- Email: [email protected]
Food intake and physical activity: A comparison of three measures of dieting
Addictive Behaviors, Vol. 19, No. 4, pp. 401-409, 1994 Copyright 0 1994 Elsevier Science Ltd Printed in the USA. All rights reserved 0306-4603194 $6.00 + .CQ
Pergamon
FOOD INTAKE AND PHYSICAL ACTIVITY: A COMPARISON THREE MEASURES OF DIETING SIMONE A. FRENCH,*
ROBERT W. JEFFERY,*
OF
and RENA R. WINGt
School of Public Health, University of Minnesota, Minneapolis, MN ‘/‘Western Psychiatric Institute and Clinic, University of Pittsburgh School of Medicine, Pittsburgh, PA *Division
of Epidemiology,
Abstract - We examined food preferences, dietary intake, and physical activity in dieters and nondieters using three different methods of dieting classification. One hundred three women and 99 men completed the cognitive restraint subscale of Stunkard and Messick’s (1985) Three Factor Eating Questionnaire (TFEQ-R) and Herman and Polivy’s (1980) Restraint Scale (RS), and answered questions about their &rent efforts to lose or maintain weight. Women identifying themselves as currently dieting to lose weight reported expending twice as much energy in physical activity compared to those reporting that they were either currently dieting to maintain weight or not dieting. There were no significant differences in dietary intake or physical activity by current dieting status in men. Women chronic dieters identified by high TFEQ-R scores reported lower total calorie intake and less frequent sweets consumption than women with low TFEQ-R scores. In men, those with high TFEQ-R scores reported consuming a greater percent of calories from protein and carbohydrate, less beef, pork, whole milk, and sweets. In women, the RS did not distinguish dieters from nondieters on any measure. In men, the RS results were similar to those from the TFEQ-R. These results show that current measures of dieting are only weakly related to behaviors thought to be indicative of dieting. Future research must develop more precise measures of dieting in order to examine the relationship between self-reports of dieting and behaviors thought to be related to dieting.
Dieting to lose weight is a widespread practice in the United States (French & Jeffery, 1994; Jeffery, Adlis, & Forster, 1991; Williamson, Serdula, Anda, & Byers, 1992). Concern has recently arisen over the possibility that dieting may be more harmful to health than helpful (Garner & Wooley, 1991; Lissner, Andres, Muller, & Shimokata, 1990; Polivy & Herman, 1985; Striegel-Moore, Silberstein, & Rodin, 1986; Wooley & Wooley, 1984). A problem in assessing the health benefits and hazards associated with dieting efforts has been identifying a valid measure of the construct of “dieting.” Epidemiologic studies have typically assessed dieting efforts using a single direct question about whether the respondent is currently trying to lose weight (DiPietro, Williamson, Caspersen, & Eaker, 1993; Horm & Anderson, 1992; Stephenson, Levy, Sass, & McGarvey, 1987; Williamson et al., 1992) or is currently dieting to lose weight (Jeffery et al., 1991; Forman, Trowbridge, Gentry, Marks, & Hogelin, 1986; Rand & Kuldau, 1991). Psychologists investigating dieting behavior have developed multiitem scales that incorporate additional behavioral and psychological characteristics of chronic dieting efforts (Herman & Polivy, 1980; Stunkard & Messick, 1985; Van Strien, Frijters, Bergers, & Defares, 1986). The Restraint Scale (RS) has been hypothesized to identify unsuccessful dieters (Herman & Polivy, 1980; Heatherton, Herman, Polivy, King, & McGree, 1988), while the Three Factor Eating Questionnaire (TFEQ) has been hypothesized to identify successful dieters (Heatherton et al., Research supported by grants 5ROl HL 41332,5ROl HL 41330, and 5T32 HL 07328 from the National Heart, Lung and Blood Institute to Robert W. Jeffery and Rena R. Wing. Requests for reprints should be sent to Simone A. French, PhD, Division of Epidemiology, School of Public Health, University of Minnesota, 1300 South Second Street, Suite 300, Minneapolis, MN 55454-1015. 401
402
S. A. FRENCH
et al.
1988; Stunkard & Messick, 1985). An examination of item content suggests that the RS focuses on previous dieting and weight fluctuations, while the cognitive restraint subscale of the TFEQ (TFEQ-R) assesses habitual behaviors related to restriction of dietary intake. While few studies have provided empirical data on the validity of self-reported dieting behaviors, some research suggests that self-reports of dieting to lose weight are correlated with reported dietary and physical activity patterns. In a study involving a population-based sample of working adults, self-reports of dieting to lose weight were cross-sectionally related to lower self-reported intake of several highfat/high-calorie foods, including sweets, meat, soft drinks, french fries, and dairy products. Dieting was also related to higher self-reported levels of high-intensity physical activity (French, Jeffery, & Forster, 1994). In addition, individuals who reported initiating a weight-loss diet reported decreased consumption of sweets, soft drinks, and french fries, and increased frequency of moderate and high intensity physical activity (French et al., 1994). Studies that have examined the dietary composition of participants in weight-loss programs show that people on self-selected calorie-restricted diets spontaneously reduce consumption of high-fat foods such as meat, fat, dairy products, high-sugar and snack foods (Brightwell, Foster, Lee, & Naylor, 1979; Nowalk & Wing, 1985; Ritt, Jordan, & Levitz, 1979). Finally, two studies of high school students found that self-reports of trying to lose weight were related to lower reported total energy intake (Field, Wolf, Herzog, Cheung, & Colditz, 1993), and self-reports of greater caloric restriction and meal skipping (Rosen & Poplawski, 1987). Self-reported physical activity was unrelated to dieting frequency in the former study (Field et al., 1993) and positively associated with dieting efforts in the latter study (Rosen & Poplawski, 1987). Studies examining the dietary intakes of chronic dieters identified by the RS suggest that this measure is unrelated to total calorie intake (Heather-ton et al., 1988; Kirkley, Burge, & Ammerman, 1988; Laessle, Tuschl, Kotthaus, & Pirke, 1989b). However, chronic dieters identified by the TFEQ-R report lower total calorie intake than nondieters (Heatherton et al., 1988; Kirkley et al., 1988; Laessle et al., 1989a; 1989b; Tuschl, Platte, Laessle, Stichler, & Pirke, 1990; Van Strien et al., 1986; Wardle & Beales, 1987). Analyses of the association between the TFEQ-R scale and specific food preference and consumption indicate that chronic dieters identified by the TFEQ-R do not differ from nondieters in their intake of meat, vegetables, fruits, starches, sweets, breads, cakes, or margarine, but they report less preference for and consumption of oils, mayonnaise, butter, and high-fat dairy products (Laessle et al., 1989a; Sunday, Einhorn, & Halmi, 1992; Tuschl, Laessle, Platte, & Pirke, 1990). These conflicting results may be related to at least two factors that differ between studies. The first is study population. Studies that have found differences between dieters and nondieters in specific food patterns have been comprised of either population-based adult samples, or adults who are participating in weight-loss programs. The studies reporting fewer differences between dieters and nondieters have been done primarily with young, female, college students, a considerably more homogeneous group whose dietary habits may have less interindividual variation. The second factor that differs between the studies is the method used to identify dieters. The studies that have found differences in specific dietary intake have used a direct questioning strategy to assess current dieting efforts, while those that have found fewer differences have relied upon composite scales such as the RS and the TFEQ-R that assess chronic dieting efforts. While chronic dieters may be more likely to be dieting at any given time than non-chronic dieters, they may not have been actively
Dieters’ food intake
403
dieting during the period that their food intake was measured (Lowe, 1993; Lowe, Whitlow, Jz Bellwoar, 1991). Composite scales may also have lower correlations with eating behavior than a direct, single-item measure of dieting simply because these scales assess things other than eating (e.g., prior weight loss, weight concerns). In addition, the correlations may be attentuated because the time period covered by the composite scales differs from that of the single, direct question. To date, no studies have examined relationships between multiple measures of dieting and dietary intake in the same group of individuals. The present study was conducted to examine the relationship between three measures of dieting and macronutrient intake, specific food intake, food preferences, and physical activity levels. It was hypothesized that dieters would report lower total calorie intake, lower percent of total calories from fat, lower intake of high-fat foods such as sweets, meat, fats, and high-fat dairy products, higher intake of low-fat foods such as salads, fruits, and vegetables, and higher levels of physical activity than nondieters. In addition, based on previous research, it was hypothesized that measures of current dieting would be more strongly related to intake of low-fat foods and avoidance of high-fat foods than measures of chronic dieting.
