PSYCHOLOGICAL PATHWAYS INFLUENCING BODY FAT CHANGES AMONG RESTRAINED EATERS: STRESS, EATING, AND EXERCISE BEHAVIORS LAI ZHAOXIU B.Soc.Sci.(Hons), NUS A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF SOCIAL SCIENCES DEPARTMENT OF PSYCHOLOGY NATIONAL UNIVERSITY OF SINGAPORE 2011 i Acknowledgements This research was funded by the National University of Singapore (WBS No.: R-581000-069-112). I would like to give my heartfelt thanks to Dr Why Yong Peng, who has been a very encouraging and patient supervisor. If not for his expert guidance, this study would not have been possible. I am grateful to A/P Fabian Lim, Professor David Koh, Ms Vivian Ng, Ms Sharon Sng, Ms Margaret Yap, Mr Leong Kah Wai, and Dr Low Yen Ling for sharing with me their expertise and for assisting me with various technical aspects of this study. I would also like to thank Ms Cinda Ong, Mr Raymond Peh, and Ms Salome Antonette Rebello for all their assistance with the Food Intake Assessment program. I would also like to thank Deborah, Jiaoyu, Eleanor, Clara, Dorothy, Dania, Wan Ling, and Shi Ting for their efforts during the tedious data collection and data processing. Many thanks to Oliver, Manisha, and Yongzhi for their help with other aspects of the study, such as helping me make sense of my incoherent ideas. And last but not the least, To Mum: Thank you for your tender loving care, and of course, for the home-cooked meals. To Yj: Thank you for being my confidant, counselor, and cheerleader. And also, for scolding me—because of that, I managed to finish my Discussion. ii Table of Contents Acknowledgements i Table of Contents ii Summary iv List of Tables v List of Figures vi Chapter 1: Introduction 1.1. Obesity: A pressing concern 1 1.2. Body fat gain: Who is at risk and when? 2 1.3. Stress and body fat gain: Eating behavior as an explanation 4 1.4. Stress and body fat gain: Exercise behavior as an explanation 8 1.5. Aims and hypotheses 10 Chapter 2: Method 2.1. Participants 13 2.2. Measures 16 2.3. Design 20 2.4. Procedure 20 2.5. Data analysis 22 Chapter 3: Results 3.1. Manipulation checks 24 3.2. MLM analyses 25 iii Chapter 4: Discussion 4.1. Summary of findings 35 4.2. Body fat gain: Restrained eaters may be at risk during stress 35 4.3. Stress and body fat gain: Eating behavior is not an explanation 36 4.4. Stress and body fat gain: Exercise behavior may be an explanation 39 4.5. Stress and body fat gain: Irregular exercise and weight cycling as another explanation 42 4.6. Limitations and improvements 43 4.7. Future directions 45 4.8. Conclusion 45 References 46 iv Summary A prospective study was conducted to test eating (energy intake) and exercise behaviors (energy expenditure) as explanations of stress-induced body fat gain among high restrained eaters (dieters), a group identified as being vulnerable to obesity. In addition, this study attempted to replicate stress-induced overeating among high restrained eaters. Restrained eating was also examined as a moderator of the relationship between stress and exercise behavior. Participants were 24 female high restrained eaters and 24 female low restrained eaters with a mean age of 19.19 years (SD = 0.67). Psychological stress, body fat, energy intake, and weekly exercise were measured at three time points: the low stress baseline, high stress pre-examination, and low stress recovery. High restrained eaters showed an increase in body fat during high stress while low restrained eaters showed no significant change. The increase in body fat was not due to energy intake but was possibly due to a decrease in energy expenditure, as inferred from decreased exercise levels. Previous findings of stress-induced overeating among high restrained eaters were not replicated. Restrained eating was a significant moderator of the relationship between stress and exercise behavior. High restrained eaters showed a decrease in exercise levels during high stress and an increase in exercise levels during low stress. On the other hand, low restrained eaters showed no significant changes in exercise levels. In conclusion, obesity prevention programs for restrained eaters should not only focus on regulating energy intake but should also target physical exercise. v List of Tables Table 1 Means and Standard Deviations of Dependent Variables for Total Sample and by Restrained Eating and Time (N = 48) 16 Table 2 MLM Testing Differences Between High Restrained and Low Restrained Eaters’ BF% Trends (N = 48) 27 Table 3 MLM Testing Differences Between High Restrained and Low Restrained Eaters’ Energy Intake Trends (N = 48) 30 Table 4 MLM Testing Differences Between High Restrained and Low Restrained Eaters’ Weekly Exercise Trends (N = 48) 33 vi List of Figures Figure 1 Psychological pathways of body fat investigated in this study. 2 Figure 2 Responses throughout recruitment process. 14 Figure 3 Changes in perceived stress levels across time points. Error bars represent standard errors. 25 Figure 4 Changes in body fat percentage across stress levels by restrained eating group. ―*‖ and ―**‖ indicate that adjacent points are significantly different, p < .05 and p < .01 respectively. 28 Figure 5 Changes in weekly exercise across stress levels by restrained eating group. ―*‖ and ―**‖ indicate that adjacent points are significantly different, p < .05 and p < .01 respectively. 34 1 Chapter 1 Introduction 1.1. Obesity: A Pressing Concern Obesity refers to the condition of having a high amount of body fat that can threaten one’s health (World Health Organization, 2011a). Many studies have uncovered links between excessive body fat and a number of diseases including type II diabetes, cardiovascular diseases, some cancers, high blood pressure, osteoarthritis, and chronic back pain (Field et al., 2001; Guh et al., 2009). For example, an obese individual has seven times greater risk of developing type II diabetes compared to a normal-weight individual (Abdullah, Peeters, de Courten, & Stoelwinder, 2010). Strazzullo and colleagues (2010) have found that obese individuals were 1.50 times more likely to get ischemic strokes. Around the world, more and more people are becoming obese. In several developed countries, the rates of obesity have risen to alarming levels. The estimated prevalence rates among those aged 15 years and older range from 8% to 25% in Germany, Italy, France, and UK (World Health Organization, 2011b). In USA, prevalence rates have reached epidemic levels, with 50% of the population being obese (World Health Organization, 2011b). Even in developing countries, the rates of obesity are increasing (World Health Organization, 2011a). The rising rates of obesity is a pressing concern, given the severe health implications and substantial medical spending associated with obesity (Finkelstein, Trogdon, Cohen, & Dietz, 2009). It is important to prevent this public health problem from escalating further by developing primary interventions. To devise effective 2 interventions, the pathways influencing the development of obesity must first be understood. For example, psychological pathways influencing body fat gain can be examined. As shown in Figure 1, this study investigates the effects of psychological stress on body fat via the pathways of eating (energy intake) and exercise behaviors (energy expenditure), among high restrained eaters (dieters) and low restrained eaters (non-dieters). In other words, this study examines whether health behaviors that regulate energy balance can explain stress-induced changes in body fat among high restrained eaters and low restrained eaters (if such changes were observed). At the same time, the study also examines if stress-induced changes in eating and exercise behaviors differed across the two groups. The primary purpose is to delineate the psychological pathways influencing body fat gain among high restrained eaters. Eating behavior (energy intake) Restrained eating Body fat × Psychological stress Exercise behavior (energy expenditure) Figure 1. Psychological pathways of body fat investigated in this study. 1.2. Body Fat Gain: Who is at Risk and When? Some people may be more prone to putting on weight than others. Restrained eaters are one such group that has been identified as being particularly vulnerable (van Strien, Engels, & van Staveren, 2006). Restrained eating, otherwise known as ―chronic 3 dieting‖, is the cognitive control of food intake to maintain or achieve an ideal body weight (Herman & Polivy, 1980). This construct originated from interest in the etiology of obesity (Greeno & Wing, 1994; Herman & Mack, 1975; Herman & Polivy, 1975). Paradoxically, restrained eating may promote weight gain. It has been positively associated with weight and body fat percentage in cross-sectional (Beiseigel & NickolsRichardson, 2004) and prospective studies (Drapeau et al., 2003; Lowe et al., 2006; Vella-Zarb & Elgar, 2009). On the other hand, there are also studies showing no significant relationships between restrained eating and obesity (de Lauzon-Guillain et al., 2006; Hays et al., 2002). The equivocal findings may point to a complex relationship between restrained eating and obesity. Perhaps, restrained eaters are vulnerable to body fat gain only under certain circumstances, such as during psychological stress. Both cross-sectional and prospective studies have demonstrated positive links between stress-related factors such as socioeconomic status and job strain with indicators of obesity such as body mass index and waist-to-hip ratio (Economos, Hildebrandt, & Hyatt, 2008; Kouvonen, Kivimäki, Cox, Cox, & Vahtera, 2005; Roberts, Troop, Connan, Treasure, & Campbell, 2007; Rosmond & Björntorp, 1999; Rosmond, Lapidus, & Björntorp, 1996). Psychological stress may increase the vulnerability of restrained eaters to body fat gain by disrupting their health behaviors (Steptoe, 1991). For example, stress may promote overeating among restrained eaters (Greeno & Wing, 1994) or it may cause a reduction in their frequency of physical exercise (e.g., Ng & Jeffery, 2003). Given that previous studies done on restrained eating and obesity have not considered the influence of stress, it might 4 be informative to examine changes in the body fat of restrained eaters across varying stress levels. 1.3. Stress and Body Fat Gain: Eating Behavior as an Explanation One psychological pathway affecting a restrained eater’s body fat during stress could be stress-induced changes in eating behavior and a resultant increase in energy intake. For example, stress has been associated with higher dietary fat intake (Ng & Jeffery, 2003), a greater frequency of fast food consumption (Steptoe, Lipsey, & Wardle, 1998), a greater frequency of snacks consumption (Conner, Fitter, & Fletcher, 1999), and increased energy intake (Michaud et al., 1990; Wardle, Steptoe, Oliver, & Lipsey, 2000). Eating more when under psychological stress may promote greater weight gain (Epel et al., 2004). This may be attributed to a positive energy balance that encourages body fat accumulation (Nieuwenhuizen & Rutters, 2008). 