Document Type



The purpose of this study was to develop regression equations for predicting density and % body fat. Subjects, 532 males, were hydro- statically weighed, and circumference measurement recorded for thigh, hips (buttocks), iliac, abdomen, and chest. Age (range = 21 to 78 yrs), weight (56.2 to 152.8 kg), and height (146.1 to 193.0 cm) were also recorded. Percent body fat ranged from 2.0 to 48.8%. Step-wise multiple regression was used to select the best set of predictors of body density and % of body fat from the eight independent (predictor) measures. Capitalization on chance was negligible due to the favorable subject to predictor ratio (58 subjects per predictor). The regression equation (N = 462) developed for predicting % body fat was: % body fat = —47.371817 + (0.57914807 x abdomen) + (0.25189114 x hips) + (0.21366088 x iliac) — (0.35595404 x weight)[S.E.E. = 3.6, R = .839, R2adj = .702] Using the above equation on a cross-validation sample (N = 70) produced a predicted mean (±SD) of 25.6 ± 6.6% (validation sample mean = 26.0 ±6.6%) and a total error (S.E.) of 4-4%. The use of three girth measurements and total body weight enabled us to develop equations for predicting body density and % body fat that are comparable in accuracy to skinfold calipers and should provide a viable alternative to skinfold measurements.