METHOD
Subjects and procedure One hundred three women and 99 men between the ages of 25 and 62 participated in the study. Participants were recruited from the university community, and were required to be living with a partner, friend, or spouse/family member in order to participate in the study. Subjects were recruited through newspaper advertisements and were paid $20.00 to participate in a questionnaire study on eating and exercise habits. During the first session, height and weight were measured and a packet of questionnaires was completed by each participant. Subjects returned for a second questionnaire session 2 weeks later during which some of the questionnaires were completed a second time. Measures Current dieting status. Current dieting status was determined by responses to two questions. Current dieters to lose weight were comprised of subjects who answered “yes” to the question “Are you currently dieting to lose weight?“. Current dieters to maintain weight were comprised of subjects who answered “yes” to the question “Are you currently dieting to maintain your present weight?“. Nondieters were classified as those who reported that they were not currently dieting to lose or maintain weight. The 2-year test-retest reliability of the single item “Are you currently dieting to lose weight” has been shown to be 0.32 in a sample of working adults (French et al., 1994). This indicates that current dieting efforts are related to future dieting efforts, although the relationship is not strong. Chronic dieting status. Chronic dieting status was assessed using two different instruments: the Restraint Scale (RS; Herman & Polivy, 1980) and the Three Factor Eating Questionnaire’s Cognitive Restraint subscale (TFEQ-R; Stunkard & Messick, 1985). The RS is hypothesized to be related to a history of unsuccessful dieting, while the TFEQ is thought to identify successful dieters (Heatherton, Herman,
404
S. A. FRENCH et al.
Polivy, King, & McGree, 1988). The RS is valid in discriminating those who will overeat in laboratory settings under certain conditions (Heatherton et al., 1988; Ruderman, 1986). Its 2-year test-retest reliability is high (r = 0.74), however it has been shown to be less stable among unrestrained eaters than restrained eaters (r = 0.37 vs. r = 0.58, respectively; Klesges, Klem, Epkins, & Klesges, 1991). The TFEQ-R is valid in discriminating those who consume fewer calories in naturalistic settings (Heatherton et al., 1988; Kirkely et al., 1988; Laessle et al., 1989a; Tuschl, Platte, Laessle, Stichler, & Pirke, 1990; Van Strien et al., 1986; Wardle & Beales, 1987). Restrained eaters identified by the TFEQ-R also show depleted fat cell size, which indirectly validates their reports of caloric restriction (Stunkard, 1981). The lmonth test-retest reliability has been shown to be high (r = 0.98; Stunkard & Messick, 1985). Dietary intake. Usual dietary intake was measured using the Block Food Frequency Questionnaire (FFQ; Block et al., 1986). Respondents indicate their typical serving size (small, medium, or large) and frequency of consumption during the previous 6 months (daily, weekly, monthly, yearly, or rarely/never) for 64 food items or groups of items. Daily macronutrient intake and frequency of intake of individual food items and food groupings are calculated. The FFQ has demonstrated reliability and validity (Block, 1982; Block et al., 1986). For this report, total calories per day and percent of calories from fat, protein, alcohol, carbohydrates, and sweets were examined. In addition, frequency of consumption of the food groups beef, cheese, butter, hot dogs, pork, whole milk, sweets, salad, and fruit were examined. Low-fat food intake. In order to specifically examine low-fat food intake, a composite index of low-fat food consumption was created by averaging intake of 22 items from the Block FFQ containing less than 17% calories from fat. Item analysis reduced this pool to 17 items: apples, cantaloupe, oranges, orange juice, grapefruit, other fruit juices, beans, tomatoes, broccoli, spinach, greens, carrots/mixed vegetables, green salad, sweet potatoes, rice, dark bread, and fortified cereal. The Cronbath’s alpha coefficient for this scale is .72, indicating high internal reliability. Low-fat food preferences. In order to specifically examine low-fat food preferences, a composite index of low-fat food preference was created in a similar manner to the low-fat food intake index described above. Respondents indicated their preference for each food listed on the Block FFQ by marking a number corresponding to their preference (1 = don’t like at all, 5 = like very much). Item analysis of the 22 foods containing less than 17% calories from fat reduced the set of items to 21: apples, cantaloupe, oranges, orange juice, grapefruit, other fruit juices, beans, tomatoes, broccoli, spinach, greens, carrots/mixed vegetables, green salad, sweet potatoes, rice, dark bread, fortified cereal, high fiber cereal, other cold cereal, cooked cereal, and skim milk. The Cronbach’s alpha coefficient for this scale is .83, indicating high internal reliability. Physical activity. Habitual physical activity was measured using the Paffenbarger physical activity recall (Paffenbarger, Wing, & Hyde, 1978). The Paffenbarger recall estimates average weekly total calories expended in physical activity over the previous 6 months. Respondents indicate the number of flights of stairs climbed, city blocks walked, and participation in sports or recreational activities (average weekly frequency and duration per session). Concurrent validity has been demonstrated by
Dieters’ food intake
405
the measure’s correlation in predicted directions with body weight (Paffenberger, Hyde, Wing, & Hsieh, 1986). Predictive validity has been demonstrated by the measure’s prospective relationship with coronary heart disease morbidity and mortality (Paffenbarger et al., 1978; Paffenbarger et al., 1986). Body Mass Index. Body Mass Index (BMI) was calculated as weight in kilograms/ height in meters2. Weight was measured in street clothing without shoes, using a balance beam scale. Height was measured without shoes using a wall-mounted ruler and head block. Statistical analysis Prior to analysis, the specific food items, food groups, and physical activity scores were transformed to normalize their skewed distributions. The FFQ items were log transformed, and the Paffenbarger scores were square root transformed. Multiple regression analysis was conducted using the Statistical Analysis System’s PROC GLM (SAS, 1985). Separate sets of analyses were performed for each dieting measure. Current dieting status was entered as a three-level categorical independent variable (weight-loss dieting, weight-maintenance dieting, not dieting). The RS and TFEQ-R were entered as continuous independent variables. Mean comparisons in the tables are based on sex-specific median splits. Age and BMI were entered as covariates in each analysis. Separate analyses were conducted for men and women. RESULTS
Sample characteristics The average age of the men and women in the study was 35.8 (SD = 8.2) years and 36.8 (SD = 9.3) years, respectively. Three quarters of the men and half of the women had completed a college degree, and 60% were currently married. About 44% of the men and 19% of the women were in professional or administrative occupations, and 40% of the men and 75% of the women were in clerical positions. Average BMI for men and women was 26.7 (SD = 5.2) and 25.4 (SD = 4.6), respectively. More women than men reported currently dieting to lose or maintain weight (37% vs. 22%). Women scored higher than men on both the RS (M = 14.6, SD = 5.5, vs. M = 11.O, SD = 5.0) and on the TFEQ-R (M = 9.1, SD = 4.2, vs. M = 5.9, SD = 4.2). These values are comparable to RS and TFEQ-R scores reported in college students (RS: Klesges et al., 1991; Ruderman & Christensen, 1983; Wardle, 1986) and adults (RS: Johnson, Lake, & Mahan, 1984; TFEQ-R: Stunkard & Messick, 1985; Stunkard 8z Wadden, 1990). In the present sample, the partial correlation (age and BMI partialled) between the RS and the TFEQ-R was 0.41 in women and 0.38 in men. In men, 40% of high TFEQR (above sex-specific median) and 9% of low TFEQ-R (below sex-specific median) scorers were currently dieting to lose or maintain weight. In women, the corresponding numbers were 59% and 16%, respectively. In men, 38% of high- and 8% of lowRS scorers reported currently dieting to lose or maintain weight. In women, the corresponding numbers were 44% and 31%, respectively. Current dieting status Table 1 shows age- and BMI-adjusted means in women and men for significant comparisons on total calories, intake of macronutrients and specific foods, and physical activity levels by dieting status. Twenty-three women reported currently dieting
