Journal Of Community Health Vol. 16, No. 3, June 1991
BODY MASS INDEX AND PERCEIVED W E I G H T STATUS IN YOUNG ADULTS John P. Sciacca, Ph.D., M.P.H.; Christopher L. Melby, D.H. Sc., M.P.H.; Gerald C. Hyner, Ph.D.; Amy C. Brown, Ph.D., R.D.; Paul L. Femea, D.N. Sc. ABSTRACT: Body Mass Index (BMI) was calculated from self-reported height and weight for 1,123 university students who returned a questionnaire mailed to a 10 percent r a n d o m sample of the entire u n d e r g r a d u a t e population of a large midwestern university. Seventeen percent of the females and 20 percent of the males were determined to be in excess of normal BMI standards. However, significantly more women (40%) considered themselves overweight in comparison to men (24%). Also, significantly more women (53%) than men (20%) reported experiencing discomfort due to excessive weight. Inaccurate perceptions of body image are common among individuals with eating disorders. T h e r e is a higher incidence of eating disorders among college-age women than among their male peers. A distorted body image as reflected by perceived overweight may serve as a marker for individuals at risk for eating disorders.
Body-image misperception exists in both anorexia nervosa and bulimic patients. In fact, an inappropriate perception of body image is one of the DSM-III diagnostic criteria for anorexia nervosa patients.' Bulimics, as well, are likely to perceive themselves as overweight even though their body weight does not differ significantly from that of nonbulimics. ~ There is a strong relationship between perception of overweight and harmful eating disorders such as anorexia nervosa and bulimia? Such disorders affect thousands of young American women and can result in a variety of health problems. 4'5,6 John P. Sciacca is Assistant Professor of Health Education; Amy C. Brown is Assistant Professor of Food and Nutrition Science; both are in the Department of Health, Physical Education, Recreation, Food and Nutrition Science, School of Health Professions, Northern Arizona University, Flagstaff, AZ 86011. Christopher L. Melby is Associate Profi~ssor of Nutritional Science, Dept. of Food Science and Human Nutrition, Colorado State University, Fort Collins, CO 80523. Gerald C. Hyner is Associate Professor of Health Promotion and Paul L. Femea is Associate Professor of Nursing, both at Purdue University, West Lafayette, IN 47906. Requests for reprints should he addressed to: .John Sciacca, Dept. of HPERF&N, School of Health Professions, Box 6012, Northern Arizona University, Flagstaff, AZ 86011. © 1991 ttuman Sciences Press, Inc.
159
160
JOURNAL OF COMMUNITY HEALTH
An i m p o r t a n t aspect of the therapeutic m a n a g e m e n t o f eating disorders involves teaching the individual to relinquish an unrealistic p e r c e p t i o n o f ideal body weight and to accept a body image consistent within a healthy weight range, r T h e early recognition o f individuals at risk o f eating disorders is i m p o r t a n t in the p r i m a r y prevention of anorexia nervosa a n d bulimia. ~''~' Early recognition o f risks involves determ i n i n g dissatisfaction with a n o r m a l body weight. This study e x p l o r e d the extent to which a large sample o f college students perceived themselves as overweight in relation to their weight status as d e t e r m i n e d by sell-reported height a n d weight values. T h e m a g n i t u d e o f dissatisfaction with c u r r e n t weight status was also examined. T h e s e findings may be helpful to educators, dietitians, counselors, a n d o t h e r health professionals c o n c e r n e d with the prevention, identification, a n d t r e a t m e n t o f eating disorders.
METHODS Questionnaire
A questionnaire was designed to access college students' health-related knowledge, attitudes, and behaviors, as well as satisfaction with and use of their campus health services. Students also were asked to report their height and weight when wearing usual indoor clothing without shoes. The instrument was developed by the investigators, inspected for content validity by a panel representing a broad spectrum of university health services and organizations, and modified according to consensus. The survey instrument was pilot tested with approximately 50 students enrolled in an undergraduate course in order to determine readability, clarity, and time necessary to complete the questionnaire. The survey population consisted of students (n = 2, 610) who were selected randomly through a computer that was programmed to pool 10 percent of undergraduate students at a large midwestern university. The students received the mailed questionnaire accompanied by a cover letter explaining the purpose of the survey. They were assured of confidentiality, and were requested to complete the survey and to return it to the investigators within one week. As an incentive for participation, four $25 prizes were offered to students chosen randomly from among those who completed the questionnaire. Nonrespondents received a postcard reminder two to three weeks after the initial mailing. Only questionnaires received within six weeks of the original mailing were included in the data analysis. A total of 1,226 out of 2,610 undergraduates returned the questionnaire. Of these, 103 did not provide complete data, leaving a total of 1,123
John P. Sciaccaet al.
161
usable questionnaires (43% response rate). Because of possible bias associated with this moderate response rate, survey respondents were compared to the university's entire undergraduate population on selected sociodemographic characteristics. There were no differences between respondents and the entire undergraduate population regarding year in school, major area of study, residence, and age. The survey respondents appeared to be representative of the undergraduate student body, although females were slightly over-represented in the respondent group (48.1%) in relation to their actual percentage of enrollment on campus (41.3%), while males were under represented (51.9% vs. 58.7%). Ninety-three percent of the respondents were Caucasian, compared to 94 percent of the total undergraduate population, and over 90 percent of the students were between the ages of 18 and 22.
Comparison Between Self-Reported Height and Weight and Actual Height and Weight Because height and weight in this study were self-reported and were not actually measured by the investigators, an additional 70 undergraduates from the same population were sampled and asked to report their height and weight without previous knowledge that these parameters would be measured. Within 24 hours of their self-report, height and weight were assessed (minus shoes) by means of a balance-beam scale with an attached height rod. Correlation coefficients and slopes of the regression lines were computed between reported height and measured weight (r = .981, p < .0001, slope = 1.026) and between reported weight and measured weight (r = .987, p < .0001, slope = 1.002). The slopes of the regression lines were not significantly different from one another, and there were no differences between the mean reported indices and the mean measured indices. Extrapolating these results to our sample suggested that the use of reported height and weight was an acceptable means of estimating the true Body Mass Index (BMI) for this population.
Body Mass Index Measurements BMI was determined for each respondent using the Quetelet index [weight(kg)/height(m)2] '°. The BMI appears to be the most satisfactory measure of relative body weight in the absence of body composition measures. Bray" has established BMI standards for health to define acceptable ranges Ibr height-weight combinations (Table 1).
Statistical Analysis Frequency data were tested [br significance with the Chi-square test. The one-way analysis of variance (ANOVA) procedure was used to compare
162
J O U R N A L OF COMMUNITY H E A L T H
TABLE 1 Frequency a n d Percentage of Students in Four Weight Categories Based O n Body Mass I n d e x (BMI) Calculated Using Self-Reported H e i g h t a n d Weight ~ Males B M I Weight Category
Underweight Acceptable Overweight Obese Totals
Females
BMI +
n
%
BM1 +
n
%
<20.0 20.0-25.0 25.1-30.0 >30.0
48 416 101 13 578
8.3 72.0 17.5 2.2 100
< 19.0 19.0-24.0 24.1-30.0 >30.0
76 378 74 17 545
13.9 69.4 13.6 3.1 100
+BMI - Weight (kg) Height (m) ~
the means of dependent variables between groups. Post hoc analyses were performed with the Scheffe multiple comparison test when the F ratio for the ANOVA was significant at p < .05.
RESULTS Actual Weight Status T h e n u m b e r a n d percentage o f female a n d male students categorized as u n d e r w e i g h t , acceptable weight, overweight, a n d obese acc o r d i n g to the standards established by Bray (11) are shown in Table 1. T h e majority o f both females and males were classified as n o r m a l (acceptable weight); only 16.7 percent of females a n d 19.7 percent o f males e x c e e d e d n o r m a l BMI standards. Very few female or male students were obese, as d e t e r m i n e d by a BMI greater than 30. A h i g h e r p e r c e n t a g e o f females (n = 76, 13.9%) was u n d e r w e i g h t c o m p a r e d to males (n = 48, 8.3%, x 2 = 9.0, d f = 1, p < .01). As expected, the m e a n BMI was greater for males (mean = 23.1, SD = 2.7) than for females (mean = 21.8, SD = 3.2, F(1,1122) = 56.4, p < .001).
