndrome. The construct of interes t in the current study is the frequency with which individuals engage in nighttime eating. Therefore, items addressing this construct were pulled from the scale. The remaining items were experimenter generated. An expert panel rated the items as a good mea sure of night eating. Responses were summed to yield a total night eating score (possible total = 35). Internal consistency for the scale was good in the current sample (Alpha = .78). Brewer et al. (2003) Measures To explore the possibility that the findi ngs from Brewer et al. (2003), were due to poor measurement, the original four questions used in their study were included as a separate measure for the sake of comparison. No reliability information is available for these items. The four items assessed ea ting before bedtime (How often do you eat during the two hours before your bedtime?), eating between meals (How often do you eat between meals?) feeling hungry within three hours of eating (How often do you feel really hungry even though you have eate n adequately in the last three hours?) and eating beyond satiation (How often do you keep eating at meals even though you are not hungry anymore?). These items were all rated on a 5 point Likert type scale from rarely or never to nearly everyday. Ea ting Expectancies Inventory Expectancies were assessed with the Eating Expectancies Inventory (EEI; Hohlstein et al., 1998). The EEI is a five subscale measure of learned expectancies related to eating. The scale was originally designed for use with eatin g disordered populations, but has applications to the problem of obesity. The five subscales are: Eating

PAGE 24

16 helps manage negative affect (18 items), Eating is pleasurable and useful as a reward (6 items), Eating leads to feeling out of control (4 items), Eati ng enhances cognitive competence (2 items), and Eating alleviates boredom (4 items). Coefficient alphas for the five factors range from .78 to .94. Various subscales of the EEI have been shown to correlate highly with several other eating measures includin g the Disinhibition and Restraint scales of the TFEQ. The EEI has also been used to show that expectancies differ across individuals with anorexia, individuals with bulimia, and control individuals. The factor structure of the EEI has been shown to be cons istent in samples of African American and Caucasian undergraduate women (Atlas, Smith, Hohlstein, McCarthy, & Kroll, 2002). Negatively worded items were reverse coded so that a higher total score reflected more positive expectancies toward food. Responses were summed to yield a total expectancies score (possible total = 170). Subscale scores were not calculated. Alpha for the overall scale for the current sample was good (Alpha = .87). Self Reported Cue Reactivity Questionnaire In the present study, indivi duals self reported cue reactivity to food stimuli was assessed using 18 items. These items assess the strength, longevity, and ability to resist responses to external food cues. Three of the situations described in these items, such as When you see othe r people eating, have been taken directly from the External Eating scale of the DEBQ (van Strien et al., 1986). The remaining situations were experimenter generated to cover other situations in which individuals might respond to food related cues. Each si tuation is repeated three times, each time assessing one of the three aspects mentioned above. The items assessing how difficult the individual finds resisting the urge to eat were modified from the Food Craving Questionnaire (Lafay et al., 2001). The word s this strong urge to eat a particular food were changed to the desire. Responses were summed to create a total score (possible total = 90). Internal consistency for the scale in the present sample was very good. (Alpha = .93).

PAGE 25

17 Procedure Participants l ogged on to the University of South Florida Experimentrak system and were directed to a generic informed consent form. Upon providing consent to participate, they were redirected to the studys web page. They were informed that the study posed minimal risk that all information collected was completely confidential and anonymous, and that participation was voluntary. Participants then completed all of the measures in one of five sequences. The sequence was randomly selected in order to counterbalance across participants and reduce fatigue effects. Demographic information was provided at the end. The final two questions asked for self reported weight and height. Accuracy of reporting was stressed along with the anonymity of the information being provided. Par ticipants were then provided with a debriefing form. Course credit was awarded via the University of South Florida Experimentrack system after their participation.

PAGE 26

18 Results Relevant scale means and standard deviations are presented in Table 2. Table 2 Means and standard deviations for scale scores Scale Score Range M SD Eating Beyond Satiety 10 50 21.3 6.23 Snacking 5 25 17.3 4.13 Hunger 7 35 17.2 3.03 Night Eating 7 35 11.1 3.68 Brewer et al. (2003) Eating Beyond Sati ety Item 1 4 2.0 0.96 Brewer et al. (2003) Snacking Item 1 4 2.6 0.98 Brewer et al. (2003) Hunger Item 1 4 2.0 0.86 Brewer et al. (2003) Night Eating Item 1 4 2.3 1.03 Physical Activity 0 no limit 13957.0 101867.47 Eating Expectancies 34 238 110.5 37.06 Self Reported Cue Reactivity 18 90 43.9 14.62 Hypothesis 1 It was hypothesized that eating beyond satiety would be the best predictor of BMI when the effects of the other eating behaviors were controlled. A preliminary correlation matrix w as calculated to examine the relationships between each of the eating behaviors and BMI. Only eating beyond satiety was significantly correlated with BMI ( r = .15, p < .01). The eating behaviors were all correlated with each other (see Table 3). Correlatio ns among the items from Brewer et al. (2003) and key measures from the present study were examined. The pattern of significant and nonsignificant correlations was identical to that for the present study (see Table 4). Only the Brewer et al. (2003) item ask ing about eating

PAGE 27

19 beyond satiety was correlated with BMI ( r = .12, p < .05) and all four of the items were intercorrelated. Correlations between present study eating behavior measures and Brewer et al. (2003) eating behavior measures are presented in Table 5. Table 3 Correlations among the eating behaviors, BMI, physical activity, eating expectancies and self reported cue reactivity. Variable 1. 2. 3. 4. 5. 6. 7. 8. 1. BMI 2. Eating Beyond Satiety .15** 3. Snackin g .04 .20** 4. Hunger .08 .39** .45** 5. Night Eating .03 .34** .44** .41** 6. Physical Activity .03 .03 .10* .06 .05 7. Eating Expectancies .16** .59** .14** .34** .23** .03 8. Self Reported Cue Reactivity .05 .48** .22** .46** .29** .03 .42** p < .05, ** p < .01 Table 4 Correlations among the eating behaviors (Brewer et al. (2003) measures), BMI, physical activity, eating expectancies and self reported cue reactivity Variable 1. 2. 3. 4. 5. 6. 7. 8. 1. BMI 2. Eating Beyond Satiety .12* 3. Snacking .03 .28** 4. Hunger .02 .47** .43** 5. Night Eating .03 .33** .33** .30** 6. Physical Activity .03 .00 .08 .05 .09 7. Eating Expectancies .16* .48** .16** .32** .23** .03 8. Self Reported Cue Reactivity .05 .40** .16** .39** .27** .03 .42** p < .05, ** p < .01

PAGE 28

20 Table 5 Correlations among the eating behaviors: Present study measures vs. Brewer et al. (2003) meas ures Variable 1. 2. 3. 4. 5. 6. 7. 8. 1. Eating Beyond Satiety (EBS) 2. Snacking .20** 3. Hunger .39** .45** 4. Night Eating .34** .44** .41** 5. Brewer et al. EBS .74** .21** .33** .26** 6. Brewer et al. Snacking 22** .65** .37** .29** .28** 7. Brewer et al. Hunger .43** .42** .58** .39** .43** .43** 8. Brewer et al. Night Eating .32** .35** .30** .51** .33** .30** .30** p < .05, ** p < .01 A multiple regression analysis examined which of the eating be haviors predicts BMI after controlling for the effects of the other behaviors. The independent variables of interest were the four eating behaviors: snacking, eating beyond satiety, night eating, and hunger. All four eating behaviors were entered simultane ously. The overall regression equation was significant ( F (4,383) = 4.90, p < .01). Eating beyond satiety was a significant predictor of BMI ( b = 0.23, p < .001). Hunger also significantly predicted BMI ( b = 0.15, p < .05). Betas are presented in Table 6. The unadjusted R 2 for the model was .049 ( SE = 5.19). In addition, a multiple regression analysis examined whether the results from the Brewer et al. (2003) study replicate. BMI was regressed on the responses given by participants in the present study to t he four questions from Brewer et al. (2003). The overall regression equation was not significant, although the trend was strong ( F (4,383) = 2.29, p = .059). Betas are presented in Table 6 for the sake of comparison. The unadjusted R 2 for the model was .023 ( SE = 5.25).

