Self-Weighing in Weight Gain Prevention and Weight Loss Trials

Jennifer A. Linde, Ph.D., Robert W. Jeffery, Ph.D., and Simone A. French, Ph.D.
University of Minnesota, Twin Cities

Nicolaas P. Pronk, Ph.D. and Raymond G. Boyle, Ph.D.

HealthPartners Research Foundation

ABSTRACT Background: Although self-monitoring is a central tenet of behavioral approaches to changing health behavior, clinical and public health recommendations for better controlling body weight do not emphasize weight self-monitoring. Purpose: The objective was to determine whether more frequent self-weighing exerts a positive influence on weight loss or weight gain prevention. Methods: This study examined cross-sectional and longitudinal associations between self-weighing frequency and weight in two distinct groups: 1,226 adults who were enrolled in a weight gain prevention trial, and 1,800 adults who were enrolled in a weight loss trial. Results: Although the samples differed significantly in weight and baseline demographic characteristics, the distribution of baseline weighing frequencies did not differ by study. In both groups, more frequent self-weighing at baseline was associated with greater age, lower fat intake, White ethnicity, current nonsmoking status, a greater history of dieting to lose weight, and lower current body mass index. Despite similar weighing instructions, differential patterns of weighing frequency over time were observed: Weight loss dieters increased weighing over time regardless of treatment group (control or intervention), whereas weight gain preventers decreased weighing over time in the control group and increased weighing over time in intervention groups. Most important, higher weighing frequency was associated with greater 24month weight loss or less weight gain. Conclusions: Results support the idea that daily weighing is valuable to individuals trying to lose weight or prevent weight gain. Daily self-weighing should be emphasized in clinical and public health messages about weight control. Experimental studies on the effects of weighing frequency in these contexts are recommended. (Ann Behav Med 2005, 30(3):210216)

INTRODUCTION In the United States, the number of adults who are classified as overweight (body mass index [BMI] 25 kg/m2) or obese (BMI 30 kg/m2) has dramatically increased in the past 25 years (1). Similarly, interest in the health benefits of weight conThis research was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants 1R01-DK45361 and 1R01-DK53826. Reprint Address: J. A. Linde, Ph.D., Division of Epidemiology, School of Public Health, University of Minnesota, Twin Cities Campus, 1300 South Second Street, Suite 300, Minneapolis, MN 55454-1015. E-mail: linde@epi.umn.edu 2005 by The Society of Behavioral Medicine.

trol has increased (2). Behavioral control has been identified as a critical element of successful weight control, and numerous programs have been developed that provide guidelines for modifying behavior to facilitate weight control (2,3). Daily self-monitoring of important behaviors (e.g., eating, physical activity) has been emphasized within behavioral programs as a useful tool for weight control (3). It is interesting to note that self-assessment of body weight on a daily basis is not as widely accepted and recommended in weight control programs. Many clinical programs recommend self-weighing only weekly, and public health recommendations for weight control developed by the U.S. Department of Agriculture and the Centers for Disease Control do not recommend self-weighing at all (4,5). In population surveys, regular self-weighing has been identified as one self-monitoring strategy often employed by those successful at losing weight (6,7). Weekly weight checks are a part of organized group weight loss programs (e.g., Weight Watchers), and weight self-monitoring instructions have been incorporated into cognitive-behavioral obesity treatment programs as well (8,9). However, studies evaluating the specific effects of weighing frequency are rare, and those that exist are not directly applicable to the domain of obesity treatment. For example, the issue of weighing frequency was studied in the context of inpatient treatment for anorexia nervosa (10) or in normal weight or college student volunteers (11,12). As a step toward clarifying the value of weight self-monitoring recommendations in clinical weight control programs and in a broader public health context, we evaluated self-weighing practices over time in two distinct populations: a clinically obese sample enrolled in a weight loss program and a more general sample of individuals volunteering for a weight gain prevention study. The purpose was to develop a clearer understanding of how different frequencies of self-monitoring of body weight may facilitate or impede weight control efforts. Frequency of self-weighing was examined in relation to body weight at baseline and to change in weight over time. It was hypothesized that more frequent weighing would be associated with lower body weight or with better weight outcomes (greater weight loss or reduced weight gain). Other weight control and health-related behaviors were also considered here, to determine associations of self-weighing within the health behavior domain, with the hypothesis that more frequent self-weighing would be associated with engagement in healthy behaviors (e.g., fat intake regulation, greater physical activity, or nonsmoking). As regular weight monitoring is more likely to be familiar to those with weight control experience, we expected to observe a positive association between self-weighing frequency and previous participation in weight control. Finally, to elucidate poten-