406
S. A. FRENCH et al.
Table 1. Age and Body Mass Index - Adjusted means and standard errors for dietary intake and physical activity by dieting status Current dieting status Weight-loss diet Not dieting Women N Physical activity (kcal/wk) Men N % Protein calories % Fat calories Three Factor
lS;;.O
77 16.8 (l.O)A 34.7 (1.1)
Cognitive
Restraint
Low restraint
51 1446.8 (72.3) 11.6 (1.5) 8.9 (1.7)
51 1650.8 (72.3) 15.9 (1.5) 16.0 (1.7)
50.1 1.9 0.4 0.2 11.8
56 16.6 45.7 2.8 0.9 1.7 14.7
(1.4) (0.3) (0.2) (0.6) (1.7)
& Polivy
Restraint
(1.2) (1.2) (0.3) (0.1) (0.5) (1.5)
P
.03 .02 .0002 .003 .04 .Ol .03 .009 .005
Scale
Low restraint
47 1.9 (0.3) 0.4 (0.2) 0.4 (0.6)
All food groups are in servings per week. Different superscripts cantly different means (p < .05). Note.
.02 .03
Subscale
High restraint
High restraint Men N Beef Pork Whole milk
.0002
il.8 (2.6)B 41.6 (2.8)*
43 19.4 (1.4)
Herman
15 1444.9 (346.8)
::.5 (2.9)*,s 30.9 (3.0)
Questionnaire
Women N Total calories % Sweets calories Sweets Men N % Protein calories % Carbohydrate calories Beef Pork Whole milk Sweets
p
23 2983.5 (281.1)*
(165.2)
Eating
Maintenance diet
52 2.8 (0.3) 0.9 (0.1) 1.7 (0.6)
P
.04 .005 ,002
within rows indicate signifi-
to lose weight, 15 reported currently dieting to maintain weight, and 65 reported not currently dieting. In women, current dieters to lose weight reported expending almost twice as many calories from physical activity as those dieting to maintain weight or not dieting. Eleven men were currently dieting to lose weight, 11 were currently dieting to maintain weight, and 77 were not currently dieting. In men, current dieters to maintain weight reported significantly higher percent calories from protein and fat than dieters to lose weight and nondieters. None of the other dietary intake or physical activity measures differed by current dieting status. Chronic dieting status Three reported and less None of
Factor Eating Questionnaire. Women who scored significantly lower total calorie intake, lower percent frequent sweets consumption in comparison with the other measures differed significantly by TFEQ-R
high on the TFEQ-R calories from sweets, low TFEQ-R women. score. Men with high
Dieters’ food intake
407
TFEQ-R scores reported a significantly greater percent of calorie intake from protein and carbohydrate, and less frequent consumption of beef, pork, whole milk, and sweets than men with low TFEQ-R scores. None of the other measures differed significantly by TFEQ-R score. Restraint score. Women who scored high on Herman & Polivy’s (1980) RS scale did not differ significantly from unrestrained women on any of the dietary-intake or the physical-activity measures. In men, those with high RS scores reported significantly less frequent consumption of beef, pork, and whole milk than those with low RS scores. No other dietary intake or physical activity measures were significantly different by RS score. DISCUSSION
This study is the first to examine the behavioral correlates of dieting using three different measures of dieting: direct questions about current dieting and two scales measuring chronic dieting. It was initially hypothesized that dieters would report a lower intake of high-fat foods, higher intake of low-fat foods, and higher levels of physical activity than nondieters. The TFEQ-R was related to lower total calorie intake, lower percent of calories from sweet foods, and lower frequency of sweets consumption in women. In men, the TFEQ-R was related to a greater percent calories from protein and carbohydrates, lower frequency of beef, pork, whole milk, and sweets. These results are consistent with previous studies using the TFEQ-R that reported a lower total calorie intake and lower consumption of butter and high fat dairy products. The RS was unrelated to intake in women, which is consistent with previous findings and supports the argument that the RS measures unsuccessful dieting, while the TFEQ measures successful dieting (Heatherton et al., 1988). In contrast, the RS was related to intake in men. The significant findings in men suggests that there may be differences between men and women in the relationship between current dieting and TFEQ-R and RS scores. In men, both the RS and the TFEQ-R discriminated current dieters. By contrast, only the TFEQ-R distinguished current dieters in women. Current weight loss dieters did not report lower intake of high-fat foods or higher intake of low-fat foods, a finding discrepant with previous studies (Brightwell et al., 1979; French et al,, 1994; Nowalk & Wing, 1985; Ritt et al., 1979). Limitations in the present study may account for this discrepancy. Specific information about the duration of the current dieting episode was not measured. Thus, the length of time that current dieters had been on their diets, and whether the time period of the current dieting episode was consistent with the period covered by the dietary intake and physical activity measures is not known. In addition, the small number of current dieters may have limited the power to detect differences. Measures of both current and chronic dieting were weakly associated with reported dietary intake and physical activity. This suggests that these dieting measures may reflect weight concern more than behavioral changes that incur a negative energy deficit. Additional research is needed to define more precise measures of dieting, including specific behavioral changes made, their duration, and frequency, in order to further elucidate the potential effects of weight-loss dieting on health outcomes.
S. A. FRENCH et al.
408
REFERENCES Block, G. (1982). A review of validations of dietary assessment methods. American Journal ofEpidemiology, 115,492-505. Block, G., Hartman, A. M., Dresser, C. M., Carroll, M. D., Gannon, .I., & Gardner, L. (1986). A databased approach to diet questionnaire design and testing. American Journal of Epidemiology, l24, 453-469.
Brightwell, D. R., Foster, D., Lee, S., & Naylor, C. S. (1979). Effects of behavioral and pharmacological weight loss programs on nutrient intake. American Journal of Clinical Nutrition, 32, 2005-2008. DiPietro, L., Williamson, D. F., Caspersen, C. J., & Eaker, E. (1993). The descriptive epidemiology of selected physical activities and body weight among adults trying to lose weight: The Behavioral Risk Factor Surveillance System Survey, 1989. International Journal of Obesity, 17, 69-76. Field, A. E., Wolf, A. M., Herzog, D. B., Cheung, L., & Colditz, G. A. (1993). The relationship of caloric intake to frequency of dieting among preadolescent and adolescent girls. Journal of the American Academy
of Child and Adolescent
Psychiatry,
32, 1246-1252.