PAGE 29

21 Table 6 Standardized regression coefficients and p values for the regression predicting BMI from all four eating behaviors and from the Brewer et al. (2003) items Present Study Measures Brewer Measures Variable Sta ndardized Coefficient ( b ) p value Standardized Coefficient ( b ) p value Eating Beyond Satiety 0.23 < .001 0.17 .005 Snacking 0.01 .885 0.05 .417 Hunger 0.15 .011 0.08 .173 Night Eating 0.05 .415 0.02 .759 A logistic regression was also performed t o predict obesity from all four eating behaviors entered simultaneously. In this analysis, the BMI variable was dichotomized to compare obese (BMI of 30 or above; N = 53) and normal weight individuals (BMI between 18.5 and 24.9; N = 240). Underweight (BMI 18 and below; N = 22) and overweight (BMI between 25 and 29.9; N = 72) individuals were removed from this analysis. The overall model was not significant ( c 2 (4) = 6.34, ns ). The logistic regression was repeated using the Brewer et al. (2003) measures. Once again, the overall model was not significant ( c 2 (4) = 3.71, ns ). Logistic regressions were conducted comparing only normal weight (N = 240) and overweight individuals (N = 72). In this case, the overall model was significant ( c 2 (4) = 22.42, p < .001). B oth eating beyond satiety ( OR = 1.22, p < .01; see Table 7) and night eating ( OR = 0.81, p < .01) predicted overweight. There was a trend for hunger to predict obesity ( OR = 0.82, p = .069). The Cox and Snell R 2 for the model was .21. When conducted using the Brewer et al. (2003) measures, the overall model was once again significant ( c 2 (4) = 14.53, p < .01). Eating beyond satiety ( OR = 2.89, p < .01) continued to predict overweight and hunger ( OR = 0.42, p < .05) also predicted overweight. The Cox and Snel l R 2 for the model was .14.

PAGE 30

22 Table 7 Odds ratios for the eating behaviors predicting normal weight versus overweight Present Study Measures Brewer Measures Variable OR p OR p Eating Beyond Satiety 1.22 .004 2.89 .005 Snacking 1. 01 .936 0.79 .473 Hunger 0.82 .069 0.42 .044 Night Eating 0.81 .008 0.76 .322 All of the above described linear and logistic regressions were repeated including physical activity as a control variable. The logistic regression comparing normal weight and obese individuals on present study measures of eating behaviors was significant when physical activity was controlled ( c 2 (5) = 11.56, p < .05). Only eating beyond satiety was a significant predictor of obesity ( OR = 1.06, p < .05). No other results differ ed after controlling for physical activity. One final set of logistic regressions compared overweight and obese individuals on all four eating behaviors and on the items from Brewer et al. (2003). Neither model was significant. Hypothesis 2 It was hypothes ized that race/ethnicity would moderate the relationship between the eating behaviors and BMI. Race/ethnicity was included in this analysis as a set of two dummy variables. In the first vector, African American Non Hispanic was given a value of 1 and the o ther two groups were coded as 0. In the second vector, Caucasian Non Hispanic was given a value of 1 and the other groups were coded as 0. Two interaction terms were created for each analysis by multiplying the dummy coded vectors by the eating behavior sc ore. BMI was regressed first on the eating behavior and then race/ethnicity was added to look for a main effect of race/ethnicity. Finally, the interaction terms were added to determine whether race/ethnicity moderates the

PAGE 31

23 relationship between BMI and the eating behavior. At each step of the model, the R 2 change was examined to determine if there was a significant effect. There was a main effect of race/ethnicity on BMI in each of the analyses. This main effect was explored with a oneway ANOVA where the in dependent variable was race/ethnicity and the dependent variable was BMI. Post hoc Tukey tests revealed that African American Non Hispanic ( M = 27.28, SD = 5.99) individuals had higher BMIs than Caucasian Non Hispanic ( M = 23.93, SD = 5.06) and Hispanic ( M = 23.48, SD = 5.11) individuals. The latter two groups did not differ significantly from each other (see Table 1). In the first analysis, eating beyond satiety was a significant predictor of BMI ( b = 0.18, p < .001) but there was no significant interactio n ( R 2 change = .005; F change (1,2) = .94, ns ). There was no main effect for snacking ( b = 0.03, ns ), nor was there a significant race/ethnicity by snacking interaction ( R 2 change = .000; F change (1,2) = 0.59, ns ). There was no significant main effect of n ight eating ( b = 0.06, ns ), nor was there a significant interaction ( R 2 change = .001; F change (1,2) = 0.75, ns ). There was no main effect of hunger ( b = 0.06, ns ). The interaction terms significantly predicted BMI ( R 2 change = .017; F change (1,2) = 3.47 p < .05). The hunger by race/ethnicity interaction was examined by calculating a simple regression of BMI on hunger for each race/ethnicity group and then plotting the slopes (see Figure 1). Figure 1 shows a positive association between hunger and BMI fo r African American Non Hispanic individuals. For Hispanic individuals, there was a small negative relationship and there was essentially no relationship between hunger and BMI in Caucasian Non Hispanic individuals. None of the individual regression models was significant, but Betas will be presented for the sake of interpretation of the interaction effect (see Table 8). In sum, hypothesis 2 was supported only for hunger.

PAGE 32

24 Figure 1 Regression slopes for the simple regression of BMI o n hunger for each race/ethnicity group Table 8 Standardized regression coefficients for simple regression of BMI on hunger for each race/ethnicity group Group Standardized Coefficient ( b ) F df p value African American Non Hispa nic 0.25 2.93 1, 43 .094 Caucasian Non Hispanic 0.09 2.23 1, 285 .136 Hispanic 0.21 2.50 1, 54 .120

PAGE 33

25 The relationship between race/ethnicity and the eating behaviors was explored further. An ANOVA was conducted for each eating behavior wherein race/et hnicity was the independent variable and the eating behavior was the dependent variable. For eating beyond satiety, there was a main effect of race/ethnicity ( F (2,385) = 3.32, p < .05; see Table 9). Post hoc Tukey tests revealed that African American Non H ispanics ( M = 19.04, SD = 4.80) had significantly lower eating beyond satiety scores than Caucasian Non Hispanic ( M = 21.59, SD = 6.45) individuals but did not differ significantly from Hispanic individuals ( M = 21.46, SD = 5.80). Caucasian Non Hispanic an d Hispanic individuals also did not differ significantly. There was also a main effect for snacking ( F (2,385) = 3.79, p < .05). Post hoc Tukey tests revealed that the only significant difference was that Hispanic ( M = 13.52, SD = 2.73) individuals had sign ificantly lower snacking scores than Caucasian Non Hispanic ( M = 15.02, SD = 3.83) individuals. There was no main effect for hunger ( F (2,3850 = 2.78, p = .063), although the trend was strong. The same was true of night eating ( F (2,385) = 2.88, p = .057). H ypothesis 3 It was hypothesized that the eating behaviors would mediate the relationship between eating expectancies and BMI. A prerequisite for mediation analysis is that the independent, dependent and mediator variables all be correlated. Therefore, eati ng expectancies, BMI, and eating beyond satiety were correlated to determine if this criterion was met (see Table 3). Eating expectancies were significantly correlated with eating beyond satiety ( r = .59, p < .001) and with BMI ( r = .16, p < .001). As repo rted above, eating beyond satiety and BMI were significantly correlated ( r = .15, p < .01). BMI was regressed on eating beyond satiety and eating expectancies. With both variables in the model, neither eating beyond satiety ( b = 0.10, p = .12), nor eating expectancies ( b = 0.10, p = .11) were significant predictors of BMI. However, there was a fairly substantial drop in the beta weight for eating expectancies when the mediator was added to the regression (from b = 0.155 to b = 0.099). A Sobel test (Sobel, 1 982) was performed as a check on the mediational model. The Sobel test confirmed that mediation was not significant ( z = 1.55, p = .12).

