Development and accuracy of self-reported weight in elementary school students

BACKGROUND: Self-reported measures are important tools in large-scale obesity prevention/treatment interventions, yet the ability of elementary school children to correctly self-report weight is unknown. The purpose of this study was to determine elementary school students' ability to accurately self-report their weight. METHODS: Self-reported and measured weight was recorded for first, third, and fifth grade students (N = 277) in Denver in May 2010. Self-reported values between the 3rd-99th percentiles of body weight-for-age were deemed “reasonable” and included in analysis. We compared mean measured and self-reported weight and the correlation between these variables for each grade level. Linear regression using weight-for-age z-score and the difference between measured and self-report determined if weight status influenced self-report accuracy. RESULTS: A reasonable value for self-reported weight was provided by 36, 71 and 87% of first, third, and fifth graders, respectively. All three grades significantly underreported weight. Mean values in kg (S.D.) for measured versus self-reported weight were 27.3 (5) vs. 24.6 (5.6), 34.8 (9.1) vs. 31.3 (6.1) and 43.9 (11) vs. 40.6 (9.8) for first, third and fifth graders, respectively. The correlation coefficient between measured and self-reported weight was 0.42, 0.72 and 0.93 in first, third and fifth graders, respectively, demonstrating improvement with age. In third and fifth graders, as weight-for-age z-score increased, self-reported accuracy decreased (r² of 0.183 and 0.09, respectively). CONCLUSIONS: Self-reported weight is not an appropriate assessment in first grade students; should be interpreted cautiously in third grade students; and is acceptable in fifth grade students. ACKNOWLEDGEMENT: Supported by NICHD/NCI/NIDDK R01HD057229