1.3.1. Literature on stress-induced eating among restrained eaters The phenomenon of stress-induced eating among restrained eaters has attracted considerable attention from researchers. Much research has been done to test the hypothesis that restrained eaters would eat more under stress than when not under stress while unrestrained eaters would show minimal changes in their food intake. Such studies have typically been done on female participants because larger proportions of restrained eaters are found among female participants than male participants (Rand & Kuldau, 1991). In an experimental study by Rutledge and Linden (1998), participants were first exposed to stress from cognitive tasks and then presented with food. Restrained eaters who experienced high negative affect ate more cookies and crackers than restrained 5 eaters who experienced low negative affect. In contrast, unrestrained eaters did not show significant changes in the amount eaten as a function of negative affect. Other experimental studies have also found comparable results (Heatherton, Herman, & Polivy, 1991; Polivy & Herman, 1999; Polivy, Herman, & McFarlane, 1994; Wallis & Heatherington, 2004). On the other hand, there have also been studies that did not uncover any stress-induced overeating among restrained eaters (Herman & Polivy, 1975; Oliver, Wardle, & Gibson, 2000). Apart from experimental studies, there have also been a few naturalistic studies done to test the hypothesis. Such studies usually assess natural stressors such as examinations and daily hassles and measure eating behavior using dietary recalls. Both supporting (O'Connor, Jones, Conner, McMillan, & Ferguson, 2008; Wardle et al., 2000) and contradictory evidence have been uncovered (Conner et al., 1999; Pollard, Steptoe, Canaan, Davies, & Wardle, 1995). To sum up, the prediction that restrained eaters will show disinhibited eating behavior during stressful times has been supported by several studies. However, a handful of studies have also shown contradictory evidence. Thus, the moderating role of restrained eating in stress-induced food intake still warrants examination. 1.3.2. Stress-induced eating: Depleted self-regulatory resources as an explanation There are a few possible explanations for restrained eaters’ stress-induced overeating. One is that restrained eaters eat more to cope with stress (Polivy et al., 1994). Specifically, eating may help to reduce anxiety from the stressor (Kaplan & Kaplan, 1957), it may help to distract from the stressor (Polivy & Herman, 1999), or it may help 6 mask the actual, more uncontrollable source of the distress (Herman & Polivy, 1988). However, studies testing these coping explanations have failed to find any convincing support for them (Herman & Polivy, 1975; Polivy & Herman, 1999; Polivy et al., 1994). Another explanation is that restrained eaters become especially sensitive to food-relevant cues in the environment when they are distressed, which makes them susceptible to overeating (Schachter, 1971; Slochower, 1976; Slochower & Kaplan, 1980). However, there is also little empirical support for this explanation (Polivy et al., 1994). On the other hand, the self-regulation literature offers a compelling explanation of restrained eaters’ stress-induced overeating. Dieting is a behavior that entails selfregulation (Baumeister, Heatherton, & Tice, 1994; Schmeichel & Baumeister, 2004). When a restrained eater is under stress, the exertion of resources to cope with the source of the stress might impair the self-regulation of other activities such as eating behavior (Baumeister & Heatherton, 1996; Baumeister et al., 1994). This may lead to overeating. Self-regulation refers to the deliberate processes that bring about changes in a person’s internal state or external behavior so that they are in line with the person’s goals (Baumeister, Vohs, & Tice, 2007; Carver, 2001). The implementation of goal-directed behavior involves the ―overriding‖ of impulses and habitual responses (Baumeister et al., 1994, p. 7). For example, a person with the goal of being slim may suppress the impulses to eat chocolates. This ―overriding‖ of impulses requires self-regulatory resources (Baumeister et al., 1994). People are not always successful in regulating themselves. For example, a person with the goal of being slim may sometimes give in to temptations to eat chocolates or to give excuses to not exercise. Baumeister and colleagues proposed that one possible cause 7 of self-regulation failure is the depletion of self-regulatory resources (Baumeister & Heatherton, 1996; Baumeister et al., 1994; Baumeister et al., 2007). In such a situation, the person becomes unable to control his or her impulses, leading to an inability to regulate behavior. Muraven and Baumeister (2000) postulate that all self-regulation tasks tap on the same pool of resources. In other words, using the resources to override impulses to play so as to stay focused on studying will leave less available for controlling one’s diet. Self-regulatory resources can also get depleted (Muraven & Baumeister, 2000). For example, after exerting self-regulation resources in studying, there may not be enough resources left for initiating and persisting in exercise. When self-regulatory resources get depleted, subsequent behaviors that require self-regulation will have a greater chance of failure (Muraven & Baumeister, 2000). These assumptions have been tested and confirmed in a number of studies (Muraven & Baumeister, 2000). Coping with psychological stress may deplete a person’s self-regulatory resources and lead to greater risks of failure in subsequent self-regulation tasks. Coping with stress typically includes trying to override negative emotions and trying to stay focused on dealing with the stressor (see review by Muraven & Baumeister, 2000). Therefore someone who is stressed may suffer from depleted resources which might in turn cause the regulation of other behaviors to fail more often. For example, a prospective study found that stress from financial difficulties was associated with a higher probability of smoking relapses (Siahpush & Carlin, 2006). Several other studies have also linked stress with a breakdown of behaviors requiring self-regulation such as abstinence from substance use (Hodgins, el-Guebaly, & Armstrong, 1995) and engagement in exercise (Ng & Jeffery, 2003; Sonnentag & Jelden, 2009). 8 Dieting is a self-regulation task that utilizes self-regulatory resources because it involves overriding impulses to eat fattening (but desirable) food (Baumeister et al., 1994; Schmeichel & Baumeister, 2004). Therefore, it is likely that dieters would show disinhibited eating under conditions of stress during which self-regulatory resources are in short supply (Kahan, Polivy, & Herman, 2003; Vohs & Heatherton, 2000). Findings from studies done on the moderating effect of restrained eating on stress-induced eating actually mirror findings from the self-regulation literature (e.g., O'Connor et al., 2008; Polivy et al., 1994; Wallis & Heatherington, 2004), suggesting that depleted selfregulatory resources may be at work. In summary, the existing literature points to the depletion of self-regulatory resources as the most likely explanation for restrained eaters’ tendency to overeat under stress. The tendency to overeat may in turn account for restrained eaters’ body fat gain under stress. 1.4. Stress and Body Fat Gain: Exercise Behavior as an Explanation Another psychological pathway that might account for the body fat gain of restrained eaters under stress is a reduction in physical exercise. Apart from increasing energy intake, psychological stress may affect the body fat of restrained eaters by simultaneously reducing energy expenditure via physical exercise. This may result in an overall energy surplus that can contribute to body fat gain (Tremblay & Therrien, 2006). Studies have found that high stress predicted lower frequency of exercise (Ng & Jeffery, 2003; Rosmond & Björntorp, 1999; Steptoe, Wardle, Pollard, Canaan, & Davies, 1996) and poorer adherence to an exercise program (Oman & King, 2000). 9 1.4.1. Rationale for examining exercise behavior: A new research area Investigating stress-induced changes in exercise behavior as a function of restrained eating not only allows a potential psychological pathway of stress-induced body fat gain to be examined, it also allows a new research area to be explored. Although there is an accumulated literature on the role of restrained eating as a moderator in stressinduced eating, no studies have examined the moderating influence of restrained eating in the relationship between stress and exercise behavior. Exercise behavior does not appear to be immediately relevant to the construct of restrained eating. But it is likely that restrained eaters will be concerned about having sufficient amounts of exercise on top of dieting, given their goal of maintaining or achieving an ideal body weight (Herman & Polivy, 1980). In fact, there is preliminary evidence showing that restrained eaters exercise more as compared to unrestrained eaters (McLean & Barr, 2003; Mclean, Barr, & Prior, 2001), suggesting that exercise may also be used as a weight control strategy. Apart from being another weight control strategy, exercise is also a behavior that taps into the limited self-regulatory resources of a restrained eater. Resources are needed to initiate exercise against impulses to do other more pleasurable leisure activities and to persist in exercise against impulses to rest (Baumeister et al., 1994; Sonnentag & Jelden, 2009). Evidence showing that exercise requires self-regulatory resources comes from Sonnentag and Jelden (2009), who found that a reduction in self-regulatory resources predicted less time spent in sports activities. Under the conditions of stress and a consequent depletion of self-regulatory resources, the frequency of exercise behavior may decrease. If high restrained eaters’ 10 resources were depleted by both stress and dieting, exercise levels may decrease by a significant extent. This study is interested to examine how stress-induced changes in exercise behavior may differ across high restrained eaters and low restrained eaters. 1.5. Aims and Hypotheses The current study investigates stress-induced changes in the eating and exercising behaviors of high restrained eaters and low restrained eaters and how these changes may influence the body fat of these individuals. Female undergraduates were measured over three time points corresponding to the low stress baseline, high stress pre-examination, and low stress recovery. 1.5.1. Aims The primary aim is to investigate the psychological pathways influencing body fat gain among high restrained eaters, a group that is vulnerable to obesity. A secondary aim is to examine whether high restrained eaters would show an increase in energy intake during stressful times to replicate past studies done on stress-induced eating. Another secondary aim is to examine whether high restrained eaters and low restrained eaters differed in the pattern of changes in exercise behavior across varying stress levels. 1.5.2. Hypotheses for energy intake High restrained eaters are hypothesized to increase their energy intake during the high stress pre-examination period compared to the low stress baseline. Energy intake is expected to decrease during the low stress recovery period. In other words, an inverted-U quadratic trend is predicted for high restrained eaters’ energy intake. During the high stress period, self-regulatory resources are likely to be directed to the higher priority tasks of emotional regulation (Tice, 2009) and studying, reducing their availability for other 11 behaviors like dieting and exercising. Therefore, the depleted supply of self-regulatory resources may increase the likelihood of disinhibited eating (Kahan et al., 2003; Vohs & Heatherton, 2000). In contrast, low restrained eaters are not expected to show changes in energy intake across the three stress levels. Since their eating behavior is not subjected to the same demands of self-regulation experienced by the high restrained eaters, their energy intake is unlikely to be affected by stress and the availability of self-regulatory resources (Muraven & Baumeister, 2000). 