The Androgyny Dimension: A Comment on Stokes, Childs, and Fuehrer

Stokes, Childs, and Fuehrer (1981) have recently studied the relationship between the Bern Sex-Role Inventory (BSRI; Bern, 1974) and certain indices of "self disclosure." However, the Stokes et al. study contains a critical methodological flaw; their multiple regression analyses were not performed in the appropriate hierarchical fashion (the main effects were not partialed from the interaction term), and their results are uninterpretable. The purpose of this note is to point out what went wrong in the Stokes et al. study and to suggest an alternative technique. Particular attention will be devoted to the relevance hierarchical multiple regression analysis (with interaction terms) has as a general data-analytic device for investigating the psychometric properties of the contemporary sex role inventories. In addition, the current status of the contemporary sex role literature is briefly reviewed.