Abnormalities in Weight Status, Eating Attitudes, and Eating Behaviors Among Urban High School Students:

Transcription

1 JOURNAL OF ADOLESCENT HEALTH 1996;18: FELLOWSHIP FORUM Abnormalities in Weight Status, Eating Attitudes, and Eating Behaviors Among Urban High School Students: Correlations with Self-Esteem and Anxiety DORIS R. PASTORE, M.D., MARTIN FISHER, M.D., AND STANFORD B. FRIEDMAN, M.D. Purpose: The study's objective was to determine the prevalence of abnormalities in weight, eating attitudes, and eating behaviors among an urban teenage population. Methods: Measures of weight, height, and blood pressure were obtained from 1,001 students attending regularly scheduled physical education classes in a New York City high school the population of which was 66% black and 23% Hispanic. The mean age was 16 years old; 55% were female and 45% were male. Approximately three quarters of the students completed the Eating Attitudes Test (EAT-26), Rosenberg Self-Esteem Scale, Spielberger Trait Anxiety Inventory, and a questionnaire on eating behaviors. Results: By percent ideal body weight (%IBW), 25% of students were obese (>20% above IBW), 18% were overweight, and 5% were underweight. Abnormal EAT scores (>21), were found in 6% of males and 15% of females. Students reported that they had engaged in the following nutritional behaviors at least three times in the past week: eating candy (59%), skipping breakfast (58%), skipping lunch (42%), eating fast foods (28%), and skipping dinner (14%). Self-esteem was high in 60%, medium in 21%, and low in 19%. Mean self-esteem was signifi- From the Division of Adolescent Medicine, Mount Sinai Medical Center; and Mount Sinai School of Medicine, New York, New York (D.R.P.), the Division of Adolescent Medicine, North Shore University Hospital, and Cornell University Medical College, New York, New York (M.F.); the Division of Adolescent Medicine, Montefiore Medical Center, and Albert Einstein College of Medicine, New York, New York (S.B.F.). Address correspondence to: Dr. Doris R. Pastore, Mount Sinai Medical Center, Adolescent Health Center, 312 East 94th Street, New York, New York Manuscript accepted November 10, I995. cantly higher (P <.0001) than established norms. Selfesteem and anxiety were inversely correlated (r =.40, P <.001), and low self-esteem and high anxiety were both associated with high EAT scores (r =.29, P <.001). Self-esteem and anxiety of obese students did not differ from those who were of normal weight. Conclusions: Among this population of urban high school students there were: (a) a large number of overweight adolescents; (b) a significant subgroup with eating attitudes suggestive of an eating disorder; and (c) high levels of self-esteem and normal levels of anxiety, which were independent of weight status. KEY WORDS: Adolescents Obesity Eating Attitudes Self-esteem Abnormalities in weight, eating attitudes, and eating disorders among adolescents are important public health concerns (1-3). As an outgrowth of a relationship fostered with the teaching staff and high school administration of a New York City high school in which we run a school-based clinic, we had the opportunity to investigate these concerns among black and Hispanic teenagers and to examine their correlations with self-esteem and trait anxiety. Obesity appears to be increasing among children and adolescents (1). Obesity, defined as - 120% X/96/$15.00 Society for Adolescent Medicine, 1996 SSDI X(95) Published by Elsevier Science Inc., 655 Avenue of the Americas, New York, NY 10010

2 May 1996 ABNORMALITIES IN WEIGHT STATUS 313 above ideal body weight, or a body mass index (BMI) > the 85th percentile, is present in approximately 21% of teenagers (1,2). This is a particular problem for inner-city teens, among whom obesity has had its greatest rate of increase (4). Extremely obese adolescents generally evolve into the heaviest adults, and so may contribute to a disproportionate share of the morbidity of adult obesity (5). This was recently confirmed by long-term follow-up of the Harvard Group Study of , in which the risk of morbidity for coronary heart disease and atherosclerosis was increased among men and women who had been overweight (defined as BMI >75th percentile) as adolescents (6). The level of obesity that needs to be treated and the risks and benefits of treatment remain controversial (7). This study evaluates the presence of weight abnormalities and their correlates as a step toward developing appropriate management policies. The prevalence of eating disorders (e.g., anorexia nervosa, bulirnia nervosa) has historically been thought to be low in urban settings (8). Whether the reported rate of eating disorders reflects lesser use of health care systems by inner-city urban adolescents, a genuine difference in risk, or a limitation of the measures used is unclear. Other authors have reported that the presence of bulimia and purging among black, low-income adolescents is as high as 12-18% (9-11). This report assesses the presence of eating abnormalities among a largely black and Hispanic, low-income teenage population. Dietary intake and eating habits were also examined. Although snacking is thought to be increased during the teenage years (12), this has not been studied recently in an urban nonclinical population. We also examined the relationship of intake patterns to weight status and eating attitudes. In addition, we assessed self-esteem and anxiety in this student population and their correlation with weight and eating attitudes. Self-esteem is thought to play a role in both anorexia nervosa and obesity. Many treatment programs include components to enhance self-esteem. SaUade (13) compared selfesteem of obese versus nonobese, low-income students in Grades 3, 5, 8, and 11 and found obese students to have poorer self-concept; others have found no difference in self-esteem between obese and nonobese groups (14,15). In contrast, Fisher et al. studied suburban female high school students and found strong intercorrelations among abnormal eating attitudes, low self-esteem, high anxiety, and increased weight concerns (16). Methods Procedures We collected data as part of a health education program entitled Health Screening Week, developed through the joint efforts of school-based clinic staff, high school faculty, school administration, and students. Each student attending his or her regularly scheduled physical education class was assessed for height, weight, visual acuity, and blood pressure. Approximately 70 high school students on a nursing assistant career track were trained to do the assessments and were supervised by high school faculty and school-based clinic staff. They obtained measurements on 1,001 students over a 5-day period. The school administration approved demographic questions on age and gender, but did not allow any questions regarding ethnicity or socioeconomic status. Information on these characteristics therefore had to be obtained indirectly, and inferences made about the likely composition of the study sample. This New York City high school is located on the Far Rockaway peninsula. The school has a total population of approximately 1,500 students, 66% of whom are black, 23% Hispanic, 8% white, and 3% Asian. Currently, more than 50% of public housing in the borough of Queens, New York City, is located in Far Rockaway. Data from students who attend the school-based clinic at this high school indicate that 66% have no health insurance, 20% have Medicaid, and 14% have private insurance. We have inferred that school-based clinic students are representative of the school as a whole, based on data from our study of violence and mental health problems among urban high school students (in this issue). A total of 26% of clinic attenders had lived in the United States < I year; 40% had lived at the same address <1 year; 45% were born in the United States; 74% described their first language as English; 70% lived with their mother; and 51% had mothers who were high school graduates (17). The 1,001 students surveyed in this study had a mean age of 16.0 years; 55% were female and 45% male. This was similar to school enrollment demographics of 50% female and 50% male. Several questionnaires were administered during the week the measurements were obtained. A total of 1,001 students, which represents the total n of this study, had height and weight measurements and received the questionnaires. Each instrument of the questionnaire was successfully completed by approximately 3/~ of the students. As per the recommendation of our biostatistician, an instrument was con-