1.5.3. Hypotheses for exercise behavior Hypotheses regarding the differential effects of stress on high and low restrained eaters’ exercise levels are more speculative, in view of the paucity of theories and research in this area. Similar to dieting, the exercise behavior of high restrained eaters is likely to demand self-regulatory resources. Therefore, high restrained eaters’ exercise levels during the high stress period are expected to show a drop from baseline levels due to the depletion of resources. Exercise levels are expected to increase during the low stress recovery period. Unlike the high restrained eaters, low restrained eaters may be less likely to use exercise behavior as a weight control strategy. Therefore, it is possible that this would not demand as much self-regulatory resources as the goal-directed exercise behavior of high restrained eaters. On this basis, low restrained eaters are not expected to show significant changes in their exercise behavior across varying stress levels. This is in line with Muraven and Baumeister’s (2000) postulation that behaviors that do not require much self-regulatory resources will not be affected by a depletion of resources. 12 1.5.4. Hypotheses for body fat Based on the predictions that high restrained eaters would show an increase in energy intake and a decrease in energy expenditure when under stress, they are expected to demonstrate an increase in body fat between the baseline and high stress preexamination period. Body fat is expected to decrease during low stress recovery in response to lower levels of energy intake and higher levels of energy expenditure. For low restrained eaters, the hypothesized consistency of their energy intake and energy expenditure across varying stress levels is expected to translate into minimal body fat changes. 13 Chapter 2 Method 2.1. Participants The detailed recruitment process is represented by Figure 2. Prior to the study, 305 Chinese female first year undergraduates from the National University of Singapore completed the Dutch Eating Behavior Questionnaire Restrained Eating subscale (DEBQR; van Strien, Frijters, Bergers, & Defares, 1986) and the Beck Depression Inventory (BDI; Beck, Rush, Shaw, & Emery, 1979). Only female students of Chinese ethnicity were sampled because of considerations of ethnic differences in diet. One hundred and fourteen participants (37.38 %) were identified as scoring one standard deviation above or one standard deviation below the mean on the DEBQ-R (M = 2.45, SD = 0.83). Of these participants, those who have indicated that they were vegetarians or were on special diets were excluded. Those with scores of 16 and above on the BDI, which indicated that they had moderate and severe depression symptoms (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961), were also excluded. The remaining participants were contacted by phone and screened for other physical or psychological conditions that might affect their eating behavior, exercise behavior, and body fat. Specifically, they were asked if they had conditions such as eating disorders, endocrine disorders, and metabolic disorders. Those without these conditions were invited to a briefing. Five participants were ineligible for the briefing because they were either on steroid medication or had gum disease. These two exclusion criteria are for another study’s procedure of saliva sampling to perform cortisol analyses. 14 305 completed DEBQ-R and BDI 114 met criteria for DEBQ-R 10 non-contactable 25 refused screening 96 to be contacted for phone screening 86 contacted 61 screened 6 declined participation 4 vegetarians/on special diets 14 scored 16 and above on BDI 3 on steroid medication 2 with gum disease 56 eligible and attended briefing 50 recruited 48 in final sample 42% of those who met DEBQ-R criteria 1 dropped out of study 1 with missing data Figure 2. Responses throughout recruitment process. Out of 56 potential participants who attended the briefing, 50 Chinese female undergraduates participated in this study for course credits. One participant had dropped out of the study and was excluded from data analyses. Another participant had missing data and was also excluded. Both were low restrained eaters. The remaining 48 participants had a mean age of 19.19 years (SD = 0.67) and a mean BMI of 20.74 kg/m2 (SD = 3.25). Half of the participants were categorized as high restrained eaters and the other half were categorized as low restrained eaters. The mean age of high restrained 15 eaters was 19.13 years (SD = 0.45) and the mean age of low restrained eaters was 19.25 years (SD = 0.85). High restrained eaters had a restrained eating score of 3.83 (SD = 0.37) while low restrained eaters had a restrained eating score of 1.34 (SD = 0.18). As seen from Table 1, the sample’s BF% was in the ―below average‖ health category, according to the recommendations for females between 20 to 29 years old (American College of Sports Medicine, 2008; p. 59). The energy intake of participants appears to be below the daily recommendation of 2,000 kcal for adults (Health Promotion Board, 2010). One sample t-tests showed that high restrained eaters’ energy intake at each of the three time points was significantly lower than 2,000 kcal, t(23) = -3.58, p = .002; t(23) = 9.12, p < .001; t(23) = -6.96, p < .001 respectively. For low restrained eaters, energy intake at the first two time points did not differ from 2,000 kcal, t(23) = -1.18, p = .251 and t(23) = -1.43; p = .166 respectively, while energy intake at the last time point was significantly different from 2,000 kcal, t(23) = -2.96; p = .007. 16 Table 1 Means and Standard Deviations of Dependent Variables for Total Sample and by Restrained Eating and Time (N = 48) Total High restrained eaters Low restrained eaters Time 0a Time 1b Time 2c Time 0 Time 1 Time 2 M (SD) M (SD) M (SD) M (SD) M (SD) M (SD) 32.44 (3.73) 33.00 (3.81) 32.31 (3.88) 29.18 (3.69) 29.33 (3.72) 29.73 (3.81) Energy 1636.74 1652.06 intake (400.95) (476.41) (kcal/day) 1443.06 (299.16) 1400.39 (422.21) 1854.92 (603.85) 1786.58 (730.94) 1683.42 (523.93) Weekly exercise 20.83 (15.76) 31.04 (24.48) 29.42 (21.60) 30.92 (22.71) 32.08 (22.02) Variable M (SD) BF% 31.00 (4.02) 28.53 (19.95) 26.92 (21.10) Note. BF% = body fat percentage. aBaseline. bPre-examination. cRecovery. 2.2. Measures 2.2.1. Beck Depression Inventory (BDI) The BDI (Beck et al., 1979) was used to identify and exclude participants with moderate and severe depression symptoms. In each of the 21 items, participants selected one statement out of four (e.g., ―I do not feel sad‖, ―I feel sad‖, ―I am sad all the time and I can’t snap out of it‖, ―I am so sad and unhappy that I can’t stand it‖) that best described how they felt in the past one week. Each statement corresponds to a score ranging from 0 to 3. Studies have shown that the scale possesses good reliability and validity in nonclinical populations (Beck & Speer, 1988; Lightfoot & Oliver, 1985). Cronbach’s alpha for this sample was .86. The scores corresponding to each selected statement were summed up to reflect the severity of depression symptoms. 17 2.2.2. Dutch Eating Behavior Questionnaire Restrained Eating scale (DEBQR) The DEBQ-R (van Strien et al., 1986) was administered to identify participants who were high or low in restrained eating. In each of the 10 items, participants rated how frequently they exhibited a restrained eating habit on a scale extending 1 (never), 2 (seldom), 3 (sometimes), 4 (often), 5 (very often). One example of an item is ―do you try to eat less at mealtimes than you would like to eat?‖ The scale had a Cronbach’s alpha of .91. Previous studies have demonstrated its adequate reliability and validity (Laessle, Reinhard, Kotthaus, & Pirke, 1989; van Strien et al., 1986). Participants’ scores on the DEBQ-R items were averaged and used as an index of restrained eating. 2.2.3. Perceived Stress Scale (PSS) The 10-item PSS measures the overall level of stress respondents experienced with reference to the last one month (Cohen, Kamarck, & Mermelstein, 1983). In this study, the time frame was modified to one week. Participants rated how frequently they felt that things were out of control, unpredictable, and overwhelming on a scale ranging 0 (never), 1 (almost never), 2 (sometimes), 3 (fairly often), 4 (very often). For example, one item asks ―in the last week, how often have you found that you could not cope with all the things that you had to do?‖ Cronbach’s alpha of the scale ranged from .82 to .87 across repeated assessments and averaged .85. The PSS has been shown to have good reliability and validity (Cohen et al., 1983). The level of perceived stress was quantified by taking the sum of the PSS items. 18 2.2.4. Three-day food diary The food diary gathered participants’ dietary information over two weekdays and one weekend day to estimate the amount of energy intake. This assessment was developed by Low (2009) for the GUSTO birth cohort study. An experimenter went through a briefing session with each participant to train her on how to make accurate dietary records. Participants had to write down (a) the name of food or drink consumed, (b) other details that were relevant (e.g., part of meat, type of vegetable, type of cooking method, brand name of packaged food products), and (c) the amount eaten. The food diary contains guidelines on making proper dietary records. There are also photographs of standard servings of food and household measurements which participants used to estimate the amounts of food and drinks they had consumed. This method of dietary assessment has been considered to be the gold standard (Thompson & Subar, 2001). The food diary records were analyzed using the computer program Food Intake Assessment to determine the amount of energy intake averaged across the three days. This program was developed by the National University of Singapore’s Centre for Molecular Epidemiology (2010) and comprises data from various sources such as laboratory analysis of local food and overseas databases e.g., USDA National Nutrient Database for Standard Reference. 2.2.5. Godin’s Leisure Time Exercise Questionnaire (GLTEQ) The GLTEQ (Godin & Shephard, 1985) was used as an index of participants’ level of weekly exercise. Participants responded to three questions asking them how many times they engaged in strenuous (e.g., running), moderate (e.g., fast walking), and mild exercise (e.g., easy walking) for at least 15 minutes in the previous one week. The 19 GLTEQ had adequate two-month test-retest reliability of .87 between baseline and preexamination and .79 between pre-examination and recovery, with an average of .83. The scale has been shown to have adequate reliability and validity (Godin & Shephard, 1985; Jacobs, Ainsworth, Hartman, & Leon, 1993). Participants’ responses were summed up to represent weekly exercise after assigning the metabolic equivalent of task weights, 9, 5, and 3 to each of the three intensities of exercise behavior (Godin & Shephard, 1985). 2.2.6. Skinfold (SKF) measures SKF measures, which assess the thickness of subcutaneous fat, were taken in order to estimate body fat percentage (BF%). A decision was made to use this method rather than BMI because the latter has been found to be a less valid indicator of body fat (Heyward & Wagner, 2004). Three trained female observers took measurements for the participants at the triceps, subscapular, biceps, and suprailiac SKF sites. Measurements were taken with Harpenden SKF calipers (British Indicators, UK) to the nearest 0.1 mm. At each site, measurements were made twice and the mean of each site was taken. The sum of the SKFs was entered into Durnin and Womersley’s equation for females (1974) to predict body density. This equation has been validated among Singaporean Chinese females (Deurenberg & Deurenberg-Yap, 2002). The equation (4.84/body density) - 4.37 (Heyward & Wagner, 2004, p. 9) was used to estimate the percentage of the total body mass which comprised of fat. This equation minimizes systematic errors in predicting BF% of Singaporeans compared to using other equations like Siri’s (1956). Following the guidelines of Gore and colleagues (1996), precision data was collected from 10 female volunteers before the study began. The three observers 20 measured each volunteer twice, with an interval of one week. The volunteers were measured at approximately the same time of the day across both sessions. The intraobserver technical error of measurement (TEM) ranged from 2.19% to 4.48%, with a mean of 3.26%. This is within the acceptable limit of 5% recommended by experts (Gore et al., 1996). The interobserver TEM was 4.15%, within the recommended target of 7.5% (Gore et al., 1996). 