PAGE 34

26 No further mediation analyses could be conducted because the remainder of the eating behaviors were not correlated wit h BMI. Hypothesis 3 was, therefore, not supported. Table 9 Race/ethnicity differences on the eating behaviors Group M SD Eating Beyond Satiety African American Non Hispanic 19.04* 4.80 Caucasian Non Hispanic 21.59* 6.45 Hispani c 21.46 5.80 Snacking African American Non Hispanic 14.44 4.76 Caucasian Non Hispanic 15.02* 3.83 Hispanic 13.52* 2.73 Hunger a African American Non Hispanic 15.27 3.59 Caucasian Non Hispanic 16.64 3.71 Hispanic 16.27 3.45 Night Eating a African Am erican Non Hispanic 12.29* 4.65 Caucasian Non Hispanic 10.89* 3.51 Hispanic 10.92 3.54 *Means differ significantly. a ANOVA was not significant. Hypothesis 4 It was hypothesized that the eating behaviors would mediate the relationship between self repor ted cue reactivity and BMI. In this case, a mediation analysis could not be conducted because BMI and self reported cue reactivity were not significantly correlated ( r = .05, ns ; see Table 3) precluding the possibility of a mediational analysis. Hypothesis 4 was, therefore, not supported.

PAGE 35

27 Hypothesis 5 It was hypothesized that eating beyond satiety would account for additional variance in BMI beyond that accounted for by a diagnosis of BED. Individuals were categorized by binge eating status. Individuals who according to the EDDS scoring algorithm, met criteria for BED or subclinical BED were placed in the binge eaters category. These two subgroups were combined due to the small number of these individuals (5 subclinical and 14 clinical). Participants who di d not meet these criteria were classified as non binge eaters. It could be that the effects of compensatory behaviors such as purging would attenuate the effects of binge eating on BMI because purging negates the caloric intake of the binge. Therefore, ind ividuals who engage in compensatory behaviors such as purging, defined as anyone who scored above 0 on any one of EDDS items 15 through 18, were removed from all subsequent analyses. BMI was regressed on binge eating status and then eating beyond satiety w as added into the model. In the first step, binge eating status was a significant predictor of BMI ( b = 0.27, p < .001). Binge eating status continued to predict BMI when eating beyond satiety was added ( b = 0.24, p < .001). Eating beyond satiety did not a ccount for a statistically significant amount of additional variance in BMI although there was a very strong trend in that direction ( b = 0.10, p = .05). An ANOVA was conducted to further examine the relationship between binge eating status and BMI. Binge eaters ( M = 30.36, SD = 7.78) had a significantly higher mean BMI than non binge eaters ( M = 23.91, SD = 4.98; F (1,375) = 28.30, p < .001). A second ANOVA examined the relationship between binge eating status and eating beyond satiety scores. Binge eaters ( M = 28.00, SD = 7.01) had a significantly higher mean eating beyond satiety score than non binge eaters ( M = 20.87, SD = 6.01; F (1,375) = 24.94, p < .001). A binge score was created next to examine binge eating as a continuous rather than a categorical ph enomenon. This score was created by taking the sum of the items on the EDDS relevant to binge eating. Items 7 and 8 were dichotomized to be on the same metric as the remaining binge items (items 5 6 and 9 14). Items 7 and 8 were recoded to have a value of 0 if the participant scored lower than 2 and a value of 1 if the participant

PAGE 36

28 scored 2 or higher. These values correspond to the cutoff on the EDDS scoring algorithm (Stice et al., 2000). The internal consistency of this binge score was high (alpha = .91). Once again, individuals who engage in compensatory behaviors were left out of all analyses. The mean binge score for all binge eaters was 2.95 ( SD = 3.36). Possible scores ranged from 0 to 10. The correlation between binge scores and BMI was small ( r = .19 p < .001) but significant. BMI was regressed on binge scores and eating beyond satiety. With both variables in the model, binge scores predicted BMI ( b = 0.15, p < .05) but eating beyond satiety did not account for a statistically significant amount of a dditional variance in BMI ( b = 0.07, ns ). Hypothesis 5 was, therefore, not supported.

PAGE 37

29 Discussion The present study asked which eating behaviors predict BMI. Both eating beyond satiety and hunger were identified as significant predictors of BMI. Brewer et al. (2003) also identified eating beyond satiety as a predictor of obesity in a sample of older Caucasian and African American women. There are several studies that have found relationships between binge eating and obesity or overweight (e.g. Picot & Lilen feld, 2003; Siqueira, Appolinario, & Sichieri, 2004). The present study was the first to examine the less severe, but likely more common phenomenon, of eating beyond satiety in the college population. Interestingly, Stice, Presnell, Shaw and Rohde (2005) failed to identify a prospective relationship between binge eating and obesity. In contrast, Brewer et al. (2003) did find that eating beyond satiety was related to age of obesity onset. However, their data were retrospective and could be subject to recall bias. It could be that, while binge eating or overeating do not cause obesity, they could act as maintaining factors. Losing weight is difficult when caloric intake is high and, therefore, behaviors such as eating beyond satiety would serve to impede weig ht loss. Because the current study is cross sectional, examining current weight rather than obesity onset, the results cannot answer this question. Alternatively, some third variable could account for both BMI and eating beyond satiety, thereby precluding a causal link between them. It is difficult to think what such a variable might be, but there could be some genetic factor that predisposes an individual to overeat and to have a high BMI. Future research will be needed to sort out possible alternative exp lanations and to further clarify the temporal relationships among these variables. It should be noted that, in the present study, only 5% of the variance in BMI was explained by eating behaviors, indicating that, while these factors do have a role to play

PAGE 38

30 there are other factors that account for much more variance in the relationship. And, while eating beyond satiety is an important predictor, the correlation between this variable and BMI is still quite low ( r = .15). This fact does not, however, diminish the importance of the findings in the present study. BMI is determined by a host of genetic and environmental factors and the more of these that can be identified, the more tools there are in the arsenal to prevent and treat obesity and overweight. One no table finding concerning eating beyond satiety is its ability to predict overweight, specifically when using the single item measure from Brewer et al. (2003). The odds ratio in this case was 2.89 indicating that individuals had an almost 3 times greater c hance of being overweight if they engaged in eating beyond satiety. This is quite a large effect and it is interesting to note that the predictive power of this single item measure appears to be greater than that of the 10 item questionnaire designed for t he present study ( OR = 1.22). This finding, however, is likely explained by the fact that the Brewer et al. (2003) item has a possible score range of 1 4 while the present study measure had a range of 10 50. The two odds ratios are not directly comparable because they are not on the same metric. The Brewer et al. (2003) measure should, therefore, not be interpreted as a stronger predictor. The present study apparently stands in contrast to the findings of Hays et al. (2002) who found that hunger did not pre dict current BMI. The present findings appear compatible with Boschi et al. (2001), who found that obese women obtained higher hunger scores than did normal weight women. However, the results of both Hays et al. (2002) and Boschi et al. (2001) must be inte rpreted with caution. The hunger scale of the TFEQ, the measure used in both studies, does not provide a pure measure of the hunger sensation, instead confounding the feeling of hunger with the behavior of eating. It is possible that hunger increases the l ikelihood that maladaptive overeating will take place, but studies have yet to confirm this possibility. Hill (1974) found that when individuals are hungry, they eat more rapidly, have longer meals, and consume more food. The model makes logical sense: inc reased hunger leads to eating beyond satiety which, in turn, results in increased caloric intake and a higher BMI. Nevertheless, the most likely