3 314 PASTORE ET AL. JOURNAL OF ADOLESCENT HEALTH Vol. 18, No. 5 sidered to be successfully completed if there were no greater than two blank answers, thus allowing us to calculate a meaningful score. The students then received a health care lecture which described the importance of knowledge and self-awareness in the areas of nutrition, weight, vision, and blood pressure. Instruments The questionnaires included the following: 1. The Eating Attitudes Test (EAT-26), a self-administered questionnaire developed by Garner and Garfinkel to measure symptoms of anorexia nerrosa (18). Each item of the instrumentmfor example, "I find myself preoccupied with food"--is a symptom frequently observed in patients with eating disorders. Respondents rate each item on a 6-point scale ranging from always to never, and a score is calculated. The EAT-26 is based on an original Eating Attitudes Test (EAT-40). The EAT-40 has an excellent internal consistency (alpha coefficient =.94) for a combined sample of patients with anorexia nervosa and normal controis (19). Sensitivity is 93% and specificity is 88%. The abbreviated EAT-26 is highly correlated with the 40-item EAT (r =.98 for anorexia nervosa patients and r =.97 for female university students) (19). Total scores on the EAT-26 are derived as a sum of the composite items, ranging from 0 to 78, with scores generally suggestive of an eating disorder. The EAT-26 consists of three factor scores: (F1) dieting--the degree of avoidance of fattening foods and preoccupation with being thinner; (F2) bulimia and preoccupation with food; and (F3) oral control--the degree of selfcontrol around eating and the perception of pressure from others to gain weight. In nonclinical populations, the EAT has been used as a screening instrument to identify individuals who are more likely to have disturbed eating patterns. 2. The Rosenberg Self-Esteem Scale is a well-studied measure of global self-esteem. Respondents rate as strongly agree, agree, disagree, or strongly disagree 10 items which include statements such as "On the whole, I am satisfied with myself" (20). The instrument has a reproducibility of.93. It was scored using a Guttman rating scale, with higher scores indicating lower self-esteem, (i.e., high = 0-1, medium = 2, low = 3-6). 3. The Spielberger Trait Anxiety scale measures. chronic anxiety (21). In responding to such items as "I have disturbing thoughts," subjects report how often they have experienced particular symptoms of anxiety on a 4-point scale (almost never to almost always). This 20-item measure is scored on a point system. High scores indicate higher levels of trait anxiety. It has a reliability coefficient of.78 for males and.85 for females. A two-page cover sheet was also administered, which included (a) demographic information, (b) questions about perceptions of weight status, and (c) a nutritional behaviors survey. The nutritional questions related to frequency of skipped meals, consumption of fast foods, amount of exercise. Data Analysis General distributions were determined for weight abnormalities using a formula for percent ideal body weight (%IBW): %IBW = [actual weight/ideal weight] x 100. IBW was determined using calculations based on adolescent-specific weight for height tables (22). Distributions also were determined for each standardized questionnaire. Chi-square analysis was used to determine variations by age, sex, and grade. Analysis of variance and two-tared t tests were used to determine significant mean score differences. Correlation coefficients were used to detect relationships among weight status and eating attitudes, self-esteem, and anxiety (23). Results Weight Status Weight status was defined using %IBW for obesity (>120%IBW), overweight ( %IBW), normal weight (86-109%IBW), and underweight (-<85%IBW). As shown in Table 1, 25% of students were obese, 18% were overweight, 52% were of normal weight, and 5% were underweight. There were no significant differences between males and females. Within the obese group, 35% were superobese (-140%IBW or BMI -95th percentile). This represents 8% of the total student population. %IBW and BMI achieved a correlation of r =.98. Table 1 also demonstrates that 28% of the study group reported that they considered themselves overweight, whereas 57% considered themselves to be normal weight and 15% considered them-

4 May 1996 ABNORMALITIES IN WEIGHT STATUS 315 Table 1. Weight Status, Weight Beliefs, and Eating Attitudes Among Urban High School Students Total Male Female (%) (%) (%) Weight status (%IBW)*(n = 1,001) Obese (->120) Overweight ( ) Normal weight (86-109) Underweight (-<85) Weight beliefs (n = 577) Believe overweight Believe normal weight Believe underweight Eating attitudes** (n = 683) Total score ~21 (abnormal) Factor scores Dieting behavior (Factor I ->18) Bulimic behavior (Factor 2 ->4) Oral control (Factor 3 ->5) * %IBW = percent ideal body weight. ** References 9 and 10. selves to be underweight. Of these, 35% of females and 19% of males considered themselves to be overweight (P <.001). Among