Forman, M. R., Trowbridge, F. L., Gentry, E. M., Marks, J. S., & Hogelin, G. C. (1986). Overweight adults in the United States: The Behavioral Risk Factor Surveys. American Journal of Clinical Nutrition, 51,253-257. French, S. A., & Jeffery, R. W. (1994). Consequences of dieting to lose weight: Effects on physical and mental health. Health Psychology, 13, 195-212. French, S. A., Jeffery, R. W., & Forster, J. L. (1994). Dieting status and its relationship to weight, dietary intake, and physical activity changes over two years in a working population. Obesity Research, 2, 135-144.
Garner, D. M., 8r Wooley, S. C. (1991). Confronting the failure of behavioral and dietary treatments for obesity. Clinical Psychology Review, 11,729-780. Heatherton, T. F., Herman, C. P., Polivy, J., King, G. A., & McGree, S. T. (1988). The (mis) measurement of restraint: An analysis of conceptual and psychometric issues. Journal of Abnormal Psychology, 97(l), 19-28. Herman, C. P., 8r Polivy, J. (1980). Restrained eating. In A. J. Stunkard (Ed.), Obesity (pp. 208-225). Philadelphia, PA: Saunders. Horm, J., & Anderson, K. (1992). Who in America is trying to lose weight? Annuls of Internal Medicine, 119, 672-676.
Jeffery, R. W., Adlis, S. A., & Forster, J. L. (1991). Prevalence of dieting among working men and women: The Healthy Worker Project. Health Psychology, 10(4), 274-281. Johnson, W. G., Lake, L., & Mahan, J. M. (1983). Restrained eating: Measuring an elusive construct. Addictive
Behaviors,
8, 413-418.
Kirkley, B. G., Burge, J. C., & Ammerman, A. (1988). Dietary restraint, binge eating and dietary behavior patterns. International Journal of Eating Disorders, 7(6), 771-778. Klesges, R. C., Klem, M. L., Epkins, C. C., & Klesges, L. M. (1991). A longitudinal evaluation of dietary restraint and its relationship to changes in body weight. Addictive Behaviors, 16, 363-368. Laessle, R. G., Tuschl, R. J., Kotthaus, B. C., & Pirke, K. M. (1989a). Behavioral and biological correlates of dietary restraint in normal life. Appetite, 12, 83-94. Laessle, R. G., Tuschl, R. J., Kotthaus, B. C., & Pirke, K. M. (1989b). A comparison of the validity of three scales for the assessment of dietary restraint. Journal of Abnormal Psychology, 98(4), 504-507. Lissner, L., Andres, R., Muller, D. C., & Shimokata, H. (1990). Bodyweight variability in men: Metabolic rate, health and longevity. International Journal of Obesity, 14, 373-383. Lowe, M. R. (1993). The effects of dieting on eating behavior: A three-factor model. Psychological Bulletin.
114. 100-121.
Lowe, M. RI, Whitlow, J. W., & Bellwoar, V. (1991). Eating regulation: The role of restraint, dieting and weight. International Journal of Eating Disorders, 10,461-471. Nowalk, M. P., & Wing, R. R. (1985). Changes in nutrient intake of hypertensives during a behavioral weight control program. Addictive Behaviors, 10,357-363. Paffenbarger, R. S., Hyde, R. T., Wing, A. L., & Hsieh, C. (1986). Physical activity, all-cause mortality, and longevity of college alumni. New England Journal of Medicine, 314, 605-613. Paffenbarger, R. S., Wing, A. L., 62 Hyde, R. T. (1978). Physical activity as an index of heart attack risk in college alumni. American Journal of Epidemiology, 108, 161-175. Polivy, J., & Herman, C. P. (1985). Dieting and binging: A causal analysis. American Psychologist, 40, 193-201.
Rand, C. S. W., & Kuldau, J. M. (1991). Restrained eating (weight concerns) in the general population and among students. International Journal of Eating Disorders, 10(6), 699-708. Ritt, R. S., Jordan, H. A., & Levitz, L. S. (1979). Changes in nutrient intake during a behavioral weight control program. Journal of the American Dietetic Association, 74, 325-330. Rosen, J. C., & Poplawski, D. (1987). The validity of self-reported weight loss and weight gain efforts in adolescents. International Journal of Eating Disorders, 6(4), 515-523.
Dieters’ food intake
Ruderman, A. (1986). Dietary restraint: A theoretical and empirical review. Psychological
409
RuNetin, 99,
247-262.
Ruderman, A., & Christensen, H. (1983). Restraint theory and its applicability to overweight individuals. Journal of Abnormal Psychology, 92, 210-215. Statistical Analysis System. (1985). SAS user’s guide: Starisrics (version 5 edition). Cary, NC: SAS Institute, Inc. Stephenson, M. Cl., Levy, A. S., Sass, N. L., & McGarvey, W. E. (1987). 1985 NHIS findings: Nutrition knowledge and baseline data for the weight loss objectives. Public Health Reports, 102(l), 61-67. Striegel-Moore, R. H., Silberstein, L. R., & Rodin, J. (1986). Toward an understanding of risk factors for bulimia. American Psychologist, 41(3), 246-263. Stunkard, A. J. (1981). “Restrained eating”: What it is and a new scale to measure it. In L. A. Cioffi (Ed.), The body weight regulatory system: Normal and disturbed mechanisms (pp. 243-251). New York, NY: Raven Press. Stunkard, A. J., & Messick, S. (1985). The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research, 29, 71-83. Stunkard, A. J., & Wadden, T. A. (1990). Restrained eating and human obesity. Nutrition Reviews, 48, 78-86.
Sunday, S. R., Einhorn, A., & Halmi, K. A. (1992). Relationship of perceived macronutrient and caloric content to affective cognitions about food in eating-disordered, restrained, and unrestrained subiects. American
Journal
of Clinical Nutrition, 55, 362-371.
Tuschl. R. J.. Laessle. R. G.. Platte. P.. & Pirke. K. M. (1990). Differences in food choice freauencies between’restrained and unrestrained eaters. Appetite, 14, 9-13. Tuschl, P., Platte, P., Laessle, R. Cl., Stichler, W., & Pirke, K. M. (1990). Energy expenditure and everyday eating behavior in healthy young women. American Journal of Clinical Nutrition, 52, 81-86.
Van Strien, T., Frijters, J. E., Bergers, G. P., & Defares, P. B. (1986). Dutch eating behavior questionnaire for the assessment of restrained, emotional and external eating behavior. International .Zourna/ of Eating Disorders,
5, 295-315.
Wardle, J. (1986). The assessment of restrained eating. Behnuior Research and Therapy, 24, 213-215. Wardle, J., & Beales, S. (1987). Restraint and food intake: An experimental study of eating patterns in the laboratory and in normal life. Behavior Research and Therapy, 25(3), 179-185. __ Williamson, D. F., Serdula, M. K., Anda, R. F., Levy, A., & Byers, T. (1992). Weight loss attempts in adults: Goals, duration and rate of weight loss. American Journal of Public Health, 82, 1251-1257. Wooley, S. C., & Wooley, 0. W. (1984). Should obesity be treated at all? in A. J. Stunkard & E. Stellar (Eds.), Earing and its disorders (pp. 185-192). New York, NY: Raven Press.
Pergamon
FOOD INTAKE AND PHYSICAL ACTIVITY: A COMPARISON THREE MEASURES OF DIETING SIMONE A. FRENCH,*
ROBERT W. JEFFERY,*
OF
and RENA R. WINGt
School of Public Health, University of Minnesota, Minneapolis, MN ‘/‘Western Psychiatric Institute and Clinic, University of Pittsburgh School of Medicine, Pittsburgh, PA *Division
of Epidemiology,
Abstract - We examined food preferences, dietary intake, and physical activity in dieters and nondieters using three different methods of dieting classification. One hundred three women and 99 men completed the cognitive restraint subscale of Stunkard and Messick’s (1985) Three Factor Eating Questionnaire (TFEQ-R) and Herman and Polivy’s (1980) Restraint Scale (RS), and answered questions about their &rent efforts to lose or maintain weight. Women identifying themselves as currently dieting to lose weight reported expending twice as much energy in physical activity compared to those reporting that they were either currently dieting to maintain weight or not dieting. There were no significant differences in dietary intake or physical activity by current dieting status in men. Women chronic dieters identified by high TFEQ-R scores reported lower total calorie intake and less frequent sweets consumption than women with low TFEQ-R scores. In men, those with high TFEQ-R scores reported consuming a greater percent of calories from protein and carbohydrate, less beef, pork, whole milk, and sweets. In women, the RS did not distinguish dieters from nondieters on any measure. In men, the RS results were similar to those from the TFEQ-R. These results show that current measures of dieting are only weakly related to behaviors thought to be indicative of dieting. Future research must develop more precise measures of dieting in order to examine the relationship between self-reports of dieting and behaviors thought to be related to dieting.
Dieting to lose weight is a widespread practice in the United States (French & Jeffery, 1994; Jeffery, Adlis, & Forster, 1991; Williamson, Serdula, Anda, & Byers, 1992). Concern has recently arisen over the possibility that dieting may be more harmful to health than helpful (Garner & Wooley, 1991; Lissner, Andres, Muller, & Shimokata, 1990; Polivy & Herman, 1985; Striegel-Moore, Silberstein, & Rodin, 1986; Wooley & Wooley, 1984). A problem in assessing the health benefits and hazards associated with dieting efforts has been identifying a valid measure of the construct of “dieting.” Epidemiologic studies have typically assessed dieting efforts using a single direct question about whether the respondent is currently trying to lose weight (DiPietro, Williamson, Caspersen, & Eaker, 1993; Horm & Anderson, 1992; Stephenson, Levy, Sass, & McGarvey, 1987; Williamson et al., 1992) or is currently dieting to lose weight (Jeffery et al., 1991; Forman, Trowbridge, Gentry, Marks, & Hogelin, 1986; Rand & Kuldau, 1991). Psychologists investigating dieting behavior have developed multiitem scales that incorporate additional behavioral and psychological characteristics of chronic dieting efforts (Herman & Polivy, 1980; Stunkard & Messick, 1985; Van Strien, Frijters, Bergers, & Defares, 1986). The Restraint Scale (RS) has been hypothesized to identify unsuccessful dieters (Herman & Polivy, 1980; Heatherton, Herman, Polivy, King, & McGree, 1988), while the Three Factor Eating Questionnaire (TFEQ) has been hypothesized to identify successful dieters (Heatherton et al., Research supported by grants 5ROl HL 41332,5ROl HL 41330, and 5T32 HL 07328 from the National Heart, Lung and Blood Institute to Robert W. Jeffery and Rena R. Wing. Requests for reprints should be sent to Simone A. French, PhD, Division of Epidemiology, School of Public Health, University of Minnesota, 1300 South Second Street, Suite 300, Minneapolis, MN 55454-1015. 401
402
S. A. FRENCH
et al.
1988; Stunkard & Messick, 1985). An examination of item content suggests that the RS focuses on previous dieting and weight fluctuations, while the cognitive restraint subscale of the TFEQ (TFEQ-R) assesses habitual behaviors related to restriction of dietary intake. While few studies have provided empirical data on the validity of self-reported dieting behaviors, some research suggests that self-reports of dieting to lose weight are correlated with reported dietary and physical activity patterns. In a study involving a population-based sample of working adults, self-reports of dieting to lose weight were cross-sectionally related to lower self-reported intake of several highfat/high-calorie foods, including sweets, meat, soft drinks, french fries, and dairy products. Dieting was also related to higher self-reported levels of high-intensity physical activity (French, Jeffery, & Forster, 1994). In addition, individuals who reported initiating a weight-loss diet reported decreased consumption of sweets, soft drinks, and french fries, and increased frequency of moderate and high intensity physical activity (French et al., 1994). Studies that have examined the dietary composition of participants in weight-loss programs show that people on self-selected calorie-restricted diets spontaneously reduce consumption of high-fat foods such as meat, fat, dairy products, high-sugar and snack foods (Brightwell, Foster, Lee, & Naylor, 1979; Nowalk & Wing, 1985; Ritt, Jordan, & Levitz, 1979). Finally, two studies of high school students found that self-reports of trying to lose weight were related to lower reported total energy intake (Field, Wolf, Herzog, Cheung, & Colditz, 1993), and self-reports of greater caloric restriction and meal skipping (Rosen & Poplawski, 1987). Self-reported physical activity was unrelated to dieting frequency in the former study (Field et al., 1993) and positively associated with dieting efforts in the latter study (Rosen & Poplawski, 1987). Studies examining the dietary intakes of chronic dieters identified by the RS suggest that this measure is unrelated to total calorie intake (Heather-ton et al., 1988; Kirkley, Burge, & Ammerman, 1988; Laessle, Tuschl, Kotthaus, & Pirke, 1989b). However, chronic dieters identified by the TFEQ-R report lower total calorie intake than nondieters (Heatherton et al., 1988; Kirkley et al., 1988; Laessle et al., 1989a; 1989b; Tuschl, Platte, Laessle, Stichler, & Pirke, 1990; Van Strien et al., 1986; Wardle & Beales, 1987). Analyses of the association between the TFEQ-R scale and specific food preference and consumption indicate that chronic dieters identified by the TFEQ-R do not differ from nondieters in their intake of meat, vegetables, fruits, starches, sweets, breads, cakes, or margarine, but they report less preference for and consumption of oils, mayonnaise, butter, and high-fat dairy products (Laessle et al., 1989a; Sunday, Einhorn, & Halmi, 1992; Tuschl, Laessle, Platte, & Pirke, 1990). These conflicting results may be related to at least two factors that differ between studies. The first is study population. Studies that have found differences between dieters and nondieters in specific food patterns have been comprised of either population-based adult samples, or adults who are participating in weight-loss programs. The studies reporting fewer differences between dieters and nondieters have been done primarily with young, female, college students, a considerably more homogeneous group whose dietary habits may have less interindividual variation. The second factor that differs between the studies is the method used to identify dieters. The studies that have found differences in specific dietary intake have used a direct questioning strategy to assess current dieting efforts, while those that have found fewer differences have relied upon composite scales such as the RS and the TFEQ-R that assess chronic dieting efforts. While chronic dieters may be more likely to be dieting at any given time than non-chronic dieters, they may not have been actively
Dieters’ food intake
403
dieting during the period that their food intake was measured (Lowe, 1993; Lowe, Whitlow, Jz Bellwoar, 1991). Composite scales may also have lower correlations with eating behavior than a direct, single-item measure of dieting simply because these scales assess things other than eating (e.g., prior weight loss, weight concerns). In addition, the correlations may be attentuated because the time period covered by the composite scales differs from that of the single, direct question. To date, no studies have examined relationships between multiple measures of dieting and dietary intake in the same group of individuals. The present study was conducted to examine the relationship between three measures of dieting and macronutrient intake, specific food intake, food preferences, and physical activity levels. It was hypothesized that dieters would report lower total calorie intake, lower percent of total calories from fat, lower intake of high-fat foods such as sweets, meat, fats, and high-fat dairy products, higher intake of low-fat foods such as salads, fruits, and vegetables, and higher levels of physical activity than nondieters. In addition, based on previous research, it was hypothesized that measures of current dieting would be more strongly related to intake of low-fat foods and avoidance of high-fat foods than measures of chronic dieting.
METHOD
Subjects and procedure One hundred three women and 99 men between the ages of 25 and 62 participated in the study. Participants were recruited from the university community, and were required to be living with a partner, friend, or spouse/family member in order to participate in the study. Subjects were recruited through newspaper advertisements and were paid $20.00 to participate in a questionnaire study on eating and exercise habits. During the first session, height and weight were measured and a packet of questionnaires was completed by each participant. Subjects returned for a second questionnaire session 2 weeks later during which some of the questionnaires were completed a second time. Measures Current dieting status. Current dieting status was determined by responses to two questions. Current dieters to lose weight were comprised of subjects who answered “yes” to the question “Are you currently dieting to lose weight?“. Current dieters to maintain weight were comprised of subjects who answered “yes” to the question “Are you currently dieting to maintain your present weight?“. Nondieters were classified as those who reported that they were not currently dieting to lose or maintain weight. The 2-year test-retest reliability of the single item “Are you currently dieting to lose weight” has been shown to be 0.32 in a sample of working adults (French et al., 1994). This indicates that current dieting efforts are related to future dieting efforts, although the relationship is not strong. Chronic dieting status. Chronic dieting status was assessed using two different instruments: the Restraint Scale (RS; Herman & Polivy, 1980) and the Three Factor Eating Questionnaire’s Cognitive Restraint subscale (TFEQ-R; Stunkard & Messick, 1985). The RS is hypothesized to be related to a history of unsuccessful dieting, while the TFEQ is thought to identify successful dieters (Heatherton, Herman,
404
S. A. FRENCH et al.
Polivy, King, & McGree, 1988). The RS is valid in discriminating those who will overeat in laboratory settings under certain conditions (Heatherton et al., 1988; Ruderman, 1986). Its 2-year test-retest reliability is high (r = 0.74), however it has been shown to be less stable among unrestrained eaters than restrained eaters (r = 0.37 vs. r = 0.58, respectively; Klesges, Klem, Epkins, & Klesges, 1991). The TFEQ-R is valid in discriminating those who consume fewer calories in naturalistic settings (Heatherton et al., 1988; Kirkely et al., 1988; Laessle et al., 1989a; Tuschl, Platte, Laessle, Stichler, & Pirke, 1990; Van Strien et al., 1986; Wardle & Beales, 1987). Restrained eaters identified by the TFEQ-R also show depleted fat cell size, which indirectly validates their reports of caloric restriction (Stunkard, 1981). The lmonth test-retest reliability has been shown to be high (r = 0.98; Stunkard & Messick, 1985). Dietary intake. Usual dietary intake was measured using the Block Food Frequency Questionnaire (FFQ; Block et al., 1986). Respondents indicate their typical serving size (small, medium, or large) and frequency of consumption during the previous 6 months (daily, weekly, monthly, yearly, or rarely/never) for 64 food items or groups of items. Daily macronutrient intake and frequency of intake of individual food items and food groupings are calculated. The FFQ has demonstrated reliability and validity (Block, 1982; Block et al., 1986). For this report, total calories per day and percent of calories from fat, protein, alcohol, carbohydrates, and sweets were examined. In addition, frequency of consumption of the food groups beef, cheese, butter, hot dogs, pork, whole milk, sweets, salad, and fruit were examined. Low-fat food intake. In order to specifically examine low-fat food intake, a composite index of low-fat food consumption was created by averaging intake of 22 items from the Block FFQ containing less than 17% calories from fat. Item analysis reduced this pool to 17 items: apples, cantaloupe, oranges, orange juice, grapefruit, other fruit juices, beans, tomatoes, broccoli, spinach, greens, carrots/mixed vegetables, green salad, sweet potatoes, rice, dark bread, and fortified cereal. The Cronbath’s alpha coefficient for this scale is .72, indicating high internal reliability. Low-fat food preferences. In order to specifically examine low-fat food preferences, a composite index of low-fat food preference was created in a similar manner to the low-fat food intake index described above. Respondents indicated their preference for each food listed on the Block FFQ by marking a number corresponding to their preference (1 = don’t like at all, 5 = like very much). Item analysis of the 22 foods containing less than 17% calories from fat reduced the set of items to 21: apples, cantaloupe, oranges, orange juice, grapefruit, other fruit juices, beans, tomatoes, broccoli, spinach, greens, carrots/mixed vegetables, green salad, sweet potatoes, rice, dark bread, fortified cereal, high fiber cereal, other cold cereal, cooked cereal, and skim milk. The Cronbach’s alpha coefficient for this scale is .83, indicating high internal reliability. Physical activity. Habitual physical activity was measured using the Paffenbarger physical activity recall (Paffenbarger, Wing, & Hyde, 1978). The Paffenbarger recall estimates average weekly total calories expended in physical activity over the previous 6 months. Respondents indicate the number of flights of stairs climbed, city blocks walked, and participation in sports or recreational activities (average weekly frequency and duration per session). Concurrent validity has been demonstrated by
Dieters’ food intake
405
the measure’s correlation in predicted directions with body weight (Paffenberger, Hyde, Wing, & Hsieh, 1986). Predictive validity has been demonstrated by the measure’s prospective relationship with coronary heart disease morbidity and mortality (Paffenbarger et al., 1978; Paffenbarger et al., 1986). Body Mass Index. Body Mass Index (BMI) was calculated as weight in kilograms/ height in meters2. Weight was measured in street clothing without shoes, using a balance beam scale. Height was measured without shoes using a wall-mounted ruler and head block. Statistical analysis Prior to analysis, the specific food items, food groups, and physical activity scores were transformed to normalize their skewed distributions. The FFQ items were log transformed, and the Paffenbarger scores were square root transformed. Multiple regression analysis was conducted using the Statistical Analysis System’s PROC GLM (SAS, 1985). Separate sets of analyses were performed for each dieting measure. Current dieting status was entered as a three-level categorical independent variable (weight-loss dieting, weight-maintenance dieting, not dieting). The RS and TFEQ-R were entered as continuous independent variables. Mean comparisons in the tables are based on sex-specific median splits. Age and BMI were entered as covariates in each analysis. Separate analyses were conducted for men and women. RESULTS
Sample characteristics The average age of the men and women in the study was 35.8 (SD = 8.2) years and 36.8 (SD = 9.3) years, respectively. Three quarters of the men and half of the women had completed a college degree, and 60% were currently married. About 44% of the men and 19% of the women were in professional or administrative occupations, and 40% of the men and 75% of the women were in clerical positions. Average BMI for men and women was 26.7 (SD = 5.2) and 25.4 (SD = 4.6), respectively. More women than men reported currently dieting to lose or maintain weight (37% vs. 22%). Women scored higher than men on both the RS (M = 14.6, SD = 5.5, vs. M = 11.O, SD = 5.0) and on the TFEQ-R (M = 9.1, SD = 4.2, vs. M = 5.9, SD = 4.2). These values are comparable to RS and TFEQ-R scores reported in college students (RS: Klesges et al., 1991; Ruderman & Christensen, 1983; Wardle, 1986) and adults (RS: Johnson, Lake, & Mahan, 1984; TFEQ-R: Stunkard & Messick, 1985; Stunkard 8z Wadden, 1990). In the present sample, the partial correlation (age and BMI partialled) between the RS and the TFEQ-R was 0.41 in women and 0.38 in men. In men, 40% of high TFEQR (above sex-specific median) and 9% of low TFEQ-R (below sex-specific median) scorers were currently dieting to lose or maintain weight. In women, the corresponding numbers were 59% and 16%, respectively. In men, 38% of high- and 8% of lowRS scorers reported currently dieting to lose or maintain weight. In women, the corresponding numbers were 44% and 31%, respectively. Current dieting status Table 1 shows age- and BMI-adjusted means in women and men for significant comparisons on total calories, intake of macronutrients and specific foods, and physical activity levels by dieting status. Twenty-three women reported currently dieting
