Gender, status, and behavior in task situations

Gender is among the most extensively investigated constructs of the social sciences. It is a basis of macrosocial organization and an individual characteristic that shapes microsocial interaction. The social psychological literature contains at least four alter native conceptions of how gender operates in face-to-face interaction: the normative-influence, the informational-influence, the weighted-averaging, and the expectation-states hypotheses. Using a rich set of experimental data on mixed-sex problem-solving discussions entailing ''feminine,'' gender-neutral, and ''masculine'' tasