On assessing “androgyny”

Research with personality instruments containing separate masculinity and femininity scales found them to be essentially orthogonal, leading to the search for methods to express their conjoint relationships with other variables. The theoretical and/or empirical assumptions underlying three methods were explored: Bem's difference model; the absolute score method; and the difference/median method, which combines features of both. The balance and difference/median methods presuppose a unique combinatory model, while the absolute method detects a number of alternative possibilities. Illustrative empirical data, using the Personal Attributes Questionnaire, were used to demonstrate that the joint contributions of M and F are not uniform and, consequently, the greater heuristic value of the absolute method.