210

Volume 30, Number 3, 2005

tial individual differences in adherence to self-weighing behaviors, demographic associations with weighing frequency were considered. METHOD Participants Data were obtained from Pound of Prevention (POP) (13), a 3-year weight gain prevention trial; and Weigh-to-Be (WTB) (14), a 2-year weight loss trial. Both studies were approved by the University of Minnesota Institutional Review Board. POP was conducted in collaboration with four local health departments in the MinneapolisSt. Paul metropolitan area, and WTB was conducted in collaboration with a large Minnesota managed-care organization. POP was composed of 998 women and 228 men recruited from the community by telephone solicitation, newspaper advertising, University of Minnesota employee mailings, and targeted recruiting of low-income women in an assistance program. There was no body weight entry criterion. Once enrolled, participants were randomized to one of three treatment conditions: an educational weight control intervention, the educational intervention plus a lottery incentive for returning monthly self-monitoring postcards, or a minimal contact control condition. Questionnaires were administered, and weight was measured at baseline and yearly for 3 years following randomization. Details of the trial are reported elsewhere (13). WTB was composed of 1,292 women and 508 men recruited from the membership roster of a large managed-care organization by posters and flyers in clinics, direct mail, and physician referrals. A BMI of 27 kg/m2 was set as the study entry criterion for weight. Once enrolled, participants were randomized to one of three treatment conditions: a telephone-based weight loss intervention, a mail-based weight loss intervention, or a usual care control condition. Participants were assessed in person at baseline; by mail at 6, 12, and 18 months; and again in person at a 2-year follow up. Details of the trial are reported elsewhere (14). Measures Demographics. At baseline, participants reported gender, age in years, education level (some college or more vs. no college), ethnicity (White vs. non-White), marital status (married vs. not married), smoking history (yes or no), current smoking status (yes or no), and whether they had ever dieted to lose weight (yes or no). Fat intake. In POP, fat intake in total grams per day was calculated from the 60-item Block Food Frequency Questionnaire, a validated food frequency assessment instrument (15,16). In WTB, a fat intake index score was calculated from the 15-item Block Screening Questionnaire for Fat (17). Higher scores indicate greater fat intake; scores above 22 indicate a diet that is high in fat. The measure has demonstrated reliability and validity in assessing relative fat intake (18). Exercise. In POP, walking was assessed by self-reported number of walking episodes per week. In WTB, walking was as-

