Validity of self-reported weight and height and predictors of weight bias in female college students

Abstract

Main objectives of the present study were to examine (i) the accuracy of using female college students’ self-reports of weight and height in estimating rates of overweight and (ii) whether dietary restraint or Body Mass Index (BMI) was the most important predictor of weight underestimation. Participants were 209 female college students who were asked to report their weight and height on a questionnaire, while they were not told that their weight would be verified. Self-report screening was highly specific (98.9%) in identifying cases of normal weight, but only moderately sensitive (48.3%) in identifying cases of overweight. While dietary restraint was not an important predictor of weight underestimation, a higher BMI was an important predictor of weight underestimation. Our findings indicate that heavier female college students strongly underestimate their weight and that exclusive reliance on self-reports of weight and height can lead to erroneous prevalence estimates of overweight among female college students.

Introduction

Increasing rates of obesity is a global public health concern in industrialized societies (James, Leach, Kalamara, & Shayeghi, 2001). Accurate monitoring of body weight is important in the diagnosis, prevention and reduction of overweight and obesity. Many studies use self-reported height and weight values to assess health and classify weight status, as it is a convenient and cost-effective method of data collection. In industrialized societies, where there is a stigma attached to being overweight, people may tend to report values of weight and height that are close to their ideal values rather than to their actual ones. Studies among both adults (Alvarez-Torices, Franch-Nadal, Alvarez-Guisasola, Hernandez-Mejia, & Cueto-Espinar, 1993; Kuskowska-Wolk, Karlsson, Stolt, & Rossner, 1989; Niedhammer, Bugel, Bonenfant, Goldberg, & Leclerc, 2000; Stewart, Jackson, Ford, & Beaglehole, 1987; Taylor et al., 2006) and (late) adolescents (Brener, McManus, Galuska, Lowry, & Wechsler, 2003; Elgar, Roberts, Tudor-Smith, & Moore, 2005; Goodman, Hinden, & Khandelwal, 2000; Wang, Patterson, & Hills, 2002) have shown that people generally overestimate height and underreport weight. Although the correlation between self-reported and measured height and weight is typically high, self-report data, when used to screen for overweight, fail to detect a substantial proportion of morbid cases. Sensitivity rates are specifically low among older adolescents, with two studies even reporting sensitivity rates as low as 54.9% (Brener et al., 2003) and 52.2% (Elgar et al., 2005), missing almost half of the overweight or obese cases.

College students are frequently used in weight-related studies, and some of these studies use self-reported height and weight data to screen for overweight. Hence, information on the accuracy of students’ self-reports in estimating rates of overweight is important. We hypothesize that, in line with previous findings among older adolescents, sensitivity rates are low among college students. Overweight college students may probably be more aware of, and eager to adapt to, the current norms relating to body weight by virtue of their high level of education and motivation to succeed in life. Previous studies among college students have not reported about sensitivity rates as a screen for overweight. They merely focused on different situational and individual-difference variables that influence the accuracy of current weight reporting, showing that women (Betz, 1994; Imrhan, Imrhan, & Hart, 1996; Jacobson & DeBock, 2001), restrained eaters (Cash, Grant, Shovlin, & Lewis, 1992; McCabe, McFarlane, Polivy, & Olmsted, 2001; Shapiro & Anderson, 2003), heavier individuals (Cash, Counts, Hangen, & Huffine, 1989; Cash et al., 1992; McCabe et al., 2001), and individuals who are not aware that they will be weighed afterwards (Cash et al., 1989; Imrhan et al., 1996) have a tendency to underreport their weight. While one study found that with Body Mass Index (BMI) controlled, dietary restraint was no longer significantly related to underreporting (Cash et al., 1992), another study found that not BMI, but restraint status was the best predictor of weight underestimation (Shapiro & Anderson, 2003). The aims of the present study were to examine (i) the accuracy of using female college students’ self-reports of weight and height in estimating rates of overweight and (ii) whether dietary restraint or BMI was the most important predictor of weight underestimation.