PAGE 39

31 reason for the findings related to hunger in the present study is suppression. Hunger acts as a suppressor vari able which according to Conger (1974 as cited in Glasnapp, 1984) is defined to be a variable which increases the predictive validity of another variable by its inclusion in a regression equation (p. 858). It is also interesting to note that the beta for hunger is negative indicating that increased hunger is associated with decreased BMI. The same is true of the odds ratio in the logistic regression. Because it is less than one, the odds ratio indicates that odds of being overweight decrease as hunger sco res increase. One possible explanation for the direction of this relationship could be that heavier individuals eat so much or so constantly that they do not experience hunger. Normal weight individuals, on the other hand, may eat less frequently or consum e smaller portions, thus allowing themselves to experience hunger. Perhaps normal weight individuals use hunger as a cue to signal when it is time to eat, while overweight and obese individuals do not give themselves the chance to experience hunger. Past r esearch on snacking has not consistently identified it as a predictor of obesity. The current study adds to the null findings in the literature, further strengthening confidence that snacking is not, in and of itself, significantly related to obesity. It c ould be that individuals who snack still consume an appropriate number of calories in a day, and therefore would not be expected to be overweight or obese. In fact, Fabry and Tepperman (1970) concluded that an infrequent meal pattern was related to a tende ncy toward obesity. Snackers may spread out their calories over the course of the day rather than eating them all at once, thus engaging in a potentially protective eating pattern. Frequent snacking may serve to reduce hunger, which, according to Hill (197 4) results in a more rapid eating rate and greater food consumption. Both of these factors very likely lead to eating beyond satiety. Eating rate could play an important role in the relationships among these eating behaviors and BMI. Sasaki, Katagiri, Tsuj i, Shimoda, and Amano (2003) found a positive relationship between BMI and eating rate. This factor was not considered in the present study but its relationship with the eating behaviors is of potential interest in future research.

PAGE 40

32 The present study confir ms the previous null results that night eating does not predict obesity (Brewer et al., 2003; Kant et al., 1995). In a logistic regression comparing normal weight and overweight individuals, however, the odds ratio for night eating was significant. The odd s ratio was less than one, indicating that the odds of being overweight decrease as night eating increases. Perhaps individuals who engage in night eating eat enough excess calories to become overweight but engage in some sort of compensatory behavior, suc h as physical activity or purging, that maintains their weight at the normal weight level. Or, it could be that as long as one eats a reasonable amount of calories in a day, weight gain will not occur, regardless of the distribution of those calories over the course of the day. Individuals who are prone to night eating may also be prone to morning anorexia (Stunkard, 1955 as cited in Napolitano, Head, Babyak, & Blumenthal, 2001), thereby simply shifting their patterns of consumption to later in the day, rat her than increasing caloric intake. In this case, night eating would not result in a higher BMI. Of note in the present study are the low correlations between physical activity and BMI, and physical activity and the eating behaviors. The correlations were negative but were all very small in magnitude with the largest being .10 (between the Brewer et al. snacking measure and physical activity). The negative direction of the correlations makes intuitive sense because one would expect greater physical activity to be associated with a lower BMI and, potentially, a healthier lifestyle and thus fewer maladaptive eating patterns. The small magnitude of the correlations suggests that physical activity is a less important determinant of BMI and eating behaviors than other factors. Brewer et al. (2003) also identified race as a moderator in the relationship between eating beyond satiety and obesity. Race/ethnicity was not found to moderate the relationship between eating beyond satiety and BMI in the present study. It is possible that the relationship does not hold in the college population; Brewer et al. (2003) examined premenopausal women. The discrepancy in findings could also be explained by the difference in the statistical techniques used. Brewer et al. (2003) cla imed that the relationship between eating beyond satiety and BMI was stronger in African American women than in Caucasian women because the odds ratio was larger. They did not,

PAGE 41

33 however, test to see if these ratios were significantly different and, in fact, the confidence intervals did overlap. The present study, on the other hand, looked for interactions with a more sensitive multiple regression technique. In contrast, race/ethnicity moderated the relationship between hunger and BMI. Regression slopes indic ated that little relationship existed between BMI and hunger in Caucasian Non Hispanic participants and there was a small, negative relationship in Hispanic participants. In African American Non Hispanic participants, there was a strong, positive relations hip between hunger and BMI. African Americans have among the highest rates of obesity in the country. Any variable that sheds light on the possible reasons for this should be considered important. Perhaps African Americans are more sensitive to hunger cues or less able to resist the feeling of hunger. It is interesting however, that while hunger and BMI were more strongly related in African American Non Hispanics than in the other two groups, African Americans also had a lower eating beyond satiety mean th at Caucasian Non Hispanics. Future research must elucidate this relationship further. Eating expectancies were found to correlate positively with BMI but, contrary to the study hypothesis, eating beyond satiety did not mediate this relationship. Both varia bles contribute to the prediction of BMI. This is the first study to systematically evaluate the effects of eating expectancies on BMI. Expectancies have long been known to predict behavior in the alcohol arena. Christiansen, Smith, Roehling, and Goldman ( 1989) found that alcohol expectancies at the age of 12 years predicted problem drinking one year later. Such findings in the obesity literature would extend the applicability of expectancy theory and make more explicit the similarities between unhealthy, a ddictive behaviors such as alcohol abuse and smoking, and a more necessary behavior like eating. It is important to note that eating is different from smoking or drinking because one cannot just go cold turkey. Eating is necessary to sustain life. Indivi duals, therefore, must be taught to adopt a healthier lifestyle rather than to go without food. Treatment goals in this case would be analogous to controlled drinking as opposed to abstinence. Controlled drinking has been suggested as a treatment goal by s ome, and may be

PAGE 42

34 effective in some cases (see Cloud, McKiernan, & Cooper, 2003 for a review). However, also important to note is that there are those individuals who have trouble stopping their food consumption when they have had enough. This finding opens the door to the examination of eating as an addictive behavior. Expectancy challenges have been found to be effective in reducing problem drinking in the college population (e.g. Darkes & Goldman, 1993). Challenging an individuals expectancies about the p ositive effects of food or eating could prove to be an effective treatment for obesity. Helping individuals to realize that food is not an effective coping mechanism and helping them to find more adaptive coping mechanisms is important for those individual s who turn to food for solace or relief from negative affect, for example. Self reported cue reactivity was correlated with all four eating behaviors. However, because cue reactivity was not significantly correlated with BMI, a mediation analysis could no t be conducted. It is curious that, in this case, self reported cue reactivity did not predict BMI. It makes sense that self reported cue reactivity would predict eating beyond satiety. Those individuals who feel stronger, longer, more difficult to resist urges when confronted with food cues or food related stimuli are also those who are more likely to eat more than they need to. However, these individuals should also be those with higher BMI. It is possible that the self reported cue reactivity scale was n ot powerful enough to detect food related cue reactivity. Studies of cue reactivity tend to use slides or in vivo exposure to cues and then measure physiological and affective reactions to those cues. A self report scale may not be able to detect these sam e reactions. In fact, participants may not even realize that these associations exist. Classical and operant conditioning operate outside of conscious realization. Therefore, a more behavioral paradigm may be required to detect such associations. Another question posed by the present study was whether eating beyond satiety represents a phenomenon distinct from binge eating or whether it represents a form of subclinical binge behavior. The results suggest that, in fact, eating beyond satiety does not accoun t for variance in BMI over and above binge eating, whether considered as a discrete diagnosis or a continuum of symptomatology. These findings suggest that