Self-Weighing Frequencies

211

sessed by self-reported city blocks walked per day, as recorded on the Paffenbarger Activity Questionnaire (19). BMI and BMI change. In POP, trained staff measured height at baseline and weight at baseline, 12 months, and 24 months. BMI was calculated using the formula (weight in kg height in meters2), and BMI changes at 12 and 24 months were calculated by subtracting 12- and 24-month BMI from baseline BMI values. In WTB, trained staff measured height and weight at baseline and weight at 24 months. Twelve-month weights were obtained by mailed questionnaires and were adjusted by linear regression model weights derived from baseline weight data for men and women (measured weight = 1.02 [self-report weight] 1.22 for men, R2 = .91; measured weight = 1.02 [self-report weight] 0.14 for women, R2 = .95) to account for underreporting bias (20). BMI changes at 12 and 24 months were calculated as described for POP. Self-weighing frequency. In POP, the item How frequently do you weigh yourself? was asked at baseline, 12, and 24 months. Response options were never, about once a year or less, every couple months, every month, every week, and every day. In WTB, the item How often do you weigh yourself? was asked at baseline, 12, and 24 months. Response options differed from POP: never, about once a year or less, every couple of months, every month, every week, every day, and more than once a day. Responses were collapsed to form five categories: never, every other month, every month, every week, and every day. Weighing Instructions by Sample Weighing instructions were similar in their emphasis on weekly weighing, although the tone differed. In POP, instructions stated, Weighing yourself is an excellent monitoring activity. Weigh yourself at least once a week. In WTB, instructions stated, Some people find it helpful to self-monitor their weight regularly. If you choose to use this tool, we recommend that you weigh yourself no more than once a week. Data Analysis Data were retrieved from university archives and were analyzed using the Statistical Analysis System, Version 8.2 (21). Means and standard deviations were calculated for continuous baseline demographic variables, and percentages were calculated for categorical baseline demographic variables. Baseline characteristics were compared using t-test and chi-square analyses. Separate general linear models were used to compare baseline, 12-month, and 24-month BMI values by weighing frequency category. Repeated measures multivariate analysis of variance (MANOVA) was used to examine changes in weighing frequency over time and by treatment group. General linear models were used to examine changes in BMI at 12 and 24 months by weighing frequency; as change scores are not typically independent of baseline levels, baseline BMI and baseline weighing frequency were included in these analyses to allow for better statistical control in our examinations of weighing frequency and weight changes (22). Effect size correlations (rY) are reported (small .10, medium = .100.24, large > .24) (23).

212

Linde et al.
RESULTS

Annals of Behavioral Medicine

significantly associated with weekly weighing at baseline (p = .047 to < .0001, rY = .06.14). In addition, being married, not smoking, or having a history of dieting were associated with weekly weighing, whereas ethnic minority status was associated with less than weekly weighing (p = .044 to < .0001, rY = .06.22). In WTB, older age and lower fat intake were again associated with weekly weighing (p < .0001, rY = .06.12), and the association between weekly weighing and greater walking activity was statistically significant here (p = .002, rY = .08). In addition, weekly weighing was again associated with not smoking and with having a history of dieting (p = .001 to .0001; rY = .08.09). Analyses were rerun using all five weighing frequency categories in the models; results were essentially unchanged, and the simplified models are presented here. Weighing Frequency and Baseline BMI Figure 1 illustrates associations between weighing frequency and baseline BMI, examined using general linear models. At baseline, all BMI means were near 27 kg/m2 for POP participants, with no significant differences by weighing frequency (p = .384 for model). In contrast, BMI values show a linear trend in WTB, with larger BMI values observed with less frequent weighing (p < .0001 for model). Significant differences in BMI by baseline weighing frequency (p < .05 for comparisons) are noted in Figure 1. Changes in Weighing Frequency Over Time Using repeated measures MANOVA, changes in weighing frequency over time and by treatment group were examined. In POP, there was neither treatment group effect at baseline (p = .660) nor a time effect on weighing frequency (p = .784). There was a significant Time Treatment Group interaction (p = .001)

Demographics Demographic characteristics, weights, and weight change data are presented in Table 1. Participants in POP and WTB differed significantly on all baseline demographic characteristics; six of the eight effect sizes were in the range of medium to large (p = .007 to < .0001, rY = .05.61). POP participants were younger on average and were more likely to be women or from an ethnic minority group. They were less likely to be married and more likely to have attended at least some college. They were also less likely to have ever smoked or to have ever dieted to lose weight; however, they were twice as likely to be current smokers. Weights and Weighing Frequencies As shown in Table 1, the samples also significantly differed in terms of weight, with effect sizes in the medium to large range (p < .0001, rY = .15.50). POP participants were overweight on average (mean BMI > 25) and gained weight over time, and WTB participants were obese on average (mean BMI > 30) and lost weight over time. Table 2 also illustrates that baseline weighing frequencies were nearly identical in the two groups (p = .792 to .948, rY = .001.01). However, by 12 and 24 months into the program, WTB participants showed less never-weighing and more weekly or daily weighing, whereas POP participants showed a steady increase in the number who reported never weighing themselves. Baseline Weighing Frequency Associations Tables 2 and 3 present associations of weekly versus less than weekly weighing frequency with baseline characteristics in each of the samples. In POP, older age and lower fat intake were