2.3. Design This study utilized a longitudinal prospective design with three time points separated by 10-week intervals. The baseline corresponded to the beginning of a new academic semester (i.e., August or January), during which students were likely to be experiencing low stress. The second time point occurred three weeks before university examinations (i.e., November or April), during which students were likely to experience relatively higher stress. The third time point, recovery, occurred in the university vacation period (i.e., January or June), during which students were likely to experience low stress. 2.4. Procedure Approval for conducting the study was obtained from the National University Singapore’s Institutional Review Board (approval no.: NUS 912). Prior to the actual study, female undergraduates completed the DEBQ-R and the BDI. Information on their height and weight was also obtained. Following the screening process, eligible participants attended a briefing session. They were told that the study was on the relationships between personality, stress, and health but were not informed about specific hypotheses. The participants who had decided to take part in the study were given detailed instructions on making records in the three-day food diary followed by a 21 practice. They were reminded not to alter their diet while keeping records and to record diet information at the time of eating. None of the participants knew their restrained eating status before the completion of the study. Participants were assigned two weekdays and one weekend day to make the food diary records. In total, three sets of three-day food diaries were obtained from each participant across the time points. Saliva samples were also taken at each time point for cortisol analyses. Neither the saliva sampling procedure nor the results of the analyses will be discussed further since they were done for the purposes of a different study. Following the food diary recording, participants returned to the laboratory for further testing within one week from the first day of the recordings. The laboratory sessions were conducted for one participant at a time. Using a standardized interview protocol, trained female interviewers clarified any missing information or problematic records with the participants for each day of the food diary records. As far as possible, each of the seven interviewers involved was assigned to the same participant throughout the study. Participants also completed the PSS and GLTEQ which were administered in a counterbalanced order. SKF measurements were then taken with participants attired in sports bra and low waist pants. They had been reminded to avoid exercising 24 hours before the laboratory session, to avoid showering just before they come, and to drink sufficient amounts of water. It has been suggested that these factors may affect SKF measures (Gore et al., 1996). SKFs at the triceps, subscapular, biceps, and suprailiac sites were measured according to the protocol recommended by the International Society for the Advancement of Kinanthropometry 22 (Marfell-Jones, Olds, Stewart, & Carter, 2006). Each observer was assigned to measure the same participant throughout the study. In total, participants attended three laboratory sessions corresponding to the three time points. At each time point, reminders were sent to participants’ mobile phones to ensure that they did not forget to make records in their food diaries and to ensure that they made the required preparations for laboratory sessions. 2.5. Data analysis All analyses were done using IBM SPSS Statistics 19 with alpha level set at .05. The high restrained eating group was coded as ―0‖ and low restrained eating group was coded as ―1‖. Multilevel linear modeling (MLM) was conducted to test key hypotheses. It is a method suitable for handling nested data structures (Tabachnick & Fidell, 2007), such as that found here where repeated measures (level 1) are nested within participants (level 2). The level 1 and level 2 models described below were fitted for all three dependent variables, BF%, energy intake, and weekly exercise. The error variance covariance matrices for level 1 and level 2 were specified to have diagonal structure and identity structure respectively. The criterion for choosing the error matrix was that it had to give small fit indices relative to other possible error matrices in all three models. The diagonal structure was chosen because it gave the best deviance statistic (-2 Restricted Log Likelihood), AIC, and BIC indices across the models. No centering was done for the level 1 and level 2 predictors because their zero values are meaningful. The zero value of time and time2 represent the low stress baseline while the zero value of restrained refers to the high restrained eating group. Restricted maximum likelihood estimation was used. 23 The level 1 model regressed the repeated measures of the dependent variable on time and time2 so as to examine if the dependent variable demonstrated a linear and quadratic trend in each participant. In other words, this modeled the expectation that stress has an impact on the dependent variable. The unexplained variance in the repeated measures was captured by rit. yit = β0i + β1i timeit + β2i time2it + rit (1) i = participant 1,…, participant 48 and t = time 0, time 1, time 2 The level 2 model was fitted to examine the study’s hypotheses about whether differences in participants’ quadratic trends for the dependent variables could be explained by restrained eating. In other words, the model tested for the moderating role of restrained eating. Hence, the level 1 intercepts and regression coefficients were regressed on restrained eating. The parameter estimates, γ01, γ11, and γ21 represented the difference in the intercepts, linear trends, and quadratic trends of the dependent variable between high restrained eaters and low restrained eaters. γ00, γ10, and γ20 represented the intercept, linear trend, and quadratic trend of the high restrained eaters respectively. Time and time2 were assumed to be fixed effects in the model because their levels were not randomly selected at the beginning of the study. The unexplained interindividual variance in the intercept was represented by µ0i. β0i = γ00 + γ01 restrainedi + µ0i (2) β1i = γ10 + γ11 restrainedi (3) β2i = γ20 + γ21 restrainedi (4) i = participant 1,…, participant 48 24 Chapter 3 Results 3.1. Manipulation Checks A 2 × 3 mixed design ANOVA with restrained eating as the between-subject factor and time as the within-subject factor was run for perceived stress. This was to check if stress levels varied in expected directions across baseline, pre-examination, and recovery. Time had a significant main effect, F(2, 92) = 3.64, p = .03, partial η2 = .07, while the interaction between restrained eating and time was not significant, F(2, 92) = 1.63, p = .201, partial η2 = .03. Restrained eating had a significant main effect, F(1, 46) = 5.10, p = .029, partial η2 = .10. On average, high restrained eaters tended to report a higher level of stress (M = 17.00) compared to low restrained eaters (M = 14.00). Contrast analysis was done in GLM to probe the significant time effect, using the contrast weights 1, -1, 0 and 0, 1, -1. The first contrast compared stress level at baseline with that during pre-examination, F(1, 47) = 3.25, p = .078, partial η2 = .065. The second contrast compared stress level during pre-examination with that during recovery, F(1, 47) = 6.45, p = .01, partial η2 = .12. As seen from Figure 3, perceived stress increased between baseline and pre-examination and then dropped to a lower level during recovery (Ms = 15.31, 16.65, 14.54 respectively). 25 18.00 Perceived Stress 17.00 16.00 15.00 14.00 13.00 Baseline Pre-examination Recovery Time Figure 3. Changes in perceived stress levels across time points. Error bars represent standard errors. 3.2. MLM Analyses 3.2.1. Analyses to test hypotheses for body fat MLM was conducted (see Equations 1 to 3 in Method) to test the predictions that high restrained eaters would show stress-induced changes in BF% while low restrained eaters would not. At the same time, this allowed the first step of mediation testing to be conducted (Baron & Kenny, 1986); that is, whether the predictor (stress) was associated with the dependent variable (BF%). Results are displayed in Table 2. High restrained eaters had a higher baseline BF% compared to low restrained eaters, indicated by the statistical significance of γ01. The estimate for γ21 was also significant, indicating that the groups differed in their quadratic trends for BF%. 26 The simple trend for each group is shown in Table 2 and is graphically displayed in Figure 3. For low restrained eaters, the parameters were estimated by recoding high restrained eaters as ―1‖ and low restrained eaters as ―0‖ before repeating the MLM analysis (Curran, Bauer, & Willoughby, 2004). The findings show that high restrained eaters had a mean baseline BF% of 32.44% and low restrained eaters had a mean baseline BF% of 29.18%. The estimate for time2 was significant for high restrained eaters but not significant for low restrained eaters. In addition, the estimate for time was significant for high restrained eaters, pointing to an overall positive trend in BF% across the time points. Three out of four of the random parameter estimates were significant, indicating that there was still unexplained intraindividual variation in BF% and interindividual variation in baseline BF%. Contrast analysis was done in GLM to probe the significant quadratic trend of high restrained eaters, using the weights 1, -1, 0 and 0, 1, -1. The first contrast compared BF% at low stress baseline with BF% at high stress pre-examination, F(1, 23) = 7.45, p = .012, partial η2 = .25. The second contrast compared BF% at high stress pre-examination with BF% at low stress recovery, F(1, 23) = 14.85, p = .001, partial η2 = .39. Hence, high restrained eaters’ BF% increased by 0.55% between baseline and pre-examination and decreased by 0.68% between pre-examination and recovery1 while low restrained eaters’ BF% did not show significant changes. 1 A third contrast was done using the weights 1, 0, -1 to see if BF% differed between low stress baseline and low stress recovery. BF% did not differ between these two time points, F(1, 23) = 0.24, p > .05, partial η2 = .01. 27 Table 2 MLM Testing Differences Between High Restrained and Low Restrained Eaters’ BF% Trends (N = 48) Parameter Estimate SE Fixed effects for interactions Intercept, γ01 -3.27** 1.11 Time, γ11 -1.14* 0.53 Time2, γ21 0.74** 0.23 Fixed effects for high restrained eaters Intercept, γ00 32.44*** 0.79 Time, γ10 1.17** 0.37 Time2, γ20 -0.62*** 0.16 Fixed effects for low restrained eaters Intercept, γ00 29.18*** 0.79 Time, γ10 0.03 0.37 Time2, γ20 0.12 0.16 Random parameters Time 0, ri0 1.05*** 0.26 Time 1, ri1 0.20 0.14 Time 2, ri2 0.61** 0.19 13.86*** 2.92 Intercept, µ0i Note. MLM = multilevel linear modeling; BF% = body fat percentage. Deviance statistic = 533.73. *p < .05. ** p < .01. ***p < .001. 28 33.5 33.0 Body Fat (%) 32.5 * ** 32.0 31.5 31.0 30.5 30.0 29.5 29.0 28.5 28.0 Low Stress, Baseline High Stress, Preexamination Low Stress, Recovery Stress Level High Restrained Eaters Low Restrained Eaters Figure 4. Changes in body fat percentage across stress by restrained eating group. ―*‖ and ―**‖ indicate that adjacent points are significantly different, p < .05 and p < .01 respectively. 