PAGE 43

35 overeating may best be viewed as a continuum with eating beyond satiety representing a lesser degree of overeating while binge eating represents its clinical manifestation. In this case, the question becomes whether, for some individuals, this subclinical overeating might precede more severe binge episodes. It is likely that, in some individuals, escalat ion is unlikely. However, for others, there may be a tendency to start small and get bigger. In this case, identification of predictors and risk factors for eating beyond satiety becomes key. Once identified, these predictors can inform treatment and preve ntion efforts which should lead to two effects. One would be to affect BMI since eating beyond satiety is a significant predictor of BMI. The other would be to prevent future eating disorders, whether bulimia or binge eating disorder, which are both charac terized by episodes of binge eating. The hypothesized model is thus: eating beyond satiety represents a subclinical form of binge eating. In some, this subclinical symptom progresses to the point of becoming a clinically significant binge eating episode. B y treating and/or preventing eating beyond satiety, we prevent the development of future eating disorders. It should be noted, however, that there was a strong trend for eating beyond satiety to account for additional variance in BMI when binge eating was considered as a categorical variable. This finding is puzzling because it stands in contrast to what is found when binge eating is considered as a continuous variable. It is difficult to explain why this might be. Limitations There is a danger that in a sa mple of undergraduate students, the variability of weight may be limited. However, while the percentages of individuals who fall into each BMI category did not match population values, the sample in the present study had good variability ( SD = 5.29). Other limitations include the cross sectional nature of the data. No causal conclusions can be drawn, but future prospective studies can help clarify causal issues. A self report bias may also be present in the reporting of height and weight, but research has s hown that self reported weight is, in general, fairly accurate. Rowland (1990) compared actual and self reported weight and found that men tend to overreport weight

PAGE 44

36 by 0.4 kg and that women tend to underreport weight by 1.0 kg. He also found that men overr eport their height by 1.4 cm and that women overreport their height by 0.6 cm. In the present study, participants were anonymous and were reporting from the privacy of their own computers. Another potential weakness is with the BMI measurement itself. Whi le it is widely used as the standard in obesity research, it can fail to reflect other aspects of body composition such as high levels of muscularity (Stice et al., 2005) or central adiposity, which has been shown to be an independent predictor of cardiova scular reactivity even in adolescents (Goldbacher, Matthews, & Salomon, 2005). Finally, this study examined only college age females in three specific racial/ethnic groups and may not generalize to males and to members of other racial/ethnic groups. Futu re Research The present study raises more questions than it answers. Future research is needed to identify whether or not eating beyond satiety is predictive of obesity onset or, rather, if it acts as a maintaining factor. The role of hunger in the path to obesity also continues to remain unclear. While hunger is predictive of current BMI, it is also possible that it plays a role in a causal path to overweight by encouraging overeating. Future research can also identify whether individuals who show stronger reactions during cue exposure tasks are those with higher BMI. More research is also needed to determine risk factors for and predictors of eating beyond satiety. Another potentially fruitful path is in the implementation of treatment and prevention stra tegies. Incorporating training on learning to identify satiety signals and to cease eating when they are recognized may be a fertile avenue in the treatment of obesity. If, in fact, eating beyond satiety is predictive of obesity onset, then prevention effo rts focusing on teaching individuals to stop eating when they are full could prove worthwhile. The present study replicated the finding that eating beyond satiety predicts BMI and has extended this finding to the college age group. It appears, therefore, that this is a

PAGE 45

37 fairly robust phenomenon. Importantly, the present study has opened the field of eating behaviors as a potentially rewarding avenue in the study if the causes, treatments, and prevention of overweight and obesity, considerable problems for contemporary society.

PAGE 46

38 References Atlas, J. G., Smith, G. T., Hohlstein, L. A., McCarthy, D. M., & Kroll, L. S. (2002). Similarities and differences between Caucasian and African American college women on eating and dieting expectancies, bulimic symptoms, dietary restraint, and disinhibition. International Journal of Eating Disorders, 32 326 334. Basdevant, A., Craplet, C., & Guy Grand, B. (1993). Snacking patterns on obese French women [Electronic version]. Appetite, 21 17 23. Berteus Forslund, H., Lin droos, A. K., Sjostrom, L., & Lissner, L. (2002). Meal patterns and obesity in Swedish women A simple instrument describing usual meal types, frequency and temporal distribution. European Journal of Clinical Nutrition, 56 740 747. Boschi, V., Iorio, D. Margiotta, N., DOrsi, P., & Falconi, C. (2001). The three factor eating questionnaire in the evaluation of eating behavior in subjects seeking participation in a dietotherapy program. Annals of Nutrition and Metabolism, 45 72 77. Braet, C., van Strien T. (1997). Assessment of emotionally induced and restrained eating behavior in nine to twelve year old obese and non obese children. Behavior Research and Therapy, 35 863 873. Brewer, E. A., Kolotkin, R. L., & Baird, D. D. (2003). The relationship betw een eating behaviors and obesity in african american and caucasian women [Electronic version]. Eating Behaviors, 4 159 171. Christiansen, B. A., Smith, G. T., Roehling, P V., & Goldman, M. S. (1989). Using alcohol expectancies to predict adolescent dri nking behavior after one year. Journal of Consulting and Clinical Psychology, 53 93 99.

PAGE 47

39 Cloud, R. N., McKiernan, P., & Cooper, L. (2003). Controlled drinking as an appropriate treatment goal: A critique of current approaches. Alcoholism Treatment Quar terly, 21 67 82. Courneya, K. S., & Hellsten, L. M. (1998). Personality correlates of exercise behavior, motives, barriers and preferences: An application of the five factor model. Personality and Individual Differences, 24 625 633. Darkes, J. & Goldman M. S., (1993). Expectancy challenge and drinking reduction: Experimental evidence for a mediational process. Journal of Consulting and Clinical Psychology, 61 344 353. DeAngelis, T. (2004). Family size portions for one [Electronic version]. Monitor on Psychology, 35 50. Epstein, L. H., Paluch, R., & Coleman, K. J. (1996). Differences in salivation to repeated food cue s in obese and nonobese women [E lectronic version]. Psychosomatic Medicine, 58 160 164. Fabry, P., & Tepperman, J. (1970). Meal freque ncy A possible factor in human pathology. The American Journal of Clinical Nutrition, 8 1059 1068. Glasnapp, D. R. Change scores and regression suppressor conditions (1984). Educational and Psychological Measurement, 44 851 867. Goldbacher, E. M., Ma tthews, K. A., & Salomon, K. (2005). Central adiposity is related to cardiovascular reac tivity to stress in adolescents Health Psychology, 24 375 384. Hays, N. P., Bathalon, G. P., McCrory, M. A., Roubenoff, R., Lipman, R., & Roberts, S. B. (2002). Eat ing behavior correlates of adult weight gain and obesity in healthy women aged 55 65 y [Electronic version]. American Journal of Clinical Nutrition, 75 476 483. Hill, J. O. (1998). Environmental contributions to the obesity epidemic [Electronic version] Science, 280 1371 1375. Hill, S. W. (1974). Eating responses of humans during dinner meals. Journal of Comparative and Physiological Psychology, 86 652 657.

PAGE 48

40 Hohlstein, L. A., Smith, G. T., & Atlas, J. G. (1998). An application of expectancy theory t o eating disorders: Development and validation of measures of eating and dieting expectancies [Electronic version]. Psychological Assessment, 10 49 58. Jansen, A. (1998). A learning model of binge eating: Cue reactivity and exposure. Behavior Research an d Therapy, 36 257 272. Jansen, A., Broekmate, J, & Heymans, M. (1992). Cue exposure vs self control in the treatment of binge eating: A pilot study. Behavior Research and Therapy, 30 235 241. Jansen, A., Theunissen, N., Slechten, K., Nederkoorn, C., Boo n, B., Mulkens, S., & Roefs, A. (2003). Overweight children overeat after exposure to food cues [Electronic version]. Eating Behaviors, 4 197 209. Jones, B. T., Corbin, W., Fromme, K. (2001). A review of expectancy theory and alcohol consumption. Addict ion, 96 57 72. Kant, A. K., Ballard Barbash, R., & Schatzkin, A. (1995). Evening eating and its relation to self reported body weight and nutrient intake in women, CSFII 1985 86. Journal of the American College of Nutrition, 14 358 363. Lafay, L., Thom as, F., Mennen, L., Charles, M. A., Eschwege, E., Borys, J., Basdevant, A. (2001). Gender differences in the relation between food cravings and mood in an adult community: Results from the Fleurbaix Laventie Ville sant study. International Journal of Eat ing Disorders, 29 195 204. Lee, I. M., Paffenbarger, R. S., & Hsieh, C. C. (1992). Time trends in physical activity among college alumni, 1962 1988. American Journal of Epidemiology, 135 915 925. Lowe, M. R., & Maycock, B. (1988). Restraint, disinhibi tion, hunger and negative affect eating. Addictive Behaviors, 13 369 377. Marshall, H. M., Allison, K. C., OReardon, J. P., Birketvedt, G., & Stunkard, A. J. (2004). Night eating syndrome among nonobese persons. International Journal of Eating Disorder s, 35 217 222. Mattes, R. D. (1997). Physiologic responses to sensory stimulation by food: Nutritional implications. Journal of the American Dietetic Association, 97 406 413.