TABLE 1 Demographics and Body Mass Index (BMI) in the Pound of Prevention (POP) and Weigh-to-Be (WTB) Samples Variable Age (years) Gender (% women) Ethnicity (% White) Education (% some college +) Marital status (% married) Smoking history (% ever smoked) Current smokers (%) Dieting history (% ever dieted) Baseline BMI BMI change, 12 months BMI change, 24 months Baseline weighing: Never (%) Baseline weighing: Weekly/Daily (%) 12-month weighing: Never (%) 12-month weighing: Weekly/Daily (%) 24-month weighing: Never (%) 24-month weighing: Weekly/Daily (%) POP M (SD) or % 34.5 (6.5) 81 87 88 49 42 18 77 27.15 (5.9) 0.26 (2.5) 0.50 (2.1) 20 40 25 39 27 39 WTB M (SD) or % 50.7 (12.4) 72 91 78 70 47 9 86 34.20 (6.0) 0.53 (2.2) 0.48 (2.6) 20 39 12 51 16 49 t or 2 41.99 35.95 12.49 49.08 141.72 7.27 63.02 36.21 31.95 8.21 9.25 0.00 0.07 72.51 36.22 39.26 23.63 p < .0001 < .0001 .0004 < .0001 < .0001 .007 < .0001 < .0001 < .0001 < .0001 < .0001 .948 .792 < .0001 < .0001 < .0001 < .0001 rY .61 .11 .06 .13 .22 .05 .14 .11 .50 .17 .20 .001 .01 .17 .12 .13 .10

Note. For POP, N = 1,226 at baseline, 1,056 at 12 months, and 992 at 24 months. For WTB, N = 1,800 at baseline, 1,338 at 12 months, and 1,000 at 24 months.

Volume 30, Number 3, 2005

Self-Weighing Frequencies

213

TABLE 2 Associations of Weighing Frequency With Baseline Characteristics in the Pound of Prevention Sample Weighing < Weekly M (SD) or % 33.7 (6.4) 65.6 (34.0) 2.9 (2.5) 50 24 26 81 45 86 15 44 22 70 Weighing Weekly M (SD) or % 35.6 (6.4) 61.7 (32.8) 3.2 (2.6) 49 27 24 81 55 90 10 38 12 88 t or 2 5.07 1.98 1.85 1.36 rY .14 .06 .05 .03

Variable Age (years) Fat intake (g/day) Walking episodes/week Treatment group (%) Control Education only Education + lottery Gender (% female) Marital status (% married) Education (% some college +) Ethnicity (% non-White) Ever smoked (% yes) Current smokers (% yes) Dieting history (% yes)
Note.

p < .0001 .047 .065 .507

0.01 11.13 3.62 4.74 4.05 18.98 56.77

.937 .001 .057 .029 .044 < .0001 < .0001

.002 .10 .05 .06 .06 .12 .22

For weighing < weekly, N = 736. For weighing weekly, N = 490.

TABLE 3 Associations of Weighing Frequency with Baseline Characteristics in the Weigh-to-Be Sample Weighing < Weekly M (SD) or % 49.4 (12.0) 35.3 (6.5) 6.6 (9.3) 32 34 34 72 69 78 11 47 10 83 Weighing Weekly M (SD) or % 52.6 (12.6) 34.5 (6.2) 8.1 (10.3) 35 32 33 71 73 77 6 47 6 90 t or 2 5.30 2.49 3.13 1.70 rY .12 .06 .08 .03

Variable Age (years) Fat screener score Blocks walked/day Treatment group Control Mail Phone Gender (% female) Marital status (% married) Education (% some college +) Ethnicity (% non-White) Ever smoked (% yes) Current smokers (% yes) Dieting history (% yes)
Note.

p < .0001 < .0001 .002 .427

0.53 3.41 0.54 10.51 0.03 12.10 14.56

.466 .065 .461 .001 .854 .001 .0001

.02 .04 .02 .08 .004 .08 .09

For weighing < weekly, Ns range from 1,051 to 1,088. For weighing weekly, Ns range from 692 to 710.

in POP, indicating that the control group decreased weighing over time, and both intervention groups increased weighing over time. In WTB, there was neither treatment group effect at any time point (ps = .798.939) nor a significant Time Treatment Group interaction (p = .778). A significant time effect in WTB (p = .007) indicates that self-weighing frequencies increased over time in all treatment groups. Changes in BMI by Weighing Frequency Category Associations of BMI changes at 12 and 24 months with self-weighing frequency were examined using general linear models that controlled for baseline BMI and baseline weigh-

ing frequency. Results are illustrated in Figure 2. All weighing frequency terms contributed significantly to overall models (p = .0002 and p < .0001 for POP at 12 and 24 months; ps < .0001 for WTB at 12 and 24 months). For POP, daily weighing at 12 and 24 months was associated with weight losses at those time points. All other weighing categories in POP were associated with weight gains. For WTB, monthly, weekly, and daily weighing were associated with weight losses at 12 and 24 months; less frequent weighing was associated with weight gain. Results were not affected by adding demographics (age, gender, ethnicity, education level, marital status), treatment group, or relevant behaviors (fat intake, walking, smoking) to the models.

214

Linde et al.

Annals of Behavioral Medicine

FIGURE 1 Mean baseline body mass index (BMI; kg/m2) by weighing frequency category ( SE). POP = Pound of Prevention weight gain prevention sample; WTB = Weigh-to-Be weight loss intervention sample. Within each group, means with different labels are significantly different (p < .05).

DISCUSSION Although regular self-weighing has been reported as a weight control strategy by participants in population surveys (6,7) and has been considered as a part of cognitive-behavioral weight loss approaches (8,9), this self-monitoring technique has received very little additional attention from weight control researchers. The purpose of this study was to clarify the role of regular weighing in weight control, with the intent of enhancing the clinical and public health significance of regular self-monitoring of body weight. To that end, we examined associations of self-weighing frequency in two distinct populations: Overweight individuals enrolled in a weight gain prevention trial and obese individuals enrolled in a weight loss trial. Analyses indicated that regular self-weighing of at least weekly frequency was associated with lower BMI and greater weight losses over time in these two groups. The 12- and 24-month weight losses of 1.3 and 2 BMI units, respectively, that were associated with daily weighing frequency for weight loss trial participants were in the clinically significant range. Strengths of the study include the large sample sizes for the trials, which differed significantly from each other in terms of all major demographic characteristics; the repeated measures of self-weighing behavior in common across the studies; and the collection of measured weights at all but one time point across the two samples. In both populations, self-weighing was associated with other healthy behaviors such as eating less fat, getting more exercise, and not smoking. However, self-weighing did demonstrate an independent effect on weight change, as evidenced by results that did not change when behavioral covariates (e.g., walking and fat intake) were added to the statistical models. The weight control literature indicates that self-monitoring, in general, is central to the behavioral weight loss process (3) and that

FIGURE 2 Twelve- and 24-month mean changes in body mass index (BMI; kg/m2) by weighing frequency category ( SE). POP = Pound of Prevention weight gain prevention sample; WTB = Weigh-to-Be weight loss intervention sample. Within each group, means with different labels are significantly different (p < .05).

participants who monitor eating and exercise behaviors typically achieve better weight losses than those who do not monitor (24). As nutrition and physical activity recommendations form the core of most behavioral weight control programs (3), individuals who are willing to engage in and monitor these behaviors to control their weight may be likely to accept additional messages about regular weighing for the purpose of weight monitoring as well. Instructions for self-weighing in the weight gain prevention program placed greater emphasis on the importance of regular weighing. However, despite the less emphatic instructions for weight loss participants, self-weighing was more strongly associated with weight change in the weight loss trial and was not associated with baseline weights in the weight gain prevention trial. These outcomes may have resulted from the greater emphasis on achievement of weight loss for those engaged in the weight loss trial as well as from the greater interest in weight loss and weight loss behaviors for participants who sought to en-

Volume 30, Number 3, 2005

ter a weight loss trial. Future weight loss trial participants may derive even greater benefit from regular self-weighing if instructions are revised to promote regular weighing more definitively. Other differences in weight gain prevention and weight loss trial results were observed as well. Again, despite more emphatic weighing instructions in the weight gain prevention trial, more weight loss trial participants were able to maintain regular weighing behaviors than weight gain prevention participants, as evidenced by increases over time in weekly and daily weighing percentages in WTB and increases in never-weighing percentages in POP. It may be that weight gain prevention participants were less likely to maintain engagement in a self-monitoring behavior closely linked to weight than those involved in a weight loss trial, who may have entered that trial with greater expectations of engaging in self-monitoring for weight control, perhaps due to their significantly greater experience with dieting to lose weight. Different patterns in weighing frequency changes over time between treatment groups in the two studies were also observed, such that only the control group in POP decreased weighing significantly over time. All intervention groups in POP and WTB received weighing instructions, and the control group in POP did not receive such instructions. However, the control group in WTB was given the option to seek weight management services through their managed care organization, and participants in the control group appear to have done so, for they had lost weight on average over the first 12 months of the study (14). Likewise, unpublished participation data from WTB indicate that 34% of the control group did engage in a variety of weight control activities during the study (25). (Sherwood et al., 2004). It is likely that even participants in the WTB control group were exposed to weight monitoring messages during their participation in the trial or had their general awareness of weight-related concerns raised as a result of study participation, both of which may explain increases in weighing frequency over time in this group. Similarly, participants in the POP no-contact control group, in the absence of any formal education or instruction in weight control during the trial (13), may have lacked exposure to messages that encouraged them to continue to engage in any weight control-relevant behavior and thus failed to maintain these behaviors even if performed at the start of the study. The study design was limited by several assessment-related factors, including reliance in both trials on self-reports of weighing frequency, which may be prone to either over- or underestimation bias, restriction of the range of self-weighing categories to the five that were common across both studies (rather than using the full range of six in WTB and seven in POP), and failure to use open-ended items to assess weighing frequency in either study. In addition, attrition was high in both samples, particularly in WTB, which had a 74% response rate at 12 months and a 56% response rate by 24 months; in comparison, response rates for POP were 86% at 12 months and 81% at 24 months. Recognizing that this study design is limited also by cross-sectional analyses of self-weighing, BMI, and BMI change, a randomized controlled trial that manipulates frequency of self-weighing

Self-Weighing Frequencies

215

within a weight loss or weight gain prevention program would be a stronger test of the links between self-weighing behavior and weight change. Based on the differences in findings between these programs, different types of trials might lend themselves to varying approaches to self-monitoring of weight, perhaps in terms of frequency instructions (e.g., instructing participants to weigh themselves weekly to forestall larger weight gains with time, or daily to improve chances of weight gain reversals in the context of weight gain prevention vs. instructing daily weight monitoring for best chances at larger weight changes in the context of weight loss) or in terms of message framing (e.g., discussing weighing as a healthy lifestyle tool along with regular exercise and fruit and vegetable consumption for weight gain prevention vs. discussing weighing as one more self-monitoring tool to be used along with rigorous calorie and exercise tracking for weight loss). Likewise, self-weighing may benefit from a similar examination in adults whose BMI is in the normal range to test its applicability as an overweight and obesity prevention tool. As the results of this study suggest, self-weighing is important for both weight loss and weight gain prevention efforts. In light of these findings, the efficacy of regular weight tracking, in the form of self-monitoring of weight, should be tested rigorously; and clinical as well as public health recommendations for regular weighing should be considered. REFERENCES
(1) Flegal KM, Carroll MD, Ogden CL, Johnson CL: Prevalence and trends in obesity among US adults, 19992000. Journal of the American Medical Association. 2002, 288:17231727. (2) Jeffery RW, Drewnowski A, Epstein LH, et al.: Long-term maintenance of weight loss: Current status. Health Psychology. 2000, 19(Suppl.):516. (3) Wing RR: Behavioral approaches to the treatment of obesity. In Bray GA, Bouchard C, James WPT (eds), Handbook of Obesity. New York: Marcel Dekker, 1998, 855878. (4) United States Department of Agriculture: Nutrition and Your Health: Dietary Guidelines for Americans. Retrieved May 4, 2004 from http://198.102.218.57/dietaryguidelines/dga2000/ document/frontcover.htm (5) Centers for Disease Control: Overweight and Obesity. Retrieved May 4, 2004 from http://www.cdc.gov/nccdphp/dnpa/ obesity/ (6) McGuire MT, Wing RR, Klem ML, Hill JO: Behavioral strategies of individuals who have maintained long-term weight losses. Obesity Research. 1999, 7:334341. (7) McGuire MT, Wing RR, Klem ML, Seagle HM, Hill JO: Long-term maintenance of weight loss: Do people who lose weight through various weight loss methods use different behaviors to maintain their weight? International Journal of Obesity. 1998, 22:572577. (8) Cooper Z, Fairburn CG: A new cognitive behavioural approach to the treatment of obesity. Behaviour Research and Therapy. 2001, 39:499511. (9) Cooper Z, Fairburn CG: Cognitive-behavioral treatment of obesity. In Wadden TA, Stunkard AJ (eds), Handbook of Obesity Treatment. New York: Guilford, 2002, 465479. (10) Touyz SW, Lennerts W, Freeman RJ, Beumont PJV: To weigh or not to weigh? Frequency of weighing and rate of weight gain