those of normal weight, 58% considered themselves to be of normal weight and 33% considered themselves overweight. Among students who were % of IBW, 23% considered themselves to be of normal weight, whereas 74% of obese students considered themselves to be overweight. Nearly all of the superobese students considered themselves to be overweight. No sex differences were found among obese and superobese students for this issue. Eating Attitude Scores Eating Attitudes Test (EAT-26) scores are also presented in Table 1. Total scores suggestive of an eating disorder (EAT ->21) were found to be present in 12% of the total population; this included 15% of the females and 6% of the males. Females in this populations had a mean score of 10.1, which was not significantly different from the mean of 11.9 established for normal controls. In contrast, the mean score of 8.3 for males in this study was significantly greater (P <.05) than the mean of 5.2 found among other populations of high school boys (24,25). Factor analysis of the EAT responses revealed high F1 (dieting) scores, reflecting dieting behaviors and a dissatisfaction with body shape, in 6% of the students. High F2 (bulimia and food preoccupation) Table 2. Self-Esteem and Anxiety Among Urban High School Students Scores Total (%) Male (%) Female (%) Self-esteem (n = 711) High (0-1) Medium (2) Low (3-6) Anxiety (n = 691) Low (20-30) Medium (30-40) High (40-60) scores were indicated by 10% of females and 10% of males. F3 (oral control) scores, which reflect selfcontrol about food as well as social awareness about weight, were high in 25% of the students studied. The sample F3 score of 2.0 [+3.1 standard deviation (SD)] was significantly greater (P <.001) than the published norm of 1.9 (+2.1 SD) for established controls. F1 and F2 scores did not differ from norms for established controls. Self-Esteem and Anxiety Scores Data from the self-esteem and anxiety measures are presented in Table 2. High self-esteem Was reported by 60% of students, medium self-esteem was reported by 21%, and low self-esteem by 19%. The mean score of 1.4 achieved by these students was significantly greater (P <.01) than the mean of 1.9 established for normal controls (indicating that these students had higher self-esteem than that reposed in the literature for other populations) (20). In addition, the mean score for males of 1.3 indicated significantly higher self-esteem than that of our female population (P <.05). Only 17% of these inner-city urban adolescents reported high anxiety. Medium anxiety was reported by 50% and low anxiety by 33%. Female students were significantly (P <.05) more anxious than male students, but neither the male nor the female groups differed significantly from published norms. Nutritional Behaviors Table 3 indicates the percentage of students who report that they had engaged in the nutritional behaviors listed at least three times in the prior week. Females were significantly (P <.05) more likely than males to report skipping breakfast and lunch. Only 43% of students reported that they ate three meals

5 316 PASTORE ET AL. JOURNAL OF ADOLESCENT HEALTH Vol. 18, No. 5 Table 3. Nutritional Behaviors (More Than Three Times in Prior Week) (n = 753)* Male (%) Female (%) Behavior Total (%) (n = 309) (n = 444) Skipped breakfast 59 49* 66 * Skipped lunch ~ 45 ~ Skipped dinner Ate three meals per day ~ 34* Ate fast foods Ate candy ~ 64 ~ Drank soda ~ 44~ Drank coffee Smoked cigarettes (~5-10) Ate fruits/vegetables Ate bread/grains $ 62 $ Drank milk * Exercised (at least 30 min) 39 54* 29 ~ * Chi-square analysis used to compare differences by gender. * P <.05; ~ P <.001. per day. A total of 28% indicated that they ate fast foods at least three times in the past week. Correlation Coefficient Analysis The interrelationships among weight, eating attitude, self-esteem, and anxiety were assessed using Spearman coefficient analysis. As Table 4 indicates, there were no significant relationships between %IBW and the EAT-26, self-esteem, or anxiety scores. However, there was a positive correlation of r =.29 between EAT-26 scores, and both lower self-esteem and higher anxiety (P <.001). Lower self-esteem was also positively correlated with higher anxiety, at r =.40 (P <.001). The positive correlation of higher EAT scores with lower self-esteem and of lower selfesteem with higher anxiety did not differ by gender. However, higher EAT scores were associated with higher anxiety scores at a correlation coefficient of r =.33 among females (P <.001), whereas among males a correlation