PAGE 49

41 National Center for Health Statistics. (2004a). Overweight Prevalence Retriev ed April 14, 2004 from http://www.cdc.gov/nchs/fastats/overwt.htm National Center for Health Statistics. (2004b). Table 68: Overweight, obesity and healthy weight among persons 20 years of age and over, according to sex, age, race, and Hispanic origin: United States, 1960 62, 1971 74, 1976 80, 1988 94, and 1999 2000 Retrieved April 14, 2004 from http://www.cdc.gov/nchs/fastats/overwt.htm. National Center for Health Statistics. (2004c). Table 69 : Overweight children and adolescents 6 19 years of age according to sex, age, race, and Hispanic origin: United States, selected years 1963 65 through 1999 2000 Retrieved April 14, 2004 from http://www.cdc.gov/nchs/fastats/overwt.htm. National Heart, Lu ng and Blood Institute. (1998). Clinical guidelines for the identification, evaluation, and treatment of overweight and obesity in adults: The evidence report [Electronic version]. Bethesda (MD): National Heart, Lung and Blood Institute. Retrieved Novembe r 11, 2003 from http://www.nhlbi.nih.gov/guidelines/obesity/ob_gdlns.htm Napolitano, M. A., Head, S., Babyak, M. A., & Blumenthal, J. A. (2001). Binge eating disorder and night eat ing syndrome: Psychological and behavioral characteristics. International Journal of Eating Disorders, 30 193 203. Nederkoorn, C. Smulders, F. T.Y., & Jansen, A. (2000). Cephalic phase responses, craving, and food intake in normal subjects. Appetite, 35 45 55. Nicklas, T. A., Baranowski, T., Cullen, K. W., & Berenson, G. (2001). Eating patterns, dietary quality and obesity Journal of the American College of Nutrition, 20 599 608. Ortega, R. M., Redondo, M. R., Zamora, M. J., Lopez Sobaler, A. M., & A ndres, P. (1995). Eating behavior and energy intake in overweight/obese and normal weight Spanish elderly. Annals of Nutrition and Metabolism, 39 371 378. Paffenbarger, R. S., Wing, A. L., & Hyde, R. T., (1978). Physical activity as an index of heart at tack risk for college alumni. American Journal of Epidemiology, 108 161 175.

PAGE 50

42 Picot, A. K., & Lilenfeld, L. R. R. (2003). The relationship among binge severity, personality psychopathology, and body mass index. International Journal of Eating Disorders, 3 4 98 107. Rowland, M. L. (1990). Self reported weight and height. American Journal of Clinical Nutrition, 52 1125 1133. Sasaki, S., Katagiri, A., Tsuji, T., Shimoda, T., & Amano, K. (2003). Self reported rate of eating correlates with body mass index i n 18 y old Japanese women. International Journal of Obesity & Related Metabolic Disorders, 27 1405 1410. Siqueira, K. S., Appolinario, J. C., & Sichieri, R. (2004). Overweight, obesity and binge eating in a non clinical sample of five Brazilian cities. O besity Research, 12 1921 1924. Sobel, M. E. (1982). Asymptotic intervals for indirect effects in structural equations models. In S. Leinghart ( Ed.), Sociological methodology (pp. 290 312). San Francisco: Jossey Bass. Stice, E., Fisher, M. & Martinez, E (2004). Eating disorder diagnostic scale : Additional evidence of reliability and validity. Psychological Assessment, 16 60 71. Stice, E., Presnell, K., Shaw, H., & Rohde, P. (2005). Psychological and behavioral risk factors for obesity onset in adoles cent girls: A prospective study. Journal of Consulting and Clinical Psychology, 73 195 202. Stice, E., Telch, C. F., & Rizvi, S. L. (2000). Development and validation of the Eating Disorder Diagnostic Scale: A brief self report measure of anorexia, buli mia, and binge eating disorder [Electronic version]. Psychological Assessment, 12 123 131. Stunkard, A. J., & Messick, S. (1985). The three factor eating questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Res earch, 29 71 83. Summerbell, C. D., Moody, R. C., Shanks, J., Stock, M. J., & Geissler, C. (1995). Sources of energy from meals versus snacks in 220 people in four age groups. European Journal of Clinical Nutrition, 49 33 41.

PAGE 51

43 Togo, P., Osler, M., Soren son, T. I. A., & Heitmann, B. L. (2001). Food intake patterns and body mass index in observational studies. International Journal of Obesity, 25 1741 1751. Tuomisto, T., Tuomisto, M. T., Hetherington, M., & Lappalainen, R. (1998). Reasons for initiation and cessation of eating in obese men and women and the affective consequences of eating in everyday situations [Electronic version]. Appetite, 30 211 222. van Strien, T., Frijters, J. E. R., Bergers, G. P. A., & Defares, P.B. (1986). The Dutch eating be havior questionnaire (DEBQ) for assessment of restrained, emotional, and external eating behavior. International Journal of Eating Disorders, 5 295 315. Yanovski, J. A., & Yanovski S. Z. (1999) Recent advances in basic obesity research [Electronic versi on]. Journal of the American Medical Association, 282 1504 1506.

PAGE 52

44 Appendices

PAGE 53

45 Appendix A: Paffenbarger Physical Activity Questionnaire 1. How many city blocks or their equivalent do you normally walk each day ? _____blocks/day (Let 12 blocks = 1 mile) 2. What is your usual pace of walking? (Please check one.) a. _____Casual or strolling (less than 2 mph) b. _____Average or normal (2 to 3 mph) c. _____Fai rly brisk (3 to 4 mph) d. _____Brisk or striding (4 mph or faster) 3. How many flights of stairs do you climb up each day ? _____flights/day (Let 1 flight = 10 steps) 4. List any sports or recreation you have actively participated in during the past year Please rem ember seasonal sports or events. Average Time/Episode Sport, Recreation, or Other Physical Activity Number of Times/year Hours Minutes Years Participation a. b. c. d. e. f. 5. Which of these statements best expresses yo ur view (Please check one.) a. _____I take enough exercise to keep healthy. b. _____I ought to take more exercise. c. _____Dont know

PAGE 54

46 Appendix A (Continued) 6. At least once a week do you engage in regular activity akin to brisk walking, jogging, bicycling, swimming, etc. long enough to wor k up a sweat, get your heart thumping, or get out of breath? _____No Why not? _____________________________________________ _____Yes How many times per week? _____ Activity:_____ 7. When exercising in your usual fashion, how would you rate your usual level of exertion (degree of effort)? (Please circle one number.) _____0 Normal _____0.5 Very very weak (just noticeable) _____1 Very weak _____2 Weak _____3 Moderate _____4 Somewhat Strong _____5 Strong (heavy) _____6 _____7 Very strong _____8 _____9 ___ __10 Very very strong (almost maximal)

PAGE 55

47 Appendix A (Continued) 8. On a usual weekday and a weekend day, how much time do you spend on the following activities? Total each day should add up to 24 hours. Usual Weekday Hours/day Usual Weekend Day Hours/day a. Vigorous activity (diggi ng in the garden, strenuous sports, jogging, aerobic dancing, sustained swimming, brisk walking, heavy carpentry, bicycling on hills, etc.) b. Moderate activity (housework, light sports, regular walking, golf, yard work, lawn mowing, painting, repairing light carpentry, ballroom dancing, bicycling on level ground, etc.) c. Light activity (office work, driving car, strolling, personal care, standing with little motion, etc.) d. Sitting activity (eating, reading, desk work, watching TV, listening to radio, etc.) e. Sleeping or reclining