3.2.2. Analyses to test hypotheses for energy intake Since high restrained eaters showed an increase in BF% in response to high stress, it was appropriate to test if this relationship could be explained by energy intake and weekly exercise (energy expenditure). Hence, MLM was conducted for energy intake (Equations 1 to 3 in Method). This also allowed testing of the hypothesized difference in the pattern of changes in energy intake across varying stress levels between high restrained eaters and low restrained eaters. 29 As seen from Table 3, there was no significant difference between the two groups in their baseline energy intake, indicated by the non-significant γ01 estimate2. The groups did not differ in their linear trends in energy intake or their quadratic trends in energy intake, as shown by the non-significant γ11 and γ21 estimates. High restrained eaters had a baseline energy intake of 1652.06 kcal/day. They did not show significant changes in energy intake across the stress levels, as indicated by the non-significant estimates associated with time and time2 for the group. Since there were no significant differences between the two groups’ trends in energy intake, it was not necessary to examine the simple trend for low restrained eaters. All the random parameter estimates for this model were significant, indicating that there was still unexplained intraindividual variance in energy intake as well as interindividual variance in baseline energy intake. 2 An independent samples t-test that was done on energy intake averaged over the three time points show that high restrained eaters consumed 276.47 kcal/day less than low restrained eaters, t(46) = -.2.52, p = .015, d = 0.73. 30 Table 3 MLM Testing Differences Between High Restrained and Low Restrained Eaters’ Energy Intake Trends (N = 48) Parameter Estimate SE Fixed effects for interactions Intercept, γ01 202.86 158.54 Time, γ11 241.23 348.19 Time2, γ21 -100.57 163.71 Fixed effects for high restrained eaters Intercept, γ00 1652.06*** 112.11 Time, γ10 -292.16 246.21 83.16 115.76 Time2, γ20 Random parameters Time 0, ri0 224409.93*** 56004.07 Time 1, ri1 228124.01*** 57160.17 Time 2, ri2 149551.54*** 42171.57 77225.89* 31201.80 Intercept, µ0i Note. MLM = multilevel linear modeling. Deviance statistic = 2132.13. *p < .05. ** p < .01. ***p < .001. 3.2.3. Analyses to test hypotheses for exercise behavior To test the hypothesis that exercise (energy expenditure) explained the stressinduced body fat gain of high restrained eaters, MLM was conducted (Equations 1 to 3 in Method). This also allowed testing of the hypothesis that high restrained eaters and low 31 restrained eaters would differ in the pattern of changes in weekly exercise across varying stress levels. The results are displayed in Table 4. High restrained eaters and low restrained eaters were similar in their baseline weekly exercise levels, as shown by the nonsignificant estimate for γ01. More importantly, the groups were different in their quadratic trends for weekly exercise, as indicated by the significant estimate for γ21. The simple trends of the two groups are presented in Table 4 and in Figure 4. The parameter estimates for low restrained eaters were obtained by recoding high restrained eaters as ―1‖ and low restrained eaters as ―0‖ before repeating the MLM analysis (Curran et al., 2004). The estimate for time2 was significant for high restrained eaters but not for low restrained eaters. The significance of the random parameters indicated that there was still unexplained variance in the repeated measurements and unexplained interindividual variance in baseline weekly exercise. Similar to what was done for BF%, contrast analysis was conducted to probe the significant quadratic trend among high restrained eaters. The first contrast compared weekly exercise during low stress baseline and weekly exercise during high stress preexamination, F(1, 23) = 6.78, p = .016, partial η2 = .23. The second contrast compared weekly exercise during high stress with weekly exercise during low stress recovery, F(1, 23) = 14.34, p = .001, partial η2 = .38. Hence, high restrained eaters showed a marked decrease in weekly exercise between low stress baseline and high stress pre-examination period and an increase during low stress recovery3. On the other hand, low restrained 3 A third contrast was done to test if weekly exercise levels differed between low stress baseline and low stress recovery. No significant difference was found, F(1, 23) = 1.52, p > .05, partial η2 = .06. 32 eaters did not demonstrate significant changes in their level of weekly exercise across varying stress levels. The MLM analyses showed that there was a significant association between Restrained Eating × Time2 and weekly exercise, establishing weekly exercise as a potential explanation for high restrained eaters’ body fat gain during high stress. Therefore, the next step was to test for a significant association between weekly exercise and BF%, after controlling for Restrained Eating × Time2. An attempt was made to test this by running an MLM analysis using Type I sums of squares, but the model did not reach convergence. 33 Table 4 MLM Testing Differences Between High Restrained and Low Restrained Eaters’ Weekly Exercise Trends (N = 48) Parameter Estimate SE Fixed effects for interactions Intercept, γ01 2.50 6.08 Time, γ11 15.90** 4.94 Time2, γ21 -8.31** 2.60 Fixed effects for high restrained eaters Intercept, γ00 Time, γ10 Time2, γ20 26.92*** 4.30 -14.23*** 3.50 8.15*** 1.84 Fixed effects for low restrained eaters Intercept, γ00 29.42*** 4.30 Time, γ10 1.67 3.50 Time2, γ20 -0.17 1.84 Random parameters Time 0, ri0 77.46* 23.13 Time 1, ri1 19.07 16.65 Time 2, ri2 170.62*** 40.07 Intercept, µ0i 366.54*** 79.58 Note. MLM = multilevel linear modeling. Deviance statistic = 1137.36. *p < .05. ** p < .01. ***p < .001. 34 34 Weekly Exercise 32 30 28 ** 26 * 24 22 20 18 Low Stress, Baseline High Stress, Preexamination Low Stress, Recovery Stress Level High Restrained Eaters Low Restrained Eaters Figure 5. Changes in weekly exercise across stress levels among by restrained eating group. ―*‖ and ―**‖ indicate that adjacent points are significantly different, p < .05 and p < .01 respectively. 35 Chapter 4 Discussion 4.1. Summary of Findings The current study was conducted with the primary aim of delineating psychological pathways that may influence the body fat gain of restrained eaters. A secondary aim is to replicate previous findings of stress-induced overeating among restrained eaters. Another secondary aim is to investigate stress-induced changes in exercise behavior as a function of restrained eating. Overall, the study’s hypotheses are partially supported. In line with expectations, high restrained eaters showed an increase in body fat between low stress baseline and high stress pre-examination and a decrease in body fat during low stress recovery. On the other hand, low restrained eaters did not show much change in body fat across varying stress levels. High restrained eaters’ gain in body fat may have been due to a decrease in exercise levels (energy expenditure) rather than an increase in energy intake; contrary to predictions, high restrained eaters were similar to low restrained eaters in their pattern of energy intake across varying stress levels and did not show significant changes in energy intake when under stress. In line with predictions, high restrained eaters showed a decrease in exercise levels in between low stress baseline and high stress pre-examination and an increase in exercise levels during low stress recovery. In contrast, low restrained eaters did not show significant changes in exercise across the different stress levels. 4.2. Body Fat Gain: Restrained Eaters may be at Risk During Stress The findings imply that high restrained eaters may be vulnerable to gaining body fat during stressful periods while low restrained eaters seem to be relatively less 36 vulnerable to gaining body fat across varying stress levels. High restrained eaters’ vulnerability to body fat gain during psychological stress is in line with the positive relationships between stress and body weight which has been uncovered before (Economos et al., 2008; Roberts et al., 2007; Rosmond & Björntorp, 1999). The finding that high restrained eaters’ body fat does change as a function of stress may be a possible explanation for the equivocal results of studies done on restrained eating and obesity which did not consider the influence of stress (de Lauzon-Guillain et al., 2006; Drapeau et al., 2003). Given that contemporary living frequently involves circumstances and events that have the potential to elicit psychological stress, restrained eaters may be at risk for developing obesity. Extrapolating from the current findings of an overall positive linear trend in body fat, restrained eaters who experience prolonged stressors and/or a large number of stressful periods may accumulate small amounts of body fat from each stressful episode that may add up to substantial gains. In summary, the findings indicate that stress may play an important role in the body fat gain of restrained eaters and may possibly encourage obesity in the long run. The subsequent sections will discuss the health behaviors that may account for the stressinduced body fat gain of restrained eaters. 4.3. Stress and Body Fat Gain: Eating Behavior is not an Explanation In this study, psychological stress did not promote excessive energy intake among high restrained eaters, contradicting past findings of positive associations between stress and energy intake (Michaud et al., 1990; Wardle et al., 2000). Hence, eating behavior is ruled out as an explanation for high restrained eaters’ increased body fat under stress. In 37 other words, the energy surplus that had facilitated high restrained eaters’ body fat accumulation was not a result of increased energy intake. An explanation for why eating behavior did not mediate the relationship between stress and body fat among high restrained eaters (i.e., why there was an absence of stress-induced overeating) is suggested in the following section. No stress-induced overeating among restrained eaters: Self-regulatory resources may have been resistant to depletion With regards to the moderating role of restrained eating in the relationship between stress and energy intake, this study did not find support for it. Previous findings of stress-induced overeating among high restrained eaters (Heatherton et al., 1991; Polivy et al., 1994; Rutledge & Linden, 1998) were not replicated. At the same time, this lack of overeating among high restrained eaters corresponds with the null results of a number of studies (Conner et al., 1999; Oliver et al., 2000; Pollard et al., 1995). The current findings seem to suggest that not all restrained eaters are vulnerable to stress-induced overeating. With regards to the low restrained eaters, the ―immunity‖ of their energy intake to the effects of stress as found by previous studies (e.g., Rutledge & Linden, 1998) was replicated. According to the self-regulation literature, the explanation for the findings among low restrained eaters is quite straightforward. Low restrained eaters’ eating behavior did not demand self-regulatory resources and hence it was not affected by the depletion of resources during examination stress (Muraven & Baumeister, 2000). On the other hand, high restrained eaters’ lack of overeating in response to stress, or more precisely, their successful dieting even when facing stress, may have been due to their self-regulatory 38 resources being resistant to depletion. Before discussing the role of self-regulatory resources further, it is necessary to point out the possibility that successful dieters may have been selected for the high restrained eaters’ group in this study due to the scale used. The nature of the items comprising of the DEBQ-R may result in the scale capturing predominantly chronic dieters with a low propensity towards overeating, as opposed to novice dieters with