A method of deriving subgroups of a population: A study of craniofacial taxonomy

Abstract

Subgroups within a population are often difficult to discover and describe except by subjective methods. In this study, cluster analysis (numerical taxonomy) methods were used on selected craniofacial measurements obtained from 308 North American White children of both sexes in the age range 6–18 to derive categories of skeletal facial types. Two different cluster analysis approaches were used in conjunction with a separate overall evaluation of facial balance, an independent measure of maxillo-mandibular relationship (AB/FOP), and a traditional classification (Angle). The categories derived rest on corroborative and overlapping evidence from each of those methods. The categories were examined to determine if it is possible to classify a sample by means of cluster analysis, the size and discreteness of each class, how they compare with the Angle classification of the same sample, and the percentage of individuals that may be identified unequivocally by this classification scheme. Five categories were obtained. Labeled Category A — Category E, they show some correspondence to the Angle Classes I, II, and III, but categories A, B, and C appear to be subgroups, heretofore undetected, of Angle Class II. Categories D and E correspond to Angle Classes I and III, respectively. The categories are more realistic and informative than the Angle classes. Each category is reported along with its distinguishing skeletal characteristics.