Relations Between Dietary Restraint and Patterns of Alcohol Use in ...

Copyright 2000 by the Educational Publishing Foundation 0893-164X/00/$5.00 DOI: 10.1037//0893-164X.14.I.77

Psychology of Addictive Behaviors 2000, Vol. 14, No. 1, 77-82

Relations Between Dietary Restraint and Patterns of Alcohol Use in Young Adult Women Jan M. Baker and Fred J. Boland Queen's University

Sherry H. Stewart and Maria Angelopoulos Dalhousie University

The present study examined relations between dietary restraint and self-reported patterns of alcohol use, including separate assessment of quantity and frequency of alcohol consumption. One hundred seventy-six female university undergraduates completed the Restraint Scale (RS) and measures of their usual quantity and frequency of alcohol consumption over the past year. Quantity and frequency self-reports were scored separately and were also used to calculate 3 additional drinking variables: a composite weekly alcohol consumption score (drinks per week), a binge drinking categorical variable (where participants were classified as either binge drinkers or non-binge drinkers), and a yearly excessive drinking score (number of times in the past year that each participant consumed at least 4 alcoholic beverages per drinking occasion). RS scores were significantly positively correlated with scores on 4 of the 5 drinking behavior measures (i.e., quantity, drinks per week, binge drinking, and yearly excessive drinking, but not frequency). Thus, chronic dieting appears to be related to a relatively heavy drinking pattern that can be characterized as potentially risky, due to its established associations with adverse health and social consequences.

Clinical studies suggest a high comorbidity between eating and alcohol use disorders (see Holderness, Brooks-Gunn, & Warren, 1994; Wilson, 1993, for reviews). Women with eating disorders display above-average rates of alcohol abuse diagnoses (e.g., Jones, Cheshire, & Moorhouse, 1985). Similarly, in female alcoholics, comorbid eating disorder rates far exceed

prevalence estimates for eating disorders in the general female population (e.g., Taylor, Peveler, Hibbert, & Fairburn, 1993). In such comorbid patients, eating disorders tend to precede the emergence of problem drinking. However, the precise mechanisms underlying the statistical and temporal associations between eating and alcohol use disorders remain unknown. This article explores the potential contribution of dietary restraint as an explanatory variable for this association. Dietary restraint refers to the tendency to exhibit chronic dieting owing to concerns about body weight and shape and is a risk factor for (Polivy & Herman, 1985), and symptom of (Diagnostic and Statistical Manual of Mental Disorders; DSM-IV; American Psychiatric Association, 1994), the clinical eating disorders of anorexia nervosa and bulimia nervosa. Research with nonclinical samples supports dietary restraint as a factor associated with increased alcohol use and misuse. Xinaris and Boland (1990) have found a significant correlation between undergraduate women's Restraint Scale (RS; Polivy, Herman, & Walsh, 1978)

Sherry H. Stewart and Maria Angelopoulos, Clinical Psychology Program, Department of Psychology, Dalhousie University, Halifax, Nova Scotia, Canada; Jan M. Baker and Fred J. Boland, Department of Psychology, Queen's University, Kingston, Ontario, Canada. This study was supported in part by a grant from the Social Sciences and Humanities Research Council of Canada. We wish to extend our thanks to J. Grant MacLeod, Paula McPherson, and Sarah Barton Samoluk for their research assistance. Correspondence concerning this article should be addressed to Sherry H. Stewart, Clinical Psychology Program, Department of Psychology, Dalhousie University, Life Sciences Centre, 1355 Oxford Street, Halifax, Nova Scoria, Canada B3H 4J1. Electronic mail may be sent to [email protected]. 77