406
S. A. FRENCH et al.
Table 1. Age and Body Mass Index - Adjusted means and standard errors for dietary intake and physical activity by dieting status Current dieting status Weight-loss diet Not dieting Women N Physical activity (kcal/wk) Men N % Protein calories % Fat calories Three Factor
lS;;.O
77 16.8 (l.O)A 34.7 (1.1)
Cognitive
Restraint
Low restraint
51 1446.8 (72.3) 11.6 (1.5) 8.9 (1.7)
51 1650.8 (72.3) 15.9 (1.5) 16.0 (1.7)
50.1 1.9 0.4 0.2 11.8
56 16.6 45.7 2.8 0.9 1.7 14.7
(1.4) (0.3) (0.2) (0.6) (1.7)
& Polivy
Restraint
(1.2) (1.2) (0.3) (0.1) (0.5) (1.5)
P
.03 .02 .0002 .003 .04 .Ol .03 .009 .005
Scale
Low restraint
47 1.9 (0.3) 0.4 (0.2) 0.4 (0.6)
All food groups are in servings per week. Different superscripts cantly different means (p < .05). Note.
.02 .03
Subscale
High restraint
High restraint Men N Beef Pork Whole milk
.0002
il.8 (2.6)B 41.6 (2.8)*
43 19.4 (1.4)
Herman
15 1444.9 (346.8)
::.5 (2.9)*,s 30.9 (3.0)
Questionnaire
Women N Total calories % Sweets calories Sweets Men N % Protein calories % Carbohydrate calories Beef Pork Whole milk Sweets
p
23 2983.5 (281.1)*
(165.2)
Eating
Maintenance diet
52 2.8 (0.3) 0.9 (0.1) 1.7 (0.6)
P
.04 .005 ,002
within rows indicate signifi-
to lose weight, 15 reported currently dieting to maintain weight, and 65 reported not currently dieting. In women, current dieters to lose weight reported expending almost twice as many calories from physical activity as those dieting to maintain weight or not dieting. Eleven men were currently dieting to lose weight, 11 were currently dieting to maintain weight, and 77 were not currently dieting. In men, current dieters to maintain weight reported significantly higher percent calories from protein and fat than dieters to lose weight and nondieters. None of the other dietary intake or physical activity measures differed by current dieting status. Chronic dieting status Three reported and less None of
Factor Eating Questionnaire. Women who scored significantly lower total calorie intake, lower percent frequent sweets consumption in comparison with the other measures differed significantly by TFEQ-R
high on the TFEQ-R calories from sweets, low TFEQ-R women. score. Men with high
Dieters’ food intake
407
TFEQ-R scores reported a significantly greater percent of calorie intake from protein and carbohydrate, and less frequent consumption of beef, pork, whole milk, and sweets than men with low TFEQ-R scores. None of the other measures differed significantly by TFEQ-R score. Restraint score. Women who scored high on Herman & Polivy’s (1980) RS scale did not differ significantly from unrestrained women on any of the dietary-intake or the physical-activity measures. In men, those with high RS scores reported significantly less frequent consumption of beef, pork, and whole milk than those with low RS scores. No other dietary intake or physical activity measures were significantly different by RS score. DISCUSSION
This study is the first to examine the behavioral correlates of dieting using three different measures of dieting: direct questions about current dieting and two scales measuring chronic dieting. It was initially hypothesized that dieters would report a lower intake of high-fat foods, higher intake of low-fat foods, and higher levels of physical activity than nondieters. The TFEQ-R was related to lower total calorie intake, lower percent of calories from sweet foods, and lower frequency of sweets consumption in women. In men, the TFEQ-R was related to a greater percent calories from protein and carbohydrates, lower frequency of beef, pork, whole milk, and sweets. These results are consistent with previous studies using the TFEQ-R that reported a lower total calorie intake and lower consumption of butter and high fat dairy products. The RS was unrelated to intake in women, which is consistent with previous findings and supports the argument that the RS measures unsuccessful dieting, while the TFEQ measures successful dieting (Heatherton et al., 1988). In contrast, the RS was related to intake in men. The significant findings in men suggests that there may be differences between men and women in the relationship between current dieting and TFEQ-R and RS scores. In men, both the RS and the TFEQ-R discriminated current dieters. By contrast, only the TFEQ-R distinguished current dieters in women. Current weight loss dieters did not report lower intake of high-fat foods or higher intake of low-fat foods, a finding discrepant with previous studies (Brightwell et al., 1979; French et al,, 1994; Nowalk & Wing, 1985; Ritt et al., 1979). Limitations in the present study may account for this discrepancy. Specific information about the duration of the current dieting episode was not measured. Thus, the length of time that current dieters had been on their diets, and whether the time period of the current dieting episode was consistent with the period covered by the dietary intake and physical activity measures is not known. In addition, the small number of current dieters may have limited the power to detect differences. Measures of both current and chronic dieting were weakly associated with reported dietary intake and physical activity. This suggests that these dieting measures may reflect weight concern more than behavioral changes that incur a negative energy deficit. Additional research is needed to define more precise measures of dieting, including specific behavioral changes made, their duration, and frequency, in order to further elucidate the potential effects of weight-loss dieting on health outcomes.
S. A. FRENCH et al.
408
REFERENCES Block, G. (1982). A review of validations of dietary assessment methods. American Journal ofEpidemiology, 115,492-505. Block, G., Hartman, A. M., Dresser, C. M., Carroll, M. D., Gannon, .I., & Gardner, L. (1986). A databased approach to diet questionnaire design and testing. American Journal of Epidemiology, l24, 453-469.
Brightwell, D. R., Foster, D., Lee, S., & Naylor, C. S. (1979). Effects of behavioral and pharmacological weight loss programs on nutrient intake. American Journal of Clinical Nutrition, 32, 2005-2008. DiPietro, L., Williamson, D. F., Caspersen, C. J., & Eaker, E. (1993). The descriptive epidemiology of selected physical activities and body weight among adults trying to lose weight: The Behavioral Risk Factor Surveillance System Survey, 1989. International Journal of Obesity, 17, 69-76. Field, A. E., Wolf, A. M., Herzog, D. B., Cheung, L., & Colditz, G. A. (1993). The relationship of caloric intake to frequency of dieting among preadolescent and adolescent girls. Journal of the American Academy
of Child and Adolescent
Psychiatry,
32, 1246-1252.
