I conduct an experiment to test the impact of relative power differential on the relationship between gender and role-taking accuracy. First in an 80 subject study, and then in the current study, role-taking accuracy is conceptualized as the accuracy with which one can predict the behavior of another or others. In Study 1, I examined self-evaluative measures of role-taking ability and found that self-evaluative measures of role-taking do not correlate with actual role-taking accuracy. In addition, women were more accurate role-takers than were men in same-gender dyads regardless of the existence of a prior relationship between the two individuals. This prior experimental research showed that female friends were much more accurate role-takers than were male friends. In fact, female strangers were more accurate role-takers than were male friends. It is my conjecture however, that role taking ability is not directly connected to gender; rather I propose that it is a situationally prompted ability based on the need for individuals of relatively less power to predict the behavior of individuals with relatively more power. In other words, while women are, indeed, better role takers, this is not a general ability; rather it is prompted by their relatively low positions of power. In Study 2, I examine role-taking accuracy under conditions in which differential power is assigned to one member of a dyad and established through interaction. I predict that power position will account for variability in role-taking accuracy, but gender will not. I tested this hypothesis using power balanced and power-imbalanced, task-oriented, same and cross gender dyads. I found that power position does account for variation in role-taking accuracy while gender and gender composition of the dyad do not account for variation in role-taking accuracy.
