The development of measures of masculinity/femininity in psychology has reflected historical interest in categorizing gender differences. Recent measures have characterized masculinity as instrumental/agentic behavior. In this study, a traditional measure (the Minnesota Multiphasic Personality Inventory Masculinity/femininity scale) was pitted against a more recent measure (the Personal Attributes Questionnaire) in predicting instrumental behavior of mixed sex dyads in laboratory sex stereotyped tasks. Neither measure effectively predicted instrumental behavior. Rather, females performed better on the more complex but feminine stereotyped task, and males performed better on the masculine stereotyped task. The outcome of this study supports the need to view gender differences as dynamic phenomena influenced by individual choice, situational pressures, and interactional characteristics.