78

BRIEF REPORTS

scores and self-reported levels of weekly alcohol consumption (cf. Lavik, Clausen, & Pedersen, 1991). Moreover, restrained-eating women report significantly more episodes of drinking to intoxication than unrestrained-eating women, suggesting a pattern of high quantities of alcohol consumed per drinking occasion (i.e., binge drinking) among young women high in dietary restraint (Lavik et al., 1991). In addition, Stewart and Samoluk (1997) have found that university students scoring high on the RS evidence selective processing of alcohol cues, suggesting that alcohol is an area of current concern for students high in dietary restraint. In contrast, Fisher, Schneider, Pegler, and Napolitano (1991) have found a nonsignificant correlation between a measure of disordered attitudes toward food, weight, and eating and a measure of frequency of alcohol consumption in a female adolescent sample. However, they did not include measures of dietary restraint or of the quantity of alcohol consumed per drinking occasion. In fact, no study to date has separately examined the relations between dietary restraint and quantity versus frequency of alcohol use. Unfortunately, the common practice of assessing drinks per week (Quantity X Frequency) precludes the identification of drinking patterns that may be most highly associated with such alcohol problems as social harm and health risks (e.g., interpersonal conflict, academic failure, fetal alcohol syndrome; Rehm et al., 1996; Wechsler, Davenport, Dowdall, Moeykens, & Castillo, 1994). Because much research indicates that drinking quantity is associated with a greater risk for alcohol problems than is drinking frequency (Rehm et al., 1996; VogelSprott, 1983; Wechsler et al., 1994), it is important to differentiate the binge drinker (e.g., someone who consumes four to five drinks in a row on a single drinking occasion) from the frequent light drinker (e.g., someone who consumes one drink with dinner four or five times a week). Investigating the relations between dietary restraint and specific alcohol consumption patterns could also be clinically useful in identifying which particular drinking behaviors should be targets for intervention with restrained eaters. This study examines relations between RS scores and self-reported alcohol use rates, including separate assessments of quantity and

frequency of alcohol consumption. The relation between dietary restraint and binge drinking suggested by Lavik et al. (1991) is examined directly by using information from the drinking self-reports to classify each participant as either characteristically a binge drinker or a non-binge drinker (cf. Wechsler et al., 1994) and to estimate the number of times each participant drank to excess in the past year (cf. Stewart, Zvolensky, & Eifert, 1999). We predict a significant positive relation between dietary restraint and self-reported drinks per week (cf. Xinaris & Boland, 1990). On the basis of Lavik et al.'s (1991) hypothesis, we predict significant positive relations between dietary restraint and drinking quantity, the binge drinking dichotomous variable, and excessive drinking episodes in the past year. In keeping with previous work (Fisher et al., 1991), we predict no significant association between dietary restraint and drinking frequency.

Method Participants Participants were 176 women (mean age = 20.4 years, SD = 3.4) enrolled in undergraduate psychology classes at Dalhousie (n = 56) or Queen's (n = 120) Universities. We limited our investigation to young adult women because of the increased risk for eating disorders in this population (DSM—IV; American Psychiatric Association, 1994). Approximately 93% of the sample (n = 164) were of legal age for drinking in the provinces of Nova Scotia and Ontario (i.e., 19 years of age or older).

Procedure Participants provided informed consent and then completed the following self-report measures anonymously during class time in the order described. Demographics and drinking behavior measure. Participants completed a demographic measure, assessing age and current year at university. Frequency and quantity of alcohol consumption in the past year were assessed on the same questionnaire (cf. Stewart, Peterson, & Pihl, 1995). For frequency, participants reported the number of occasions per week on which they usually consumed alcohol. Those who consumed alcohol on less than one occasion weekly reported on their monthly or yearly frequency. For quantity, participants indicated the average number of alcoholic beverages they usually consumed per drinking occasion (one alcoholic

BRIEF REPORTS beverage = 12 oz beer, 4 oz wine, or 1 oz hard liquor). We used methods recommended by Sobell and Sobell (1990) for enhancing the accuracy of these retrospective self-reports (e.g., drinking questions were open-ended; confidentiality was assured). Restraint Scale. The RS is a 10-item, multiplechoice questionnaire assessing a combination of dietary restraint (i.e., chronic attempts to restrict food intake due to concerns about body weight; 6 items) and dietary disinhibition (as reflected by a history of body weight fluctuations; 4 items). Each question has either 4 or 5 possible response options. This measure has satisfactory levels of reliability and validity (Ruderman, 1986).