Perceived Weight Status Students were asked to describe their present perceived weight status according to t b u r possible categories: "underweight," "about what
John P. Sciacca et al.
163
TABLE 2
Mean Body Mass Index by Self-Described Weight Status in Males and Females. Males* Self-described Weight Status
Underweight About what is recommended for m y height Moderately overweight Great deal overweight
Females +
n
Mean B M I
SD
n
Mean B M I
SD
76 363
20.5 22.7
1.9 1.8
21 305
18.3 20.4
1.2 1.8
131
25.7
2.5
194
23.2
2.3
7
30.1
5.6
25
29.9
3.9
*Mean B M I f o r all four weight categories among males significantly difterent from each other (p < .05). +Mean BMI for all four weight categories among |emales significantly different tYom each other (p < .05).
is recommended for my height," "moderately overweight," and "a great deal overweight." The results shown in Table 2 reveal that in both males and females, mean BMIs rose for each progressively greater "perceived weight" category and differed significantly from each other (males, F(3,576 = 147.5, p < .001; females, F(3,541) = 212.5, p < .001). Among males, the percentage of students in each weight category based on calculated BMI did not differ greatly from self-described weight status. Among females, however, a significantly greater number of students classified themselves as overweight than were considered overweight according to BMI calculations (x 2 = 94.2, df = 3, p < .0001) (Figure 1). Perceived Discomfort
Students were asked to report any discomfort they had experienced during the previous three months due to "weighing too much." Mean BMI was calculated for each of the following levels of discomfort: a) no discomfort, b) very little discomfort, c) a fair amount of discomfort, and d) quite a bit of discomfort. For both males and females, BMI rose with increasing magnitude of discomfort (Table 3). When the fre-
164
JOURNAL OF COMMUNITY HEALTH
FIGURE 1 Comparison o f self-described weight status to BMI standards in females, x ~ = 94.2, d f = 3, p < .0001. 69%
70-
BMI - determined weight status
/////
60-
50p.Z
W 0 n.W
13-
I
/ ~ / ~ / / , 56%
J Self-described weight status
///// /////
/////
40-
///// ///// / / / //
30-
35%
20/4%
/////
I0-
°°
nY~88~ /
0 UNDERWEIGHT
n
°o
305
A GREAT DEAL OVERWEIGHT
MODERATELY OVERWEIGHT
ACCEPTABLE OR RECOMMENDED
OR OBESE
TABLE 3 BMI in Males and Females by Degree o f Discomfort Experienced From Weighing too Much. Amount of Discomjbrt from Weighing Too Much None
Males Females
Very Little
Fair Amount
Quite a Bit
n
Mean BMI
n
Mean BMI
n
Mean BMI
n
Mean BMI
305 107
21.9 19.7
158 145
23.7 21.6
87 169
25.4 22.2
26 120
26.2 23.2
John P. Sciacca et al.
! 65
FIGURE 2 Reported degree of discomfort due to "weighing too much" in males and females, x 2 = 181.6, df = 3, p < .0001. 70-
Moles 60-
I Fem°res
I 53 %
50-
tZ
40-
W
L)
n,,W I1
3/%
30-
27%
27% 22 %
20%
20-
10-
0
r~: 305 NO DISCOMFORT
~ : 169
a : 145 A LITTLE DISCOMFORT
FAIR A M O U N T OF DISCOMFORT
QUITE A BIT OF DISCOMFORT
quency of reported discomfort was compared, only 20 percent of the males reported a "fair amount" or "quite a bit of discomfort" from weighing too much as compared to 53 percent of the females (x 2 = 181.6, d f = 3, p < .0001) (Figure 2).