PAGE 56

48 Appendix B: Eating Disorder Diagnostic Scale Please carefully complete all questions. Over th e past 3 months Not at all Slightly Moderately Extremely 1. Have you felt fat? 0 1 2 3 4 5 6 2. Have you had a 0 1 2 3 4 5 6 definite fear that you might gain weight or become fat? 3. Has your weight 0 1 2 3 4 5 6 infl uenced how you think about (judge) yourself as a person? 4. Has your shape 0 1 2 3 4 5 6 influenced how you think about (judge) yourself as a person? 5. During the past 6 months have there been times when you felt you have eat en what other people would regard as an unusually large amount of food (e.g. a quart of ice cream) given the circumstances? YES NO 6. During the times when you ate an unusually large amount of food, did you experience a loss of control (feel you could nt stop eating or control what or how much you were eating)? YES NO 7. How many DAYS per week on average over the past 6 MONTHS have you eaten an unusually large amount of food and experienced a loss of control? 0 1 2 3 4 5 6 7 8. How many TIMES per week on average over the past 3 MONTHS have you eaten an unusually large amount of food and experienced a loss of control? 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

PAGE 57

49 Appendi x B (Continued) During these episodes of overeating and loss of control did you 9. Eat much more rapidly than normal? YES NO 10. Eat until you felt uncomfortably full? YES NO 11. Eat large amounts of food when you didnt feel physically hungry? YES N O 12. Eat alone because you were embarrassed by how much you were eating? YES NO 13. Feel disgusted with yourself, depressed, or very guilty after overeating? YES NO 14. Feel very upset about your uncontrollable overeating or resulting weight gain? YES NO 15. How many times per week on average over the past 3 months have you made yourself vomit to prevent weight gain or counteract the effects of eating? 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 16. How many times per week on average over the past 3 months have you used laxatives or diuretics to prevent weight gain or counteract the effects of eating? 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 17. How many times per week on average over the past 3 months have you fasted (skipped at least 2 meals in a row) to prevent weight gain or counteract the effects of eating? 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 18. How many times p er week on average over the past three months have you engaged in excessive exercise specifically to counteract the effects of overeating episodes? 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

PAGE 58

50 Appendix C: Ea ting Beyond Satiety Questionnaire 1. I continue to eat after I am full. 1 2 3 4 5 never sometimes often almost always always 2. I keep on eating even though I am no longer hungry. 1 2 3 4 5 never sometim es often almost always always 3. When I start eat ing, I just cant seem to stop. 1 2 3 4 5 never sometimes often almost always always 4. I eat until I feel uncomfortably full. 1 2 3 4 5 never sometimes often almost always always 5. I eat food even when I am not hungry. 1 2 3 4 5 never sometimes often almost always always 6. I feel uncomfortably full after eating. 1 2 3 4 5 never sometimes often almost always always 7. I leave some food on my plate. 1 2 3 4 5 never sometimes often almost always always 8. Even though I am full, it is hard for me to stop eating before I finish the food on my plate. 1 2 3 4 5 never sometimes often almost always always

PAGE 59

51 Appendix C (Continued) 9. I have seconds at a meal even though I am full. 1 2 3 4 5 never once a week several times daily more than a week once a day 10. When I am full, I skip dessert. 1 2 3 4 5 never sometimes often almost always always

PAGE 60

52 Appendix D: Snacking Questionnaire 1. I eat ONLY breakfast, lunch, and dinner nothing else. 1 2 3 4 5 never sometimes often almost always always 2. I eat between meals. 1 2 3 4 5 never once a week several times daily more than a week once a day 3. How often do you have a snack in the morning? 1 2 3 4 5 less than once 1 or 2 days 3 or 4 days 5 or 6 days every day a week per week per week per we ek 4. How often do you have a snack in the afternoon? 1 2 3 4 5 less than once 1 or 2 days 3 or 4 days 5 or 6 days every day a week per week per week per week 5. How often do you have a snack in the evening? 1 2 3 4 5 less than once 1 or 2 days 3 or 4 days 5 or 6 days every day a week per week per week per week 6. I am the kind of person who eats 3 meals a day: breakfast, lunch and dinner. 1 2 3 4 5 less than once 1 or 2 days 3 or 4 days 5 or 6 days every day a week per week per week per week

PAGE 61

53 Appendix E: Hunger Questionnaire 1. I get so hungry that my stomach feels like a bottomless pit. 1 2 3 4 5 never sometimes often almost always always 2. After breakfast, how long is it before you feel hungry enough to eat again? 1 2 3 4 5 0 1 hours 1 2 hours 2 3 hours over 3 hours I dont 3. After lunch, how long is it before you feel hungry enough to eat again? 1 2 3 4 5 0 1 hours 1 2 hours 2 3 hours over 3 hours I dont 4. After dinner, how long is it before you feel hungry enough to eat again? 1 2 3 4 5 0 1 hours 1 2 hours 2 3 hours over 3 hours I dont 5. I am hungry enough to eat at any time. 1 2 3 4 5 never sometimes often almost always always 6. My stomach rumbles because I am so hungry. 1 2 3 4 5 never somet imes often almost always always 7. In between meals, I feel hungry. 1 2 3 4 5 never sometimes often almost always always

PAGE 62

54 Appendix F: Night Eating Scale 1. How often do you eat a lot of food after dinner but before bedtime? 1 2 3 4 5 less than once 1 or 2 days 3 or 4 days 5 or 6 days every day a week per week per week per week 2. About how much of your daily food intake do you consume after dinnertime? 0 1 2 3 4 0 % 1 25% 25 50% 50 75% 75 100% (none) (up to a quarter) ( about half) (more than half) (all) 3. Do you have cravings or urges to eat snacks when you wake up at night? 0 1 2 3 4 Never Sometimes About half the Usually Always time 4. Do you need to eat in order to get back to sleep when you awake at night? 0 1 2 3 4 Never Sometimes About half the Usually Always Time 5. When you get up in the middle of the night, how often do you snack? 0 1 2 3 4 Never Sometimes About half the Usually Always time 6. How often do you find yourself eating or drinking (not including water or diet drinks) throughout much of the evening (after dinner)? 1 2 3 4 5 less than once 1 or 2 days 3 or 4 days 5 or 6 days every day a week per week per week per week 7. When I get up in the middle of the nig ht, I eat a snack. 1 2 3 4 5 never sometimes often almost always always

PAGE 63

55 Appendix G: Brewer et al. (2003) Items 1. How often do you eat during the two hours before your bedtime? 1 2 3 4 rarely or never occasionally (at more than once nearly least once a week) a week everyday 2. How often do you eat between meals? 1 2 3 4 rarely or never occasionally (at more than once nearly least once a week) a week everyday 3. How often do you feel really hungry even though you have eat en adequately within the last three hours? 1 2 3 4 rarely or never occasionally (at more than once nearly least once a week) a week everyday 4. How often do you keep eating at meals even though you are not hungry anymore? 1 2 3 4 rarely or never occasionally (at more than once nearly least once a week) a week everyday