coefficient of.r =.16 (P <.01) was achieved. Table 4. Relationships Among Eating Attitudes, Self- Esteem, and Anxiety Expressed as Spearman Rank- Correlation Coefficients % Ideal Body Weight EAT-26 Self-Esteem EAT Self-esteem.08.29* Anxiety *.40* * P <.001. Table 5. Analysis of Eating Attitudes, Self-Esteem, and Anxiety Scores by Weight Group Obese Normal Underweight (~120%IBW)* (86-119%) (~85%IBW) (n = 192) (n = 601) (n = 37) Mean eating attitude score $,~ Mean self-esteem score Mean anxiety score * IBW = ideal body weight. + P <.05, obese compared with underweight. P <.05, obese compared with normal weight. Analysis by Weight Groups Comparisons of EAT scores, self-esteem, and anxiety by weight group using two-tailed t tests are shown in Table 5. The mean EAT score of 11.0 for obese students, 9.0 for normal weight students, and 7.0 for underweight students differed significantly, indicating that higher-weight students had more abnormal eating attitudes than those who were normal or underweight (P <.05). This underscores the analysis of variance (ANOVA) of EAT scores across these weight groups, in which an F of 5.55 was significant to P <.005. Mean self-esteem scores did not differ significantly when obese and normal-weight groups were compared. The ANOVA self-esteem scores across all weight groups did not attain significance (F = 2.54, P <.075). The mean self-esteem of superobese students also did not differ from those of normal weight. Obese, normal, and underweight groups did not differ in mean anxiety scores (ANOVA of F = 0.19, P <.15). When the reported nutritional behaviors of obese, normal, and underweight students were compared, a significant (P <.001) difference was found between the percentage of obese and underweight students in skipping breakfast (72% vs. 30%) and eating three meals per day (32% vs. 59%). Among the obese group, females were more likely to skip breakfast, whereas males were more likely to eat three meals per day. There was no significant difference among weight groups in terms of skipping lunch, skipping dinner, or fast-food intake. No gender differences were found within weight groups for these items. Discussion This study found abnormalities in weight and eating attitudes among an inner-city urban population. Our

6 May 1996 ABNORMALITIES IN WEIGHT STATUS 317 findings strongly suggest that inner-city adolescents are not exempt from intense concerns about weight and shape, and that those who provide primary care to urban adolescents must be aware of the hidden subgroup of obese urban teenagers with abnormal eating attitudes. Abnormal eating attitudes in this context appear to manifest most commonly among obese teenagers struggling for oral control in a way that is less obvious than the struggles noted among suburban teenagers who are more often normal or underweight. This study found rates of obesity consistent with those observed in recent National Health and Nutrition Examination Surveys (1,2,26,27). The fact that 25% of students were - 120%IBW and 8% were - 140%IBW supports growing concerns about increases in the prevalence of obesity among teenagers. In our study, 15% of females and 6% of males had abnormal eating attitude scores. These findings are similar to rates more commonly described among suburban white student populations (16). It may be that a subgroup of urban teenagers are at particular risk of developing an eating abnormality which is similar in some ways but different in others from that found in suburban settings. Our findings would suggest that a subgroup of urban teenagers who are obese, consider themselves to be overweight, and have abnormal eating attitudes and somewhat lower self-esteem, are at risk for developing eating abnormalities. These individuals are similar to the suburban adolescent in that they consider themselves to be overweight. They differ in that they are in fact overweight. The development of abnormal eating attitudes and behaviors in any particular individual appears to be mediated through a variety of genetic, familial and social factors (28,29). The individual who is vulnerable to developing an eating disorder because of factors such as low self-esteem or increased anxiety may manifest abnormal eating attitudes or behaviors regardless of the setting. Factor analysis of the EAT revealed high F3 (oral control), reflecting efforts at self-control about food as well as social awareness about weight. High oral control scores may reflect a waning of the protective effect which the greater tolerance for higher weight among black