Factor analysis and multivariate scaling of anthropometric variables for the assessment of body composition

Abstract

This study was designed to identify the body composition factors measured by skinfold fat (S), girth (G), and diameter (D) variables and to provide multivariate scaling models that measure fat (F) and lean body weight (LBW) factors. A total of 7S, 11G, and 7D measurements were determined on 83 young women (YW) and 95 young men (YM). Body density (BD) was determined by the hydrostatic technique. All variables were factor analyzed, and F and LBW were used to validate isolated factors. The YM data showed one F factor and two LBW factors. LBW factor I measured muscle mass and bone structure; LBW factor II, bone D. The YW data isolated four factors: the same three of YM plus a second F factor specific to the pelvic girdle. Factor score models were used to develop scaling strategies to measure F and LBW factors. Full models of all 27 variables and restricted models of 14 S, G, and D variables were developed; the correlations between factor scores scaled from the full and restricted models ranged from 0.84 to 0.98. The multiple correlation between factor scores calculated from the restricted models and laboratory-determined F and LBW ranged from 0.90 to 0.94 for YM and from 0.88 to 0.89 for YW. The multivariate scaling models provide a valid field method of measuring body composition.