The "unisex phantom," sexual dimorphism, and proportional growth assessment

The "unisex phantom" tactic of Ross and associates attempts to examine proportional growth, expressing body measurements relative to a phantom created from arbitrary mean dimensions and the observed coefficients of variation for an arbitrary sample of adult men and women. Fundamental assumptions of the model are (1) that data treated in unisex fashion have the normal distribution required of Z-type statistics throughout the period of growth, and (2) that it is reasonable to consider anthropometric measurements in all populations (regardless of ultimate size) as growing toward the common height chosen for the phantom. The validity and possible usefulness of this approach has been tested using anthropometric data collected on 546 francophone primary school children from the Trois Rivières district who had been measured repeatedly from 6 to 12 years of age. Over this age range, the mixing of data for girls and boys does not create a bimodal distribution, and has only a marginal effect upon skewing and kurtosis; however, the requirement of a normal distribution is not satisfied by quite a number of common anthropometric measurements even at this age. Moreover, application of the unisex phantom procedure to the Trois Rivières sample does little to clarify anticipated sex-related differences in regional growth, and it is argued that univariate standardization against a power function of an arbitrary adult height may not provide the best method of examining the multivariate problem of growth. Comparison of results with data from Saskatoon, previously treated by unisex phantom methodology, reveals puzzling inconsistencies; it is suggested that interlaboratory differences of methodology rather than ethnic differences are responsible for apparent discrepancies in growth patterns. It is urged that interlaboratory validation of techniques and landmarks should precede the ascription of apparent differences in body build to constitutional factors.