Results Scoring of the Drinking Behavior Measures Five drinking behavior measures were scored for each participant. Frequency and quantity were first scored separately. Average weekly frequency and quantity were then multiplied to yield a composite drinks per week measure (cf. Stewart et al., 1995). We also used the selfreported typical drinking quantity information to calculate a binge drinking dichotomous variable: Those who reported that their usual quantity of alcohol consumption was four or more drinks per drinking occasion were classified as binge drinkers; all others were classified as non-binge drinkers. A minimum of four drinks per occasion was used as the cutoff, given this is the minimal drinking quantity associated with adverse outcomes for young women (Wechsler et al., 1994). Finally, we calculated a yearly excessive drinking measure, defined as the average number of times in the past year each student drank four or more drinks on a single drinking occasion (cf. Stewart et al., 1999).

Sample Means The mean RS score for the entire sample was 13.3 (SD = 6.7), which compared well with scores reported for previously tested nonclinical female samples (e.g., Knight & Boland, 1989). Mean drinking frequency was 1.0 (SD = 0.8) drinking occasions per week; mean drinking quantity was 4.3 (SD = 2.9) standard drinks per drinking occasion; mean drinks per week was 5.2 (SD = 6.2) standard drinks; 54.0% of the

79

sample were classified as binge drinkers; and the sample reported drinking to excess 35.0 (SD = 43.1) times in the past year, on average. These sample means were consistent with those reported for previously tested nonclinical female samples (e.g., Stewart et al., 1995; Wechsler et al., 1994; Xinaris & Boland, 1990).

Relations Between Dietary Restraint and Drinking Behavior Measures We examined relations between RS scores and the drinking behavior measures in a continuous fashion through the use of correlational-regression techniques (as opposed to extreme RS-group comparisons) to maximize statistical power and for direct comparison of our results with previous correlational findings (e.g., Xinaris & Boland, 1990). All five drinking behavior measures were significantly intercorrelated, with shared variance between measures ranging from 10.7% (drinking frequency with binge drinking) to 70.2% (drinks per week with yearly excessive drinking; allps < .001, one-tailed tests). Given the multicollinearity of the drinking behavior measures, we first performed canonical correlation between RS scores (Variable Set 1) and scores on the five drinking behavior measures (Variable Set 2). The first canonical correlation was .26 (6.5% shared variance between RS scores and the set of drinking behavior measures; p < .001). The remaining four canonical correlations were effectively zero. Next, bivariate correlations were calculated between RS scores and each of the five drinking behavior measures. Given the directional hypotheses, one-tailed tests were used, and a stringent Bonferroni-adjusted alpha level of .01 (.05/5 tests) was applied in evaluating significance. As hypothesized, RS scores were not significantly associated with the drinking frequency measure (f = .14, ns). In keeping with prediction, RS scores were significantly positively correlated with the binge drinking (r = .30, p < .001), drinking quantity (r = .26, p < .001), yearly excessive drinking (r = .21, p < .005), and composite drinks per week (r = .18, p < .01) measures. The magnitude and levels of significance of the correlations did not change when the effects of age and education level were partialed out. We also examined whether the significant

BRIEF REPORTS relations between RS scores and the binge drinking and yearly excessive drinking measures reported earlier were due to the particular cutoff used to operationalize binge drinking. For this set of analyses, we calculated a more stringent binge drinking dichotomous variable, where only those who reported that their usual quantity of alcohol consumption was five or more drinks per drinking occasion were classified as binge drinkers. Similarly, we calculated a more stringent yearly excessive drinking measure, defined as the average number of times in the past year each participant drank five or more drinks per occasion. The first canonical correlation between RS scores and the set of five drinking behavior measures was again significant (r = .24, p < .001), and the remaining four canonical correlations were again effectively zero. Moreover, RS scores were again significantly positively correlated with the more stringently coded binge drinking variable (r = .24, p < .001) and with the more stringently coded yearly excessive drinking measure (r = .20. p = .005). Thus, it can be fairly safely concluded that the significant relations between dietary restraint and binge drinking or yearly excessive drinking reported earlier are not an artifact of the particular cutoff value used to define binge drinking.