DISCUSSION The results of this study revealed a significant incongruity between females' self-perception of body weight status and their BMI-determined weight category (Table 2; Figure 1). Only 16.7 percent of the females were overweight or obese according to BMI standards established by Bray", yet 40.2 percent described themselves as either "moderately overweight" or "a great deal overweight." In contrast 19.7 percent of the males were above BMI standards for acceptable weight, and 23.9 percent classified themselves as overweight. T h e 1983 Metropolitan Life height/weight tables recommend even higher weight levels
166
JOURNAL OF COMMUNITY HEALTH
tor each height category over than 1959 tables. If these more recent recommendations were used in this study, even fewer of our respondents who described themselves as overweight would actually be categorized as such. Although the nature of this study prevents one from ascertaining if the individuals who were classified as moderately or a great deal overweight were indeed so (for example, they may have had a more active lifestyle and therefore had more muscle than fat), prior analyses suggest that they were correctly classified. Students in this study were asked how often they participated in an aerobic exercise lasting at least 30 minutes per session. Approximately 73 percent of the males and 80 percent of the females reported exercising less than three times per weeU ~. While BMI is a limited index for determining body composition, these findings provide suggestive evidence that those individuals with higher BMI values were likely to be overweight or obese rather than to have excessive muscle mass. These results support the findings of Zuckerman et al. ':~, who reported that although only 10 percent of the college students surveyed were overweight, 50 percent of the females and 13 percent of the males perceived that they were overweight. Furthermore, Bailey and Goldberg ~ reported that 85 percent of tire college women in their study indicated that they were slightly or grossly overweight despite the fact that they averaged 95 percent of their ideal weight fi)r t'rame using the 1983 Metropolitan Life height/weight tables. Rucinski,':' who studied the relationship between body image and dietary intake in competitive ice skaters, reported that male and female ice skaters seem to have different views of ideal body size. Individuals who perceive of themselves as overweight may be prone to chronic dieting and eating disorders. ~"': Jakobovits et al. (18), for example, reported that 50 percent of female college students were chronic dieters. Rosen and Gross '~' found that 63 percent of high school girls were on weight-reducing regimens, even though most of them were already of normal weight. Interestingly, the onset of bulimia is usually preceded by a period of voluntary dieting. '-'~' We found that a higher percentage of males reported that they were underweight than actually were, on the basis of calculated BMI. A smaller percentage of the females, however, reported being underweight than they actually were, according to their BMI. These findings support those of Gray, ~ who found that male college sludents perceived themselves as lighter while female students perceived themselves as heavier than their actual weight.
John P. Sciacca et al.
167
It was also found that women of normal weight reported experiencing discomfort from weighing too much (Table 3; Figure 2). Over 53 percent reported experiencing a "fair amount" or "quite a bit of discomfort" within the previous three months from weighing too much, even though over 80 percent were considered either underweight or of normal weight. Interestingly, although mean BMI increased progressively according to the reported degree of discomfort felt from weighing too much, the mean BMI for each reported level of discomfort among women was always within the range of acceptable weight. Thus it appears that many females experience discomfort from perceived excessive weight even though, in reality, their weight is normal for their height. Although the nature of the discomfort was not specifically defined, its presence may lead to the development or maintenance of fad dieting and anorexic and/or bulimic behavior, problems which are currently prevalent among college women? These findings could help explain why the incidence of eating disorders is higher among college-age women than among their male peersY ~'2~7 In summary, this study showed that the proportion of overweight and obese students, as determined by BMI standards, was similar for both males and females. Even so, a significantly higher percentage of women than men considered themselves overweight. Furthermore, significantly more women than men experienced discomfort due to perceived excessive weight. A distorted body image may serve as a marker for individuals at risk for eating disorders. Eating disorders may be prevented if early warning signs are recognized and if individuals are provided with realistic weight expectations regarding their bodies, u~ We recommend support for initiatives aimed at 1) recognizing the precursors to eating disorders, 2) establishing eating disorders/body weight curricula in schools, and 3) modifying unhealthful societal expectations regarding body weight. Unhealthful eating patterns may be changed if students' perceptions of their weight, or their responses to these perceptions, are modified. REFERENCES 1. Halmi, KA, Classification of eating disorders, lntJ Eat Dis, 2: 21-25, 1983. 2. Johnson C, Lewis, C., Love, S, Stuckey, M, and Lewis, L, Incidence and correlations of bulemic behavior in a female high school population..] Youth Adol, 13: 15-26, 1982. 3. Vann, BH: Gender, Self-Perception and Eatin~ Behavior. Arizona: Doctoral Dissertation, University of Arizona, 1987.