a high propensity towards overeating (Allison, Kalinsky, & Gorman, 1992; Ouwens, van Strien, & van der Staak, 2003). This has been cited as a reason as to why studies using the DEBQ-R have tended not to find stress-induced overeating among high restrained eaters (Ouwens et al., 2003). But why might chronic dieters have a low susceptibility towards overeating, including those that are stressinduced? The literature suggests that this may be due to the resistance of their selfregulatory resources to depletion which means a slower rate of depletion of their resources during self-regulation (Muraven & Baumeister, 2000). According to some studies (Baumeister, Gailliot, DeWall, & Oaten, 2006; Muraven, Baumeister, & Tice, 1999), one’s self-regulatory resources can become resistant to depletion in the long run if he or she repeatedly practiced self-regulation and rested following short-term depletions from practice. Chronic dieters may have a long history of alternating between dieting and disinhibited eating (Lowe, 1993), which may reflect repeated cycles of self-regulation practice and rest. Such cycles may in turn build up the resistance of their self-regulatory resources and allow them to handle more self-regulatory tasks (e.g., dieting and other non-dieting tasks) at any one time. Therefore, the high restrained eaters of this study, who may be chronic dieters with resistant self-regulatory resources, may have successfully 39 regulated themselves in both the handling of the academic stressor and the maintaining of their diets. It is also possible that stress-induced overeating did not appear to have occurred among high restrained eaters due to their underreporting of energy intake. Underreporting has been found among participants when self-report measures such as dietary records are used to assess food intake (Thompson & Subar, 2001). However, it is difficult to ascertain whether underreporting had occurred in this study since the precise energy requirements of each participant was not measured. But even if participants had underreported their energy intakes, it is reasonable to assume that the extent of underreporting had been minimized through various precautions taken (e.g., concurrent recording, reminders to participants, post-recording interview) such that any increase in energy intake would still have been detected. To sum up, the lack of stress-induced overeating among high restrained eaters is more likely to have been due to the resistance of their self-regulatory resources to depletion. Therefore, the stamina of high restrained eaters in self-regulation may have explained why their stress-induced body fat gain did not occur via excessive eating. 4.4. Stress and Body Fat Gain: Exercise Behavior may be an Explanation High restrained eaters’ exercise levels decreased in response to psychological stress, in line with past findings of negative associations between stress and exercise frequency (Ng & Jeffery, 2003; Steptoe et al., 1996). Hence exercise behavior may be a possible explanation as to why high restrained eaters showed an increase in body fat during stress though a formal test of this mediation model could not be achieved in the current study. A decrease in exercise levels may have led to a decrease in energy 40 expenditure, which could then have resulted in a positive energy balance and an accumulation of body fat (Nieuwenhuizen & Rutters, 2008). An account of why exercise behavior may have mediated the relationship between stress and body fat gain among high restrained eaters (i.e., why there was a reduction in exercise levels) is examined in the subsequent section. Restrained eaters exercised less during stress: Depleted self-regulatory resources may be an explanation Restrained eating moderated the relationship between stress and exercise; high restrained eaters exercised less during stressful times while low restrained eaters’ exercise levels were not affected by stress. These findings are novel since this study is one of the first to have gone beyond examining restrained eaters’ stress-induced eating behavior to examine the effects of stress on another equally important weight control strategy—exercise behavior. The findings imply that researchers interested in understanding the etiology of obesity among restrained eaters may need to consider the role of stress-induced reductions in exercise behavior, in addition to stress-induced overeating. Depleted self-regulatory resources may underlie high restrained eaters’ reduction in exercise levels during high stress. Despite the fact that the self-regulatory resources of high restrained eaters may have been resistant to depletion, the exertion of limited resources in both the control of diet and in the handling of the examination stressor would have reduced the availability of resources for regulating exercise behavior, resulting in a greater likelihood of failure in the regulation of exercise (Muraven & Baumeister, 2000). After all, ―one cannot regulate everything at once‖ (Baumeister & Heatherton, 1996, p. 41 3). A question follows: Why were resources diverted to the regulation of eating behavior more so than exercise behavior? One tentative explanation is that high restrained eaters prefer using dieting as a weight control strategy compared to exercising, as found by one study (Field, Manson, Taylor, Willett, & Colditz, 2004). Hence, self-regulation failure occurred in exercise behavior rather than eating behavior. Low restrained eaters in the current study exercised as much as high restrained eaters at baseline, which is contrary to previous findings (McLean & Barr, 2003; Mclean et al., 2001). Therefore, the possibility that the low restrained eaters may have been regulating their exercise behavior cannot be ruled out with certainty. If low restrained eaters’ exercise behavior demanded self-regulatory resources, why were their exercise levels not affected by stress then? One tentative explanation is that low restrained eaters’ exercise behavior was not in conflicting demands with dieting like the high restrained eaters and hence exercise levels could be maintained with the remaining self-regulatory resources which were not exerted in the handling of the stressor. In sum, there is a limit to high restrained eaters’ success in self-regulation, despite the fact that their self-regulatory resources may be resistant to depletion. While they may have been successful in maintaining their energy intake during stress, this may have led to the exhaustion of their limited supply of self-regulatory resources and a subsequent reduction in exercise levels. In turn, this may have promoted fat gain. This psychological pathway may predict the development of obesity among restrained eaters in the long run. Specifically, sustained reductions in exercise levels and energy expenditure due to experiences of prolonged stressors and/or a large number of stressful episodes may encourage substantial gains in body fat. 42 From a practitioner’s perspective, obesity prevention programs for restrained eaters should not simply focus on regulating energy intake but should also target physical exercise. Specifically, prevention programs for restrained eaters who are rather successful at dieting may need to focus on maintaining their exercise levels, especially during stress. Initially, maintaining exercise levels may not be easy given the finite pool of selfregulatory resources. One way to overcome this is to build up the resistance of selfregulatory resources to depletion through intensive self-regulation practice (Baumeister et al., 2006; Muraven et al., 1999). In time to come, restrained eaters will then be able to sustain exercise levels even during stressful times. 4.5. Stress and Body Fat Gain: Irregular Exercise and Weight Cycling as Another Explanation There is a tentative possibility of another pathway through which stress may promote obesity among restrained eaters in the long-term: irregular exercise and weight cycling which refers to recurring cycles of weight loss and weight regain (Brownell & Rodin, 1994). This study’s results reflect the occurrence of irregular exercise and weight cycling among high restrained eaters; they demonstrated fluctuating exercise levels in response to varying stress levels and possibly because of this, they also showed a single weight cycle in which body fat increased and then decreased. According to the existing literature, weight cycling that is driven by yo-yo dieting (recurring cycles of dieting and disinhibited eating) might be a risk factor for obesity (Brownell & Rodin, 1994). The upward trend in weight may be attributed to the metabolic effects of yo-yo dieting although this is a subject of some controversy (Brownell, Greenwood, Stellar, & Shrager, 1986; Brownell & Rodin, 1994). 43 The current study is unique in that it suggests that for some dieters, weight cycling may be due to ―yo-yo exercising‖ rather than yo-yo dieting. Perhaps, yo-yo exercising, similar to yo-yo dieting, may predict future weight cycling and long-term upward trends in body weight. The findings of one study suggest that irregular exercising might actually hinder future weight loss and predict weight gain, though its exact mechanisms are not known (Williams, 2008). Future prospective studies of longer duration can investigate whether body weight cycling due to stress-induced fluctuations in exercise might predict weight gain among restrained eaters. 4.6. Limitations and Improvements One limitation of the current study is the restricted sample size; the time consuming nature of the study activities may have unavoidably discouraged participation. Insufficient number of data points at the second level may have resulted in the nonconvergence of the MLM model testing exercise as a mediator. With a longer duration of data collection, it may be possible to gather a larger sample. Another limitation is the use of self-reports in the measurement of exercise levels. The GLTEQ’s reliability is dependent on the accuracy of participants’ recall, pointing to the possibility that the findings on exercise behavior may reflect inaccurate recalls. Future studies may need to consider supplementing subjective reports of exercise with objective methods of assessment such as pedometers or accelerometers (Hankinson, 2008). The use of self-report to assess energy intake comes with the drawback of underreporting. Studies comparing self-reported energy intake (including those from three-day food diaries) with energy expenditure assessed by doubly-labelled water have 44 shown that participants underreport energy intake by approximately 4% to 37% (Hill & Davies, 2001; Thompson & Subar, 2001; Trabulsi & Schoeller. 2001). The low reported energy intakes in this study points to the possibility of underreporting. Regardless, precautions taken during the study e.g., reminders to participants to make records and the post-recording interview has likely minimized underreporting, such that confidence in the validity of findings related to energy intake remains high. This study did not include measurements of energy expenditure which might be a limitation. The use of exercise levels to infer energy expenditure and to draw conclusions about overall energy balance relies on the assumption that the other components of energy expenditure (e.g., spontaneous physical activity) did not change across the time points of the study. However, the assumption may not hold given that there is a degree of intraindividual variation in the other components of energy expenditure (Donahoo, Levine, & Melanson, 2004). Hence, to ascertain that the body fat gain of restrained eaters during stress is a result of decreased energy expenditure, energy expenditure measurements are required. Future studies can include techniques like portable forms of expiratory collection open-circuit systems to assess the components of energy expenditure (Levine, 2005). Another limitation is the limited generalizability of the findings given that this study was conducted with a sample of young and highly educated Chinese female participants. Future studies should be extended to more representative samples such as community based samples of female participants of Chinese and non-Chinese ethnicity. The caveat is that the religious practices of certain groups (e.g., Muslims) may pose challenges in the assessment of dietary behaviors. 