Discussion Using canonical correlation, we found that dietary restraint levels accounted for a significant 6.5% of the variance in our set of drinking behavior measures, after accounting for shared variance between the drinking measures. Like Xinaris and Boland (1990), we found that level of dietary restraint was positively correlated with the composite drinks per week measure (Quantity X Frequency). However, when quantity and frequency estimates were examined separately (cf. Vogel-Sprott, 1983), quantity but not frequency was significantly correlated with RS scores. Our failure to find a significant relation between dietary restraint and drinking frequency is consistent with the results of Fisher et al. (1991) that disturbed eating attitudes were unrelated to drinking frequency. We also found direct support for Lavik et al.'s (1991) suggestion that restrained-eating women may be characterized by a binge drinking profile:

Higher levels of dietary restraint were significantly associated both with a greater likelihood of being classified as a binge drinker and with a greater number of excessive drinking episodes in the past year. Taken together, our results suggest that young women high in dietary restraint do not drink more often than others, but when they do drink, they drink significantly more alcohol. Because an increased quantity of alcohol consumption has been associated with adverse social and health consequences more strongly than increased drinking frequency (e.g., Rehm et al., 1996), the present findings suggest that restrained-eating women display a relatively risky pattern of alcohol use. One possible explanation for the significant associations observed between dietary restraint and increased drinking quantity involves the notion of drinking restraint. Because women who control their dietary intake selectively avoid high-calorie foods (Francis, Stewart, & Hounsell, 1997; Knight & Boland, 1989), they may also similarly avoid alcohol given its relatively high caloric content. In fact, recent research shows that RS scores are positively associated with drinking restraint levels (Ricciardelli & Williams, 1997). Just as it has been suggested that attempts to limit caloric intake may in fact cause binge eating (Polivy & Herman, 1985), behavioral attempts to limit drinking among high-dietary-restraint women could be responsible for their higher rates of binge drinking (see Collins, 1993, for a review of the literature linking drinking restraint to increased binge drinking). The drinking restraint explanation for the current findings could be further evaluated using laboratory-based behavioral assessment methods (e.g., the alcohol taste test; Samoluk & Stewart, 1996) to determine whether the administration of a high-caloric alcohol preload leads to disinhibited alcohol consumption among restrained eaters. Another possible explanation for the current findings of significant associations between RS scores and increased heavy drinking involves a dysregulation hypothesis. It is possible that high scores on the RS may be reflecting a general dysfunction in behavioral regulation. The dysregulation hypothesis could be evaluated further by examining associations between RS scores and forms of behavioral dysregulation in addition to eating-drinking disturbances (e.g., risky sexual

