A Multivariate Method of Classif in Bod Form Variation for Sizing Women's Apparel

The purpose of this study was to develop a multivariate method for structuring a sizing system for women's apparel. The I 977 suvey of U.S. Army women was the data base. The sample was limited to 1217 subjects who were White or Black and 17-3 5 years old. The methodology underlying PS 42-70, the current U.S. sizing standard, was studied relative to recent changes in body form classification, apparel production, and merchandising which support development of multivariate methodology. Examination of PS 42-70 as a classification system for the sample showed inadequate proportioning of overall body dimensions, particularlyshoulder breadth and body segment lengths. The experimental method used components 1 and 2 to summarize body form variation as the relationship between body thickness and length. A Principal Component Sizing System (PCSS) was constructed by partitioning the distribution of principal components 1 by 2 into 26 sizes. A per subject comparison of "goodness of fit" of PS 42-70 and PCSS showed PCSS consistently closer to subjecfs' overall body proportions than PS 42-70. The multivariate relationship between body thickness and length is fundamental to body form and size variation. Our results indicate that PCSS better represents that relationship for the sample studied.