45 4.7. Future Directions One possible direction that future studies can take is to investigate the physiological pathways influencing the body fat gain of restrained eaters during stressful times. The role of the hypothalamic-pituitary-adrenal axis has featured prominently in physiological explanations of stress-induced obesity. Long-term elevations of cortisol, a stress hormone released by the axis, may facilitate the accumulation of central body fat (Björntorp, 2001). Perhaps similar physiological pathways are at work among restrained eaters, in addition to psychological pathways. 4.8. Conclusion Obesity is a preventable health condition and yet, America spends hundreds of billions of dollars treating it each year (Finkelstein et al., 2009). A better understanding of the complex pathways influencing it can contribute towards the development of effective primary interventions. The current study has identified a psychological pathway that may possibly influence the development of obesity. Stress may encourage body fat gain among some restrained eaters, possibly through reducing exercise levels (energy expenditure). Self-regulatory resources may be involved in this pathway as well. The findings suggest that interventions to prevent obesity among restrained eaters may need to intensify during stressful periods and focus on maintaining physical exercise. 46 References Abdullah, A., Peeters, A., de Courten, M., & Stoelwinder, J. (2010). The magnitude of association between overweight and obesity and the risk of diabetes: A metaanalysis of prospective cohort studies. Diabetes Research and Clinical Practice, 89, 309–319. Allison, D. B., Kalinsky, L. B., & Gorman, B. S. (1992). A comparison of the psychometric properties of three measures of dietary restraint. Psychological Assessment, 4, 391–398. American College of Sports Medicine (2008). ACSM's health-related physical fitness assessment manual. Philadelphia: Wolters Kluwer/Lippincott Williams & Wilkins. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Baumeister, R. F., Gailliot, M., DeWall, C. N., & Oaten, M. (2006). Self-regulation and personality: How interventions increase regulatory success, and how depletion moderates the effects of traits on behavior. Journal of Personality, 74, 1773– 1802. Baumeister, R. F., & Heatherton, T. F. (1996). Self-regulation failure: An overview. Psychological Inquiry, 7, 1–15. Baumeister, R. F., Heatherton, T. F., & Tice, D. M. (1994). Losing control: How and why people fail at self-regulation. San Diego, CA: Academic Press. 47 Baumeister, R. F., Vohs, K. D., & Tice, D. M. (2007). The strength model of self-control. Current Directions in Psychological Science, 16, 351–355. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression. New York, NY: Guilford Press. Beck, A. T., & Speer, R. A. (1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77– 100. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561–571. Beiseigel, J. M., & Nickols-Richardson, S. M. (2004). Cognitive eating restraint scores are associated with body fatness but not with other measures of dieting in women. Appetite, 43, 47–53. Björntorp, P. (2001). Do stress reactions cause abdominal obesity and comorbidities? Obesity Reviews, 2, 73–86. Brownell, K. D., Greenwood, M. R. C., Stellar, E., & Shrager, E. E. (1986). The effects of repeated cycles of weight loss and regain in rats. Physiology & Behavior, 38, 459–464. Brownell, K. D., & Rodin, J. (1994). Medical, metabolic, and psychological effects of weight cycling. Archives of Internal Medicine, 154, 1325–1329. Carver, C. S. (2001). Self-regulation. In A. Tesser & N. Schwarz (Eds.), Blackwell Handbook of Social Psychology: Intraindividual processes (pp. 479–498). Malden, M.A.: Blackwell. 48 Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24, 385–396. Conner, M., Fitter, M., & Fletcher, W. (1999). Stress and snacking: A diary study of daily hassles and between-meal snacking. Psychology and Health, 14, 51–63. Curran, P. J., Bauer, D. J., & Willoughby, M. T. (2004). Testing main effects and interactions in latent curve analysis. Psychological Methods, 9, 220–237. de Lauzon-Guillain, B., Basdevant, A., Romon, M., Karlsson, J., Borys, J. M., Charles, M. A., & Group, T. F. S. (2006). Is restrained eating a risk factor for weight gain in a general population? American Journal of Clinical Nutrition, 83, 132–138. Deurenberg, P., & Deurenberg-Yap, M. (2002). Validation of skinfold thickness and hand-held impedance measurements for estimation of body fat percentage among Singaporean Chinese, Malay and Indian subjects. Asia Pacific Journal of Clinical Nutrition, 11, 1–7. Donahoo, W. T., Levine, J. A., & Melanson, E. L. (2004). Variability in energy expenditure and its components. Current Opinion in Clinical Nutrition & Metabolic Care, 7, 299–605. Drapeau, V., Provencher, V., Lemieux, S., Després, J. P., Bouchard, C., & Tremblay, A. (2003). Do 6-y changes in eating behaviors predict changes in body weight? Results from the Québec Family Study. International Journal of Obesity, 27, 808–814. Durnin, J. V. G. A., & Womersley, J. (1974). Body fat assessed from total body density and its estimation from skinfold thickness: Measurements on 481 men and women aged from 16 to 72 years. British Journal of Nutrition, 32, 77–97. 49 Economos, C. D., Hildebrandt, L., & Hyatt, R. R. (2008). College freshman stress and weight change: Differences by gender. American Journal of Health Behavior, 32, 16–25. Epel, E., Jimenz, S., Brownell, K., Stroud, L., Stoney, C., & Niaura, R. (2004). Are stress eaters at risk for the metabolic syndrome? Annals of the New York Academy of Sciences, 1032, 208–210. Field, A. E., Coakley, E. H., Spadano, J. L., Laird, N., Dietz, W. H., Rimm, E., & Colditz, G. A. (2001). Impact of overweight on the risk of developing common chronic diseases during a 10-year period. Archives of Internal Medicine, 161, 1581–1586. Field, A. E., Manson, J. E., Taylor, C. B., Willett, W. C., & Colditz, G. A. (2004). Association of weight change, weight control practices, and weight cycling among women in the Nurses' Health Study II. International Journal of Obesity, 28, 1134–1142. Finkelstein, E. A., Trogdon, J. G., Cohen, J. W., & Dietz, W. (2009). Annual medical spending attributable to obesity: Payer- and service-specific estimates. Health Affairs, 28, w822–w831. Godin, G., & Shephard, R. J. (1985). A simple method to assess exercise behavior in the community. Canadian Journal of Applied Sport Sciences, 10, 141–146. Gore, C., Norton, K., Olds, T., Whittingham, N., Birchall, K., Clough, M., . . . Downie, L. (1996). Accreditation in anthropometry: An Australian model. In K. Norton & T. Olds (Eds.), Anthropometrica (pp. 395–421). Sydney: UNSW Press. 50 Greeno, C. G., & Wing, R. R. (1994). Stress-induced eating. Psychological Bulletin, 115, 444–464. Guh, D. P., Zhang, W., Bansback, N., Amarsi, Z., Birmingham, L., & Anis, A. H. (2009). The incidence of co-morbidities related to obesity and overweight: A systematic review and meta-analysis. BMC Public Health, 9. doi: 10.1186/1471-2458-9-88 Hankinson, A. (2008). Association of activity and chronic disease risk factors: Utility and limitations of objectively measured physical activity data. Journal of the American Dietetic Association, 108, 945–947. Hays, N. P., Bathalon, G. P., McCrory, M. A., Roubenoff, R., Lipman, R., & Roberts, S. B. (2002). Eating behavior correlates of adult weight gain and obesity in healthy women aged 55-65 y. American Journal of Clinical Nutrition, 75, 476–483. Health Promotion Board (2010). Nutrition. Retrieved from http://www.hpb.gov.sg/healthyliving/article.aspx?id=8958 Heatherton, T. F., Herman, C. P., & Polivy, J. (1991). Effects of physical threat and ego threat on eating behavior. Journal of Personality and Social Psychology, 60, 138– 143. Herman, C. P., & Mack, D. (1975). Restrained and unrestrained eating. Journal of Personality, 43, 647–660. Herman, C. P., & Polivy, J. (1975). Anxiety, restraint, and eating behavior. Journal of Abnormal Psychology, 84, 666-672. Herman, C. P., & Polivy, J. (1980). Restrained eating. In A. Stunkard (Ed.), Obesity (pp. 208–225). Philadelphia, PA: W.B. Saunders. 51 Herman, C. P., & Polivy, J. (1988). Excess and restraint in bulimia. In K. Pirke, W. Vandereycken & D. Ploog (Eds.), The psychobiology of bulimia nervosa (pp. 33– 41). New York, NY: Springer-Verlag. Heyward, V. H., & Wagner, D. R. (2004). Applied body composition assessment. Champaign, IL: Human Kinetics. Hill, R.J. & Davies, P.S.W. (2001). The validity of self-reported energy intake as determined using the doubly labelled water technique. British Journal of Nutrition, 85, 415–430. Hodgins, D. C., el-Guebaly, N., & Armstrong, S. (1995). Prospective and retrospective reports of mood states before relapse to substance use. Journal of Consulting and Clinical Psychology, 63, 400–407. Jacobs, D. R. J., Ainsworth, B. E., Hartman, T. J., & Leon, A. S. (1993). A simultaneous evaluation of 10 commonly used physical activity questionnaires. Medicine and Science in Sports and Exercise, 25, 81–91. Kahan, D., Polivy, J., & Herman, C. P. (2003). Conformity and dietary disinhibition: A test of the ego-strength model of self-regulation. International Journal of Eating Disorders, 32, 165–171. Kaplan, H. I., & Kaplan, H. S. (1957). The psychosomatic concept of obesity. Journal of Nervous and Mental Disease, 125, 181–201. Kouvonen, A., Kivimäki, M., Cox, S. J., Cox, T., & Vahtera, J. (2005). Relationship between work stress and body mass index among 45, 810 female and male employees. Psychosomatic Medicine, 67, 577–583. 52 Laessle, R. G., Reinhard, J. T., Kotthaus, B. C., & Pirke, K. M. (1989). A comparison of the validity of three scales for the assessment of dietary restraint. Journal of Abnormal Psychology, 98, 504–507. Levine, J. A. (2005). Measurement of energy expenditure. Public Health Nutrition, 8, 1123–1132. Lightfoot, S. L., & Oliver, J. M. (1985). The Beck Inventory: Psychometric properties in university students. Journal of Personality Assessment, 49, 434–436. Low, Y. L. (2009). Food diary from the GUSTO birth cohort study [Measurement instrument]. Unpublished instrument. Lowe, M. R. (1993). The effects of dieting on eating behavior: A three-factor model. Psychological Bulletin, 114, 100–121. Lowe, M. R., Annunziato, R. A., Markowitz, J. T., Didie, E., Bellace, D. L., Riddell, L., . . . Stice, E. (2006). Multiple types of dieting prospectively predict weight gain during the freshman year of college. Appetite, 47, 83–90. Marfell-Jones, M., Olds, T., Stewart, A., & Carter, J. E. L. (2006). International Standards for Anthropometric Assessment. Potchefstroom, South Africa: The International Society for the Advancement of Kinanthropometry. McLean, J. A., & Barr, S. I. (2003). Cognitive dietary restraint is associated with eating behaviors, lifestyle practices, personality characteristics and menstrual irregularity in college women. Appetite, 40, 185–192. Mclean, J. A., Barr, S. I., & Prior, J. C. (2001). Dietary restraint, exercise, and bone density in young women: Are they related? Medicine & Science in Sports & Exercise, 33, 1292–1296. 