Forman, M. R., Trowbridge, F. L., Gentry, E. M., Marks, J. S., & Hogelin, G. C. (1986). Overweight adults in the United States: The Behavioral Risk Factor Surveys. American Journal of Clinical Nutrition, 51,253-257. French, S. A., & Jeffery, R. W. (1994). Consequences of dieting to lose weight: Effects on physical and mental health. Health Psychology, 13, 195-212. French, S. A., Jeffery, R. W., & Forster, J. L. (1994). Dieting status and its relationship to weight, dietary intake, and physical activity changes over two years in a working population. Obesity Research, 2, 135-144.
Garner, D. M., 8r Wooley, S. C. (1991). Confronting the failure of behavioral and dietary treatments for obesity. Clinical Psychology Review, 11,729-780. Heatherton, T. F., Herman, C. P., Polivy, J., King, G. A., & McGree, S. T. (1988). The (mis) measurement of restraint: An analysis of conceptual and psychometric issues. Journal of Abnormal Psychology, 97(l), 19-28. Herman, C. P., 8r Polivy, J. (1980). Restrained eating. In A. J. Stunkard (Ed.), Obesity (pp. 208-225). Philadelphia, PA: Saunders. Horm, J., & Anderson, K. (1992). Who in America is trying to lose weight? Annuls of Internal Medicine, 119, 672-676.
Jeffery, R. W., Adlis, S. A., & Forster, J. L. (1991). Prevalence of dieting among working men and women: The Healthy Worker Project. Health Psychology, 10(4), 274-281. Johnson, W. G., Lake, L., & Mahan, J. M. (1983). Restrained eating: Measuring an elusive construct. Addictive
Behaviors,
8, 413-418.
Kirkley, B. G., Burge, J. C., & Ammerman, A. (1988). Dietary restraint, binge eating and dietary behavior patterns. International Journal of Eating Disorders, 7(6), 771-778. Klesges, R. C., Klem, M. L., Epkins, C. C., & Klesges, L. M. (1991). A longitudinal evaluation of dietary restraint and its relationship to changes in body weight. Addictive Behaviors, 16, 363-368. Laessle, R. G., Tuschl, R. J., Kotthaus, B. C., & Pirke, K. M. (1989a). Behavioral and biological correlates of dietary restraint in normal life. Appetite, 12, 83-94. Laessle, R. G., Tuschl, R. J., Kotthaus, B. C., & Pirke, K. M. (1989b). A comparison of the validity of three scales for the assessment of dietary restraint. Journal of Abnormal Psychology, 98(4), 504-507. Lissner, L., Andres, R., Muller, D. C., & Shimokata, H. (1990). Bodyweight variability in men: Metabolic rate, health and longevity. International Journal of Obesity, 14, 373-383. Lowe, M. R. (1993). The effects of dieting on eating behavior: A three-factor model. Psychological Bulletin.
114. 100-121.
Lowe, M. RI, Whitlow, J. W., & Bellwoar, V. (1991). Eating regulation: The role of restraint, dieting and weight. International Journal of Eating Disorders, 10,461-471. Nowalk, M. P., & Wing, R. R. (1985). Changes in nutrient intake of hypertensives during a behavioral weight control program. Addictive Behaviors, 10,357-363. Paffenbarger, R. S., Hyde, R. T., Wing, A. L., & Hsieh, C. (1986). Physical activity, all-cause mortality, and longevity of college alumni. New England Journal of Medicine, 314, 605-613. Paffenbarger, R. S., Wing, A. L., 62 Hyde, R. T. (1978). Physical activity as an index of heart attack risk in college alumni. American Journal of Epidemiology, 108, 161-175. Polivy, J., & Herman, C. P. (1985). Dieting and binging: A causal analysis. American Psychologist, 40, 193-201.
Rand, C. S. W., & Kuldau, J. M. (1991). Restrained eating (weight concerns) in the general population and among students. International Journal of Eating Disorders, 10(6), 699-708. Ritt, R. S., Jordan, H. A., & Levitz, L. S. (1979). Changes in nutrient intake during a behavioral weight control program. Journal of the American Dietetic Association, 74, 325-330. Rosen, J. C., & Poplawski, D. (1987). The validity of self-reported weight loss and weight gain efforts in adolescents. International Journal of Eating Disorders, 6(4), 515-523.
Dieters’ food intake
Ruderman, A. (1986). Dietary restraint: A theoretical and empirical review. Psychological
409
RuNetin, 99,
247-262.
Ruderman, A., & Christensen, H. (1983). Restraint theory and its applicability to overweight individuals. Journal of Abnormal Psychology, 92, 210-215. Statistical Analysis System. (1985). SAS user’s guide: Starisrics (version 5 edition). Cary, NC: SAS Institute, Inc. Stephenson, M. Cl., Levy, A. S., Sass, N. L., & McGarvey, W. E. (1987). 1985 NHIS findings: Nutrition knowledge and baseline data for the weight loss objectives. Public Health Reports, 102(l), 61-67. Striegel-Moore, R. H., Silberstein, L. R., & Rodin, J. (1986). Toward an understanding of risk factors for bulimia. American Psychologist, 41(3), 246-263. Stunkard, A. J. (1981). “Restrained eating”: What it is and a new scale to measure it. In L. A. Cioffi (Ed.), The body weight regulatory system: Normal and disturbed mechanisms (pp. 243-251). New York, NY: Raven Press. Stunkard, A. J., & Messick, S. (1985). The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research, 29, 71-83. Stunkard, A. J., & Wadden, T. A. (1990). Restrained eating and human obesity. Nutrition Reviews, 48, 78-86.
Sunday, S. R., Einhorn, A., & Halmi, K. A. (1992). Relationship of perceived macronutrient and caloric content to affective cognitions about food in eating-disordered, restrained, and unrestrained subiects. American
Journal
of Clinical Nutrition, 55, 362-371.
Tuschl. R. J.. Laessle. R. G.. Platte. P.. & Pirke. K. M. (1990). Differences in food choice freauencies between’restrained and unrestrained eaters. Appetite, 14, 9-13. Tuschl, P., Platte, P., Laessle, R. Cl., Stichler, W., & Pirke, K. M. (1990). Energy expenditure and everyday eating behavior in healthy young women. American Journal of Clinical Nutrition, 52, 81-86.
Van Strien, T., Frijters, J. E., Bergers, G. P., & Defares, P. B. (1986). Dutch eating behavior questionnaire for the assessment of restrained, emotional and external eating behavior. International .Zourna/ of Eating Disorders,
5, 295-315.
Wardle, J. (1986). The assessment of restrained eating. Behnuior Research and Therapy, 24, 213-215. Wardle, J., & Beales, S. (1987). Restraint and food intake: An experimental study of eating patterns in the laboratory and in normal life. Behavior Research and Therapy, 25(3), 179-185. __ Williamson, D. F., Serdula, M. K., Anda, R. F., Levy, A., & Byers, T. (1992). Weight loss attempts in adults: Goals, duration and rate of weight loss. American Journal of Public Health, 82, 1251-1257. Wooley, S. C., & Wooley, 0. W. (1984). Should obesity be treated at all? in A. J. Stunkard & E. Stellar (Eds.), Earing and its disorders (pp. 185-192). New York, NY: Raven Press.