BRIEF REPORTS behavior). A further variable that may be useful in examining the dysregulation hypothesis is that of impulsivity (i.e., failure to consider risks and consequences before acting; Fahy & Eisler, 1993). Bulimic women have been described as impulsive, and this tendency may account for the increased substance use and promiscuous behavior observed in this clinical population (e.g., Vitousek & Manke, 1994; Wiederman & Pryor, 1996). Examination of levels of impulsivity among nonclinical samples of restrainedeating women may be useful in explaining the associations between dietary restraint and heavy drinking-binge eating. The fact that RS scores accounted for a relatively small proportion of the variance in drinking behavior suggests that characteristics assessed by the RS total score alone cannot be entirely responsible for the high comorbidity between eating disorders and alcohol abuse. In fact, both bulimics and restricting anorexics show elevated dietary restraint, but only the former show increased rates of alcohol abuse (Wilson, 1993). Future research may wish to assess relations between RS components (i.e., concern for dieting vs. weight fluctuations; Heatherton, Herman, Polivy, King, & McGree, 1988) and alcohol consumption patterns. The dysregulation hypothesis would predict that the weight fluctuations (i.e., dietary disinhibition) component would be more highly associated with increased drinking quantity and binge drinking, whereas the drinking restraint hypothesis would predict that the concern for dieting (i.e., efforts aimed at intake restraint) component would be most associated with drinking quantity-binge drinking. Several potential limitations to the current study should be acknowledged. First, the correlational nature of the present findings precludes drawing conclusions that dietary restraint is causally related to heavier drinking. Second, there is the possibility that the associations observed are secondary to a relation between body weight and drinking quantity because restrained eaters may be overrepresented in higher body weight categories (Heatherton et al., 1988). However, Xinaris and Boland (1990) found a significant positive correlation between RS scores and weekly alcohol consumption self-reports even after accounting for body weight. Nonetheless, future research should

81

include body weight measures to determine whether restrained-eating women actually drink larger doses of alcohol per drinking occasion (e.g., in milliliters of alcohol per kilograms of body weight) or drink to the point of objective intoxication more frequently (Conrod, Stewart, & Pihl, 1997) than nonrestrained eaters. The present findings are also limited by the fact that alcohol consumption behavior in the past year was assessed through retrospective self-report methods, which can be subject to errors such as memory inaccuracies. The accuracy of such self-reports might be further improved through the use of the timeline-followback procedure (e.g., Sobell, Sobell, Leo, & Concilia, 1988) hi future research. Finally, our use of a relatively light drinking sample of university women may limit the generalizability of our findings. The present results thus require replication in a larger, more representative nonclinical sample that allows for an assessment of gender differences as well as including heavier drinkers. The applicability of the present findings to understanding the overlap of eating and alcohol use disorders hi clinical samples remains to be established through examination of the drinking patterns of such comorbid patients and examination of RS scores as a possible mediator of increased drinking behavior in diagnosed eating disorder patients relative to controls.

References American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Collins, R. L. (1993). Drinking restraint and risk for alcohol abuse. Experimental and Clinical Psychopharmacology, 1, 44—54. Conrod, P. J., Stewart, S. H., & Pihl, R. O. (1997). Validation of a measure of excessive drinking: Frequency per year that BAL exceeds 0.08%. Substance Use and Misuse, 32, 587-607. Fahy, T., & Eisler, I. (1993). Impulsivity and eating disorders. British Journal of Psychiatry, 162, 193-197. Fisher, M., Schneider, M., Pegler, C., & Napolitano, B. (1991). Eating attitudes, health-risk behaviors, self-esteem, and anxiety among adolescent females in a suburban high school. Journal of Adolescent Health, 12, 377-384. Francis, J., Stewart, S. H., & Hounsell, S. (1997). Dietary restraint and the selective processing of