53 Michaud, C., Kahn, J. P., Musse, N., Burlet, C., Nicolas, J. P., & Mejean, L. (1990). Relationships between a critical life event and eating behaviour in high-school students. Stress Medicine, 6, 57–64. Muraven, M., & Baumeister, R. F. (2000). Self-regulation and depletion of limited resources: Does self-control resemble a muscle? Psychological Bulletin, 126, 247–259. Muraven, M., Baumeister, R. F., & Tice, D. M. (1999). Longitudinal improvement of self-regulation through practice: Building self-control strength through repeated exercise. The Journal of Social Psychology, 139, 446–457. National University of Singapore, Centre for Molecular Epidemiology. (2010). Food Intake Assessment [Computer program]. Singapore: National University of Singapore. Ng, D. M., & Jeffery, R. W. (2003). Relationships between perceived stress and health behaviors in a sample of working adults. Health Psychology, 22, 638–642. Nieuwenhuizen, A. G., & Rutters, F. (2008). The hypothalamic-pituitary-adrenal-axis in the regulation of energy balance. Physiology & Behavior, 94, 169–177. O'Connor, D. B., Jones, F., Conner, M., McMillan, B., & Ferguson, E. (2008). Effects of daily hassles and eating style on eating behavior. 27, S20–S31. Oliver, G., Wardle, J., & Gibson, E. L. (2000). Stress and food choice: A laboratory study. Psychosomatic Medicine, 62, 853–865. Oman, R. F., & King, A. C. (2000). The effect of life events and exercise program format on the adoption and maintenance of exercise behavior. Health Psychology, 19, 605–612. 54 Ouwens, M. A., van Strien, T., & van der Staak, C. P. F. (2003). Tendency toward overeating and restraint as predictors of food consumption. Appetite, 40, 291–298. Polivy, J., & Herman, C. P. (1999). Distress and eating: Why do dieters overeat? International Journal of Eating Disorders, 26, 153–164. Polivy, J., Herman, C. P., & McFarlane, T. (1994). Effects of anxiety on eating: Does palatability moderate distress-induced overeating in dieters? Journal of Abnormal Psychology, 103, 505–510. Pollard, T. M., Steptoe, A., Canaan, L., Davies, G. J., & Wardle, J. (1995). Effects of academic examination stress on eating behavior and blood lipid levels. International Journal of Behavioral Medicine, 2, 299–320. Rand, C. S. W., & Kuldau, J. M. (1991). Restrained eating (weight concerns) in general population and among students. International Journal of Eating Disorders, 10, 699–708. Roberts, C., Troop, N., Connan, F., Treasure, J., & Campbell, I. C. (2007). The effects of stress on body weight: Biological and psychological predictors of change in BMI. Obesity, 15, 3045–3055. Rosmond, R., & Björntorp, P. (1999). Psychosocial and socio-economic factors in women and their relationship to obesity and regional body fat distribution. International Journal of Obesity, 23, 138–145. Rosmond, R., Lapidus, L., & Björntorp, P. (1996). The influence of occupational and social factors on obesity and body fat distribution in middle-aged men. International Journal of Obesity, 20, 599–607. 55 Rutledge, T., & Linden, W. (1998). To eat or not to eat: Affective and physiological mechanisms in the stress-eating relationship. Journal of Behavioral Medicine, 21, 221–240. Schachter, S. (1971). Some extraordinary facts about obese humans and rats. American Psychologist, 26, 129–144. Schmeichel, B. J., & Baumeister, R. F. (2004). Self-regulatory strength. In R. F. Baumeister & K. D. Vohs (Eds.), Handbook of self-regulation: Research, theory, and applications (pp. 84–98). New York, NY: Guilford Press. Siahpush, M., & Carlin, J. B. (2006). Financial stress, smoking cessation and relapse: Results from a prospective study of an Australian national sample. Addiction, 101, 121–127. Siri, W. E. (1956). The gross composition of the body. In J. H. Lawrence & C. A. Tobias (Eds.), Advance in biological and medical physics, IV (pp. 239–280). New York, NY: Academic Press. Slochower, J. (1976). Emotional labeling and overeating in obese and normal weight individuals. Psychosomatic Medicine, 38, 131–139. Slochower, J., & Kaplan, S. P. (1980). Anxiety, perceived control, and eating in obese and normal weight persons. Appetite, 1, 75–83. Sonnentag, S., & Jelden, S. (2009). Job stressors and the pursuit of sport activities: A day-level perspective. Journal of Occupational Health Psychology, 14, 165–181. Steptoe, A. (1991). The links between stress and illness. Journal of Psychosomatic Research, 35, 633–644. 56 Steptoe, A., Lipsey, Z., & Wardle, J. (1998). Stress, hassles and variations in alcohol consumption, food choice and physical exercise: A diary study. British Journal of Health Psychology, 3, 51–63. Steptoe, A., Wardle, J., Pollard, T. M., Canaan, L., & Davies, G. J. (1996). Stress, social support and health-related behavior: A study of smoking, alcohol consumption and physical exercise. Journal of Psychosomatic Research, 41, 171–180. Strazullo, P., D'Elia, L., Cairella, G., Garbagnati, F., Cappuccio, F. P., & Scalfi, L. (2010). Excess body weight and incidences of stroke: Meta-analysis of prospective studies with 2 million participants. Stroke, 41, e418–e426. Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics. Boston, MA: Allyn & Bacon. Thompson, F. E., & Subar, A. F. (2001). Dietary assessment methodology. In F.E. Thompson. & T. Byers (Eds.), Nutrition in the prevention and treatment of disease (pp. 3–19). San Diego, CA: Academic Press. Tice, D. M. (2009). How emotions affect self-regulation. In J. P. Forgas, R. F. Baumeister & D. M. Tice (Eds.), Psychology of self-regulation: Cognitive, affective, and motivational processes (pp. 201–215). New York, NY: Routledge. Trabulsi, J. & Schoeller, D.A. (2001). Evaluation of dietary assessment instruments against doubly labeled water, a biomarker of habitual energy intake. American Journal of Physiology-Endocrinology and Metabolism, 281, E891–899. Tremblay, A., & Therrien, F. (2006). Physical activity and body functionality: Implications for obesity prevention and treatment. Canadian Journal of Physiology and Pharmacology, 84, 149–156. 57 van Strien, T., Engels, R. C. M. E., & van Staveren, W. (2006). The validity of dietary restraint scales: Comment on Stice et al. (2004). Psychological Assessment, 18, 89–94. van Strien, T., Frijters, J. E. R., Bergers, G. P. A., & Defares, P. B. (1986). The Dutch Eating Behavior Questionnaire for assessment of restrained, emotional, and external eating behavior. International Journal of Eating Disorders, 5, 295–315. Vella-Zarb, R. A., & Elgar, F. J. (2009). The 'Freshman 5': A meta-analysis of weight gain in the freshman year of college. Journal of American College Health, 58, 161–166. Vohs, K. D., & Heatherton, T. F. (2000). Self-regulatory failure: A resource-depletion approach. Psychological Science, 11, 249–254. Wallis, D. J., & Heatherington, M. M. (2004). Stress and eating: The effects of ego-threat and cognitive demand on food intake in restrained and emotional eaters. Appetite, 43, 39–46. Wardle, J., Steptoe, A., Oliver, G., & Lipsey, Z. (2000). Stress, dietary restraint and food intake. Journal of Psychosomatic Research, 48, 195–202. Williams, P. T. (2008). Asymmetric weight gain and loss from increasing and decreasing exercise. Medicine & Science in Sports & Exercise, 40, 296–302. World Health Organization (2011a). Obesity and overweight. Retrieved from http://www.who.int/mediacentre/factsheets/fs311/en/index.html World Health Organization (2011b). WHO Global Infobase: International comparisons. Retrieved from https://apps.who.int/infobase/Comparisons.aspx [...]... informative to examine changes in the body fat of restrained eaters across varying stress levels 1.3 Stress and Body Fat Gain: Eating Behavior as an Explanation One psychological pathway affecting a restrained eater’s body fat during stress could be stress-induced changes in eating behavior and a resultant increase in energy intake For example, stress has been associated with higher dietary fat intake (Ng... explanation for restrained eaters tendency to overeat under stress The tendency to overeat may in turn account for restrained eaters body fat gain under stress 1.4 Stress and Body Fat Gain: Exercise Behavior as an Explanation Another psychological pathway that might account for the body fat gain of restrained eaters under stress is a reduction in physical exercise Apart from increasing energy intake, psychological. .. stress-induced changes in the eating and exercising behaviors of high restrained eaters and low restrained eaters and how these changes may influence the body fat of these individuals Female undergraduates were measured over three time points corresponding to the low stress baseline, high stress pre-examination, and low stress recovery 1.5.1 Aims The primary aim is to investigate the psychological pathways influencing. .. ***p < 001 28 33.5 33.0 Body Fat (%) 32.5 * ** 32.0 31.5 31.0 30.5 30.0 29.5 29.0 28.5 28.0 Low Stress, Baseline High Stress, Preexamination Low Stress, Recovery Stress Level High Restrained Eaters Low Restrained Eaters Figure 4 Changes in body fat percentage across stress by restrained eating group ―*‖ and ―**‖ indicate that adjacent points are significantly different, p < 05 and p < 01 respectively... recoding high restrained eaters as ―1‖ and low restrained eaters as ―0‖ before repeating the MLM analysis (Curran, Bauer, & Willoughby, 2004) The findings show that high restrained eaters had a mean baseline BF% of 32.44% and low restrained eaters had a mean baseline BF% of 29.18% The estimate for time2 was significant for high restrained eaters but not significant for low restrained eaters In addition,... influencing body fat gain among high restrained eaters, a group that is vulnerable to obesity A secondary aim is to examine whether high restrained eaters would show an increase in energy intake during stressful times to replicate past studies done on stress-induced eating Another secondary aim is to examine whether high restrained eaters and low restrained eaters differed in the pattern of changes in exercise. .. and a consequent depletion of self-regulatory resources, the frequency of exercise behavior may decrease If high restrained eaters 10 resources were depleted by both stress and dieting, exercise levels may decrease by a significant extent This study is interested to examine how stress-induced changes in exercise behavior may differ across high restrained eaters and low restrained eaters 1.5 Aims and. .. explained by restrained eating In other words, the model tested for the moderating role of restrained eating Hence, the level 1 intercepts and regression coefficients were regressed on restrained eating The parameter estimates, γ01, γ11, and γ21 represented the difference in the intercepts, linear trends, and quadratic trends of the dependent variable between high restrained eaters and low restrained eaters. .. the role of restrained eating as a moderator in stressinduced eating, no studies have examined the moderating influence of restrained eating in the relationship between stress and exercise behavior Exercise behavior does not appear to be immediately relevant to the construct of restrained eating But it is likely that restrained eaters will be concerned about having sufficient amounts of exercise on... the other half were categorized as low restrained eaters The mean age of high restrained 15 eaters was 19.13 years (SD = 0.45) and the mean age of low restrained eaters was 19.25 years (SD = 0.85) High restrained eaters had a restrained eating score of 3.83 (SD = 0.37) while low restrained eaters had a restrained eating score of 1.34 (SD = 0.18) As seen from Table 1, the sample’s BF% was in the ―below ... effects of psychological stress on body fat via the pathways of eating (energy intake) and exercise behaviors (energy expenditure), among high restrained eaters (dieters) and low restrained eaters. .. changes in eating and exercise behaviors differed across the two groups The primary purpose is to delineate the psychological pathways influencing body fat gain among high restrained eaters Eating. .. concern 1.2 Body fat gain: Who is at risk and when? 1.3 Stress and body fat gain: Eating behavior as an explanation 1.4 Stress and body fat gain: Exercise behavior as an explanation 1.5 Aims and hypotheses