females would offer against the intensity in dieting that has been found in white female populations. Studies by Rucker and Cash found that normal-weight college-age black females, compared with white females, held more favorable body image attitudes, cognitions, and behaviors (30). A study confirms work by other researchers who studied dieting and purging behaviors among groups of urban teenagers and consistently found that small but significant numbers were regularly using vomiting, laxatives, or diuretics, or demonstrated a concern that they had a eating disorder (9-11,32). We were unable to highlight differences between black and Hispanic students in our study, because we were denied permission by the school administration to ask students their ethnicity. Desmond's work on adolescents' perception of their weight reported that approximately 30% of students perceived their weight category inaccurately (31). Our study's findings support these results in that among those subjects ~ 120%IBW, 74% considered themselves to be overweight and 26% considered themselves to be of normal weight. Contrary to common assumptions, this study found that self-esteem among these inner-city adolescents was not low, and despite trends of increasing weight associated with somewhat lower selfesteem, the self-esteem of obese students did not differ from those of normal-weight students. Our work agrees with the findings of Wadden et al. (14) and Kaplan and Wadden (15), who noted no differences in self-esteem among weight groups. This work also confirms results obtained by Kleges, who found no differences in the global measures of assertiveness, depression, or self-consciousness between obese and nonobese groups (33). We found trait anxiety levels to be medium to low, and not to differ from established norms. This finding substantiates normal levels of self-esteem and implies that these adolescents, as a group, do not have psychological difficulties in these areas. The nutritional habits reported by the adolescents in this study are similar to those generally described by teenagers. Teenagers, for example, have often reported eating fewer than three meals per day, and obese teenagers most frequently report skipping meals. Our findings are also similar to those reported by the Hispanic Health and Nutrition Examination Survey, in which intake included more energy-dense and high-calorie foods than items of nutritive value (35). The limitations of this study pertain to the challenges of conducting research in a high school setting (see commentary) and to the instruments used. Caution must be used in the application of the EAT to a nonclinical setting (24,25,35). One must consider that the EAT was initially developed for inpatients with anorexia nervosa, and that the items may have different meaning to nonpatients. Like other selfreport measures, validity is dependent on accurate subject reporting, and those with tendencies toward

7 318 PASTORE ET AL. JOURNAL OF ADOLESCENT HEALTH Vol. 18, No. 5 abnormal eating behaviors might have particular difficulty with precise reporting. In addition, each question item in the EAT has equal weight and some studies have indicated that more discriminant analysis would yield a higher predictive value (25). Similarly, because the Rosenberg Self-Esteem Scale is a measure of global self-esteem, it may not be sufficiently sensitive to assess differences between ethnic groups or weight groups as outlined in this study. Other measures of self-concept would have allowed us to confirm our findings. In addition, some bias could have possibly been introduced by omitting students who had greater than two blank answers, but different students missed different answers and no apparent pattern was observed. In reviewing the abnormalities in weight status, eating attitudes, and eating behaviors found among these urban students, we were most surprised by the trend toward aberrant eating attitudes. The scores achieved by these inner-city adolescents on the EAT were higher than expected, and in fact approached those of the 1st- and 2nd-year university students originally used as a control group for this test (18,19). Also, we emphasize that the majority of students who were - 120%IBW considered themselves to be overweight, and it was this group which had the most abnormal eating attitudes, the least healthy eating behaviors, and the lowest self-esteem relative to the other weight groups. We would recommend that those providing primary care to urban adolescents tailor their medical and psychosocial screening to