82

BRIEF REPORTS

forbidden and non-forbidden food words. Cognitive Therapy and Research, 21, 633-646. Heatherton, T., Herman, C. P., Polivy, 1, King, G. A., & McGree, S. T. (1988). The (mis)measurement of restraint: An analysis of conceptual and psychometric issues. Journal of Abnormal Psychology, 97, 19-28. Holderness, G. C., Brooks-Gunn, J., & Warren, M. P. (1994). Co-morbidity of eating disorders and substance abuse: Review of the literature. International Journal of~ Eating Disorders, 16, 1—34. Jones, D. A., Cheshire, N., & Moorhouse, H. (1985). Anorexia nervosa, bulimia and alcoholism: Association of eating disorder and alcohol. Journal of Psychiatric Research, 19, 377-380. Knight, L. J., & Boland, F. J. (1989). Restrained eating: An experimental disentanglement of the dismhibiting variables of perceived calories and food type. Journal of Abnormal Psychology, 98, 412-420. Lavik, N. J., Clausen, S. E., & Pedersen, W. (1991). Eating behavior, drug use, psychopathology and parental bonding in adolescents in Norway. Acta Psychiatrics Scandinavia, 84, 387-390. Polivy, J., & Herman, C. P. (1985). Dieting and bingeing: Acausal analysis. American Psychologist, 40, 193-201. Polivy, J., Herman, P., & Walsh, S. (1978). Internal and external components of emotionality in restrained and unrestrained eaters. Journal of Abnormal Psychology, 87, 497-504. Rehm, J., Ashley, M. J., Room, R., Single, E., Bondy, S., Ferrence, R., & Giebrecht, N. (1996). On the emerging paradigm of drinking patterns and their social and health consequences: Summary of the proceedings of the International Conference on Social and Health Effects of Different Drinking Patterns. Addiction, 91, 1615-1621. Ricciardelli, L. A., & Williams, R. J. (1997). A two-factor model of dietary restraint. Journal of Clinical Psychology, 53, 123-131. Ruderman, A. J. (1986). Dietary restraint: A theoretical and empirical review. Psychological Bulletin, 99, 247-262. Samoluk, S. B., & Stewart, S. H. (1996). Anxiety sensitivity and anticipation of a self-disclosing speech as determinants of alcohol consumption. Psychology of Addictive Behaviors, JO, 45—54. Sobell, L. C., & Sobell, M. B. (1990). Self-report issues in alcohol abuse: State of the art and future directions. Behavioral Assessment, 12, 77-90. Sobell, L. C., Sobell, M. B., Leo, G. I., & Cancilla, A. (1988). Reliability of a timeline method: Assessing

normal drinkers' reports of recent drinking and a comparative evaluation across several populations. British Journal of Addiction, 83, 393^02. Stewart, S. H., Peterson, J. B., & Pihl, R. O. (1995). Anxiety sensitivity and self-reported alcohol consumption rates among university women. Journal of Anxiety Disorders, 9, 283-292. Stewart, S. H., & Samoluk, S. B. (1997). Effects of short-term food deprivation and chronic dietary restraint on the selective processing of appetitiverelated cues. International Journal of Eating Disorders, 21, 129-135. Stewart, S. H., Zvolensky, M. J., & Eifert, G. H. (1999). Negative-reinforcement drinking motives mediate the relation between anxiety sensitivity and increased drinking behavior. Manuscript submitted for publication. Taylor, A. V., Peveler, R. C., Hibbert, G. A., & Fairburn, C. G. (1993). Eating disorders among women receiving treatment for an alcohol problem. International Journal of Eating Disorders, 14, 147-151. Vitousek, K., & Manke, F. (1994). Personality variables and disorders in anorexia nervosa and bulimia nervosa. Journal of Abnormal Psychology, 103, 137-147. Vogel-Sprott, M. (1983). Response measures of social drinking: Research implications and applications. Journal of Studies on Alcohol, 44, 817-836. Wechsler, H., Davenport, A., Dowdall, G., Moeykens, B., & Castillo, S. (1994). Health and behavioral consequences of binge drinking in college: A national survey of students at 140 campuses. Journal of the American Medical Association, 272, 1672-1677. Wiederman, M. W., & Pryor, T. (1996). Multiimpulsivity among women with bulimia nervosa. International Journal of Eating Disorders, 20, 359-365. Wilson, G. T. (1993). Binge eating and addictive disorders. In C. G. Fairburn & G. T. Wilson (Eds.), Binge eating: Nature, assessment and treatment (pp. 97-122). New York: Guilford Press. Xinaris, S., & Boland, F. J. (1990). Disordered eating in relation to tobacco use, alcohol consumption, self-control, and sex-role ideology. International Journal of Eating Disorders, 9, 425^433.

Received June 22, 1998 Revision received February 8, 1999 Accepted February 8, 1999