168
JOURNAL OF COMMUNITY HEALTH
4. Garner, DM, Garfinkel, PE, Schwartz, D, and Thompson, M, Cultural expectations of thinness in women. Psvchol Rep, 47: 483-491, 1980. 5. Sours, JA: Starving to Death in a Sea of Objects: The Anorexia Nervosa Syndrome. New York: Aronson, 1980, p. 3. 6. Squires: The Sle~uter Balance. New York: Pinnacle, 1983. 7. Story, M, Nutrition management and dietary treatment of" bulimia. J Am Diet Assoc, 86: 517-519, 1986. 8. Slade, PD and Dewey, ME, Development and preliminary validation of SCANS: A screening instrument for identifying individuals at risk of developing anorexia and bulimia nervosa. Int J Eat Dis, 5: 517-538, 1986. 9. Williams, RL, Schaefer, CA, Shisslak, CM, Gronwaldt, VH, and Comerci, GD, Eating attitudes and behaviors in adolescent women: Discrimination normals, dieters, and suspected bulimics using the eating attitudes test and eating disorder inventory. IntJ Eat Dis, 5: 879-894, 1986. 10. Lee, J and Kolonel, L, Are body mass indices interchangeable in measuring obesity disease associations? Am J Public Health, 74: 376-377, 1984. 11. Bray, GA, Obesity in America. lntJ Obesity, 3: 363-375, 1979. 12. Melby, C, Femea, P, and Sciacca, J, Reported dietary and exercise behaviors, beliefs and knowledge among university undergraduates. Nut Res, 6: 799-808, 1986. 13. Zuckerman, DM, Colby, A, Ware, NC, and Lazerson, JS, The prevalence of bulimia among college students. Am J Public Health, 76:1135-1137, 1986. 14. Bailey, S and Goldberg, JP, Eating patterns and weight concerns of college women. J Am Diet Assoc, 89: 95-96, 1989. 15. Rucinski, A, Relationship of body image and dietary intake of competitive ice skaters. J Am Diet Assoc, 89: 98-100, 1989. 16. Polivy,J, Garner, DM, and Garfinkel, PE, Causes fi)r thin female physiques. In CP Herman, M Zanna, and ET Higgins (Eds.) Physical Appearance, Stzgma and Social Behavior: The Ontario Symposium, Volume 3. New Jersey: Lawrence Erlbaum, 1988, p. 101. 17. Fallon, AE and Rozin, P, Sex differences in perceptions of desirable body shape. J Abnorm Psychol, 94: 102-105, 1985. 18. Jakobovits, SC, Halstead, P, Kelley, L, Roe, DA, and Young, CM, Eating habits and nutrients intakes of college women over a thirty year period. J Am Diet Assoc, 71:405-411, 1977. 19. Rosen, JE and Gross, J, Prevalence of weight reducing and weight gaining in adolescent girls and boys. Health P~ychol, 6; 131-147, 1987. 20. Gandour, MJ, Bulimia: Clinical description, assessment, etiology, and treatment, lntJ Eat DIS, 3: 3-38, 1984. 21. Gray, SH, Social aspects of body image: Perception of normalcy of weight and affect of college undergraduates. Percept Mot Ski&, 45: 1035-1040, 1977. 22. Drewnowski, A, Hopkins, SA, and Kessler, RC, The prevalence of bulimia nervosa in the U.S. college student population. A m J Public Health, 78, 1322-1325, 1988. 23. Howat, PM and Saxton, AM, The incidence of bulimic behavior in a secondary and university school population.J Youth Adol, 17: 221-231, 1983. 24. Lindsey, BJ and Janz, KF, A healthy connection: Helping physical educators address eating disorders. J Phys Ed Rec Dan, 56: 41, 1985. 25. Nero, S, Bulemic symptoms: Prevalence and ethnic differences among women. IntJ Eat Diz, 4: 151-168, 1985. 26. Pyle, RL, Mitchell, JE, Eckert, ED, Halvorson, PA, Newman, PA, and Goff, GM, The incidence of bulimia in freshman college students, lntJ Eat Dis, 2: 75-85, 1987. 27. Schotte, DE and Stunkart, AJ, Bulimia vs bulimic behaviors on a college campus. JAMA, 258: 1213-1215, 1987. 28. Adams, LB and Shafer, MB, Early manifestations of eating disorders in adolesceuts: Defining those at risk. JNutrEd, 20: 307-313, 1988.