216

Linde et al.
in patients with anorexia nervosa. British Journal of Psychiatry. 1990, 157:752754. Ogden J, Evans C: The problem with weighing: Effects on mood, self-esteem and body image. International Journal of Obesity. 1996, 20:272277. Ogden J, Whyman C: The effect of repeated weighing on psychological state. European Eating Disorders Review. 1997, 5:121130. Jeffery RW, French SA: Preventing weight gain in adults: The Pound of Prevention study. American Journal of Public Health. 1999, 89:747751. Jeffery RW, Sherwood NE, Brelje K, et al.: Mail and phone interventions for weight loss in a managed-care setting: Weighto-Be one-year outcomes. International Journal of Obesity. 2003, 27:15841592. Block G, Hartman AM, Dresser CM, et al.: A data-based approach to diet questionnaire design and testing. American Journal of Epidemiology. 1986, 124:453469. Block G, Woods M, Potosky A, Clifford C: Validation of a self-administered diet history questionnaire using multiple diet records. Journal of Clinical Epidemiology. 1990, 43:13271335. Thompson F, Byers T: Dietary assessment resource manual. Journal of Nutrition. 1994, 124:2296S. Caan B, Coates A, Schaffer D: Variations in sensitivity, specificity, and predictive value of a dietary fat screener from Block

Annals of Behavioral Medicine

et al. Journal of the American Dietetic Association. 1995, 95:564568. Paffenbarger R, Wing A, Hyde R: Physical activity as an index of heart attack risk in college alumni. American Journal of Epidemiology. 1978, 108:161175. Pirie P, Jacobs D, Jeffery R, Hannan P: Distortion in self-reported height and weight data. Journal of the American Dietetic Association. 1981, 78:601606. SAS Institute: SASOnlineDoc (Version 8). Retrieved May 5, 2003 from http://www.epi.umn.edu/sas8docs/ Cohen J, Cohen P: Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences (2nd Ed.). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc., 1983. Cohen J: Statistical Power Analysis for the Behavioral Sciences (2nd Ed.). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc., 1988. Wadden TA, Letizia KA: Predictors of attrition and weight loss in patients treated by moderate and severe calorie restriction. In Wadden TA, VanItallie TB (eds), Treatment of the Seriously Obese Patient. New York: Guilford, 1992, 383410. Sherwood NE, Pronk NP, Jeffery RW, et al.: Mail and phone interventions for weight loss in a managed-care setting: Weighto-be two-year outcomes [abstract]. Annals of Behavioral Medicine. 2004, 27(Suppl.):45.

(11)

(19)

(12)

(20)

(13)

(21) (22)

(14)

(23)

(15)

(24)

(16)

(25)

(17) (18)