Document Type



This study was designed to investigate the relationship between body density and anthropometric dimensions in 8 to 11 year old boys for two samples of children located in different regions of the United States. It was hypothesized that similar anthropometric sites would be found to be the best predictors of body density in both samples leading to a common prediction equation for this age group. The samples consisted of 97 boys from Champaign-Urbana, Illinois and 86 boys from Davis. California. All subjects were measured for five skinfolds, five circumferences, six skeletal widths and body density estimated from hydrostatic weighing and pulmonary residual volume measurements. The precision of predicting body density from anthropometric dimensions was generally found to be as high as observed in adult samples. Skinfold thicknesses were better than circumferences or widths in predicting body density in each sample, although regression equations based on a combination of skinfolds, circumferences and widths resulted in slightly more precise estimates of body density. Two findings supported the development of a common equation for this age group. First, similar anthropometric sites were found to be applicable to both samples. Second, the curvilinear nature of the relation of skinfolds to body density was found in both samples. However, two additional findings prevented proposing one general equation: 1) a significant mean difference in predicting values between samples using the same anthropometric sites, and 2) a significant difference in density prediction error between samples. Separation of methodological from biological variability in the relation between density and anthropometry is needed prior to the development of a common approach to the prediction of body composition from anthropometry for this population.