PAGE 64

56 Appendix H: Eating Expectancies Inventory Read each statement and circle the number of the response which most closely matches your level of agreement. Please respond to the items in terms of what the word eating means to you. There are no right or wrong answers. Choose only one response for each item. Do not leave any items blank. 1 2 3 4 5 6 7 completely mostly slightly neither slightly mostly completely disagree disagree disagree agree nor agree agree agree disagree 1. Eating makes me feel loved. 1 2 3 4 5 6 7 2. When I am feeling depressed or upset, eating 1 2 3 4 5 6 7 can help me take my mind off my problems. 3. Eating makes me feel out of control. 1 2 3 4 5 6 7 4. Eating fills some emotional need. 1 2 3 4 5 6 7 5. When I am feeling anxious or tense, 1 2 3 4 5 6 7 eating helps me relax. 6. I dont see eating as a pleasurable event. 1 2 3 4 5 6 7 7. Eating helps me deal with feelings of inadequacy 1 2 3 4 5 6 7 abo ut myself. 8. Eating doesnt help me deal with boredom. 1 2 3 4 5 6 7 9. When I have nothing to do, eating helps 1 2 3 4 5 6 7 relieve the boredom. 10. When I eat, I often feel I am not in charge of my life. 1 2 3 4 5 6 7

PAGE 65

57 Appe ndix H (Continued) 11. When I am feeling anxious, eating does not make 1 2 3 4 5 6 7 me feel calmer. 12. Eating serves as an emotional release. 1 2 3 4 5 6 7 13. Eating seems to decrease my level of anxiety 1 2 3 4 5 6 7 if I am feeling tense or stressed. 14. Eating is a good way to celebrate. 1 2 3 4 5 6 7 15. When I do something good, eating is a way to 1 2 3 4 5 6 7 reward myself. 16. Eating isnt useful as a reward for me. 1 2 3 4 5 6 7 17. I dont get a sense of security or safety from eating. 1 2 3 4 5 6 7 18. If I have nothing planned to do during the day, 1 2 3 4 5 6 7 eating isnt something that would help me fill the time. 19. Eating helps me think and study better. 1 2 3 4 5 6 7 20. Eating is fun and enjoyable. 1 2 3 4 5 6 7 21. My eating behavior often results in a feeling 1 2 3 4 5 6 7 that I am not in control. 22. When I work hard or accomplish something, 1 2 3 4 5 6 7 eating doesnt serve as a good reward. 23. Eating is something to do when you feel bored. 1 2 3 4 5 6 7 24. Eating is a way to vent my anger. 1 2 3 4 5 6 7

PAGE 66

58 Appendix H (Continued) 25. Eating helps me avoid uncomfort able 1 2 3 4 5 6 7 social situations. 26. When I am angry at my parents, spouse or 1 2 3 4 5 6 7 friends, eating helps me get back at them. 27. When I am faced with difficult tasks, eating 1 2 3 4 5 6 7 can help me avoid doing them. 28. Eating helps me forget or block out negative 1 2 3 4 5 6 7 feelings, like depression, loneliness, or fear. 29. Eating calms me when I am feeling 1 2 3 4 5 6 7 stressed, anxious, or tense. 30. Eating c an help me bury my emotions 1 2 3 4 5 6 7 when I dont want to feel them. 31. Eating helps me work better. 1 2 3 4 5 6 7 32. Eating helps me cope with negative emotions. 1 2 3 4 5 6 7 33. Eating does not make me feel out of control. 1 2 3 4 5 6 7 34. Eating helps me deal with sadness or 1 2 3 4 5 6 7 emotional pain.

PAGE 67

59 Appendix I: Self Reported Cue Reactivity Questionnaire 1. If you walk past a bakery, how long do you desire to eat something? 1 2 3 4 5 at most, a about 30 about a several five minutes few seconds seconds minute minutes or more 2. If you see or smell something delicious, how strong is your desire to eat it? 1 2 3 4 5 not strong somewhat moderately very strong extremely at all strong strong strong 3. When your favorite foods are around the house, how difficult is it to resist eating them? 1 2 3 4 5 not at all somewhat moderately very difficult extremely difficult to difficult to difficult to to resist difficult to r esist resist resist resist 4. If you see others eating, how strong is your desire to eat too? 1 2 3 4 5 not strong somewhat moderately very strong extremely at all strong strong strong 5. If you walk past a snack bar or caf, how difficult is it to resist your desire to eat something? 1 2 3 4 5 not at all somewhat moderately very difficult extremely difficult to difficult to difficult to to resist difficult to resist resist resist resist 6. If you walk past a bakery, how strong is your desire to eat something? 1 2 3 4 5 not strong somewhat moderately very strong extremely at all strong strong strong

PAGE 68

60 Appendix I (Continued) 7. When appealing food comes up in conversation or in something you read, how long do you desire to eat so mething? 1 2 3 4 5 at most, a about 30 about a several five minutes few seconds seconds minute minutes or more 8. If you see or smell something delicious, how long do you desire to eat it? 1 2 3 4 5 at most, a about 30 about a several fiv e minutes few seconds seconds minute minutes or more 9. If you walk past a bakery, how difficult is it to resist your desire to eat something? 1 2 3 4 5 not at all somewhat moderately very difficult extremely difficult to difficult to difficult t o to resist difficult to resist resist resist resist 10. When appealing food comes up in conversation or in something you read, how strong is your desire to eat something? 1 2 3 4 5 not strong somewhat moderately very strong extremely at all stro ng strong strong 11. If you walk past a snack bar or caf, how strong is your desire to eat something? 1 2 3 4 5 not strong somewhat moderately very strong extremely at all strong strong strong 12. If you see or smell something delicious, how diffi cult is it to resist your desire to eat it? 1 2 3 4 5 not at all somewhat moderately very difficult extremely difficult to difficult to difficult to to resist difficult to resist resist resist resist

PAGE 69

61 Appendix I (Continued) 13. When your favorite foods are around the house, how strong is your desire to eat them? 1 2 3 4 5 not strong somewhat moderately very strong extremely at all strong strong strong 14. If you see others eating, how long do you desire to eat something? 1 2 3 4 5 at most, a about 30 about a several five minutes few seconds seconds minute minutes or more 15. When appealing food comes up in conversation or in something you read, how difficult is it to resist your desire to eat something? 1 2 3 4 5 not at all so mewhat moderately very difficult extremely difficult to difficult to difficult to to resist difficult to resist resist resist resist 16. If you walk past a snack bar or caf, how long do you desire to eat something? 1 2 3 4 5 at most, a about 30 about a several five minutes few seconds seconds minute minutes or more 17. If you see others eating, how difficult is it to ignore your desire to eat something? 1 2 3 4 5 not at all somewhat moderately very difficult extremely difficult to diff icult to difficult to to resist difficult to resist resist resist resist 18. When your favorite foods are in the house, how long do you desire to eat something? 1 2 3 4 5 at most, a about 30 about a several five minutes few seconds seconds mi nute minutes or more

PAGE 70

62 Appendix J: Demographic Information Please provide the following information accurately and honestly. It is very important that the information is correct. Please remember that this study is anonymous and your name will not appear anywhere on these forms. Age: ______ Year in school: _____ Freshman _____ Sophomore _____ Junior ______ Senior Major: ____________________________ Marital Status: _____ Single _____ Casually dating _____ Serious dating relationship _____ L iving with a partner _____ Engaged _____ Married _____ Separated _____ Divorced _____ Widow or widower _____ Other (Please specify: ___________________________)

PAGE 71

63 Appendix J (Continued) Are you currently dieting? Yes No Which diet plan are you following? ____ Atkins _____ Low carb _____ Low fat _____ Counting calories _____ South Beach _____ Jenny Craig _____ Weight Watchers _____ Other (Please specify: _____________________________________) Race/Ethnicity (circle only one): African American Non Hispanic Caucasian Non Hispanic Hispanic Other Weight in pounds: ___________ Height: _____________

PAGE 72

About the Author Tovah Yanover earned a degree in Scholars Electives: French and Psychology from the University of Western Ontario in June 2003. The followin g September, she began her graduate studies in Clinical Psychology at the University of South Florida as the recipient of a Presidential Doctoral Scholarship. Tovah has twice attended the annual meeting of the Society of Behavioral Medicine and was the fir st author on several posters presented at these meetings. She also has several manuscripts in preparation based on the data from these posters. In her spare time, Tovah enjoys spending time with her pet rabbit, Edgar. She loves to cook and she is an avid r ecipe collector. She also loves to travel, especially to New York City and to any European destination.