be sensitive to issues of weight, eating attitudes, and eating behaviors to detect the subgroup of teenagers vulnerable to developing eating-related abnormalities. Mental health services should recognize that inner-city adolescents, including those who may be overweight, do not necessarily have low global self-esteem. Refereylce$ 1. Gortmaker S, Dietz W, Sobol A, et al. Increasing pediatric obesity in the U.S. Am J Dis Child 1987;141: Prevalence of overweight among adolescents, United States, MMWR 1994;43: Pope H, Hudson J, Yurgehin-Todd D, et al Prevalence of anorexia nervosa and bulimia in three student populations. Int J Eating Disord 1984;3: Dietz W. You are what you eat--what you eat is what you are. J Adolesc Health 1990;11: Rimm I, Rimm A. Association between juvenile onset obesity and severe adult obesity in 73,532 women. Am J Public Health 1976;66: Must A, Jacques P, DaUal G, et al. Long-term morbidity and mortality of overweight adolescents. N Engl J Med 1992;327: Garner D, Wooley S. Confronting the failure of behavioral and dietary treatments for obesity. Clin Psychol Rev 1991;11: Kendell R, Hall D, Hailey A, et al. The epidemiology of anorexia nervosa. Psychol Med 1973;3: Emmons L. Dieting and purging behavior in black and white high school students. J Am Diet Assoc 1992;92: Klllen JD, Taylor CB, Telch MJ, et al. Self-induced vomiting and laxative and diuretic use among teenagers. JAMA 1986; 255: Ballentine M, Stitt K, Bonner J, et al. Self-reported eating disorders of black, low-income adolescents: Behavior, body weight perceptions and methods of dieting. J Sch Health 1991;61: Huenemann R, Shapiro L, Hampton M, et al. Food and eating practices of teenagers. J Am Diet Assoc 1968;53: Sallade J. A comparison of the psychological adjustment of obese vs. non-obese children. J Psychosom Res 1973;17: Wadden T, Foster G, BrowneU K, et al. Self-concept in obese and normal-weight children. J Consult Clin Psychol 1984;52: Kaplan K, Wadden T. Childhood obesity and self-esteem. J Pediatr 1986;109: Fisher M, Schneider M, Pegler C, et al. Eating attitudes, health-risk behaviors, self-esteem and anxiety among adolescent females in a suburban high school. J Adolesc Health 1991;12: Fisher M, Juszczak L, Friedman SB, et al. School-based adolescent health care: Review of a clinical service. Am J Dis Child 1992;146: Garner D, Garfinkel P. The EAT: An index of the'symptoms of anorexia. Psychol Med 1979;9: Garner D, Dimstead M, Bohr T, et al. The EAT: Psychometric features and clinical correlations. Psychol Med 1982;12: Rosenberg M. Conceiving the self. New York, Basic Books, Spielberger, C. State-Trait Anxiety Inventory for Children. New York, Consulting Psychologists Press, Inc., Baldwin BT, Wood TD. Average weight for height tables (9 and 11). In: Proudfit b-q', Robinson CH. Normal and therapeutic nutrition. New York, Macmillan, 1961:782, Hollander M, Wolfe D. Nonparametric/statistical methods. New York, John Wiley & Sons, Rosen J, Silberg N, Gross J. Eating Attitudes Test and Eating Disorder Inventory: Norms for adolescent girls and boys. J Consult Clin Psychol 1988;56: Williams R. Use of the Eating Attitudes Test and Eating Disorder Inventory in adolescents. J Adolesc Health 1987;8: The Public Health Service. Surgeon General's Report on Nutrition and Health. Washington, DC: U.S. Dept. of Health and Human Services; no. (PHS) Dawson D. Ethnic difference in female overweight: Data from the 1985 National Health Interview Survey. Am J Public Health 1988;78: Strober M. Humphrey L. Familial contributions to the etiology and course of anorexia and bulimia. J Consult Clin Psychol 1987;55: Pumariega A. Acculturation and eating attitudes in adolescent girls: A comparative and correlational study. J Am Acad Child Psychol 1986;25:276-9.

8 May 1996 ABNORMALITIES IN WEIGHT STATUS Rucker CE, Cash TF. Body images, body-size perceptions and eating behaviors among African-American and white college women. Int J Eating Disord 1992;12: Desmond S, Price J, Gray N, CYConnell J. The etiology of adolescents' perceptions of their weight. J Youth Adolesc 1986;6: Rosen J, Gross J. Prevalence of weight reducing and weight gaining in adolescent girls and boys. Health Psycho11987;6: Kleges RC. Personality and obesity: Global vs. specific measures? Behav Assess 1984;6: Murphy S, Castilli R~ Mendoza F. An evaluation of food group intakes by Mexican-American children. J Am Diet Assoc 1990;90: Williams P, Hand D, Tarnopolsky A. The problem of screening for uncommon disorders---a comment on the Eating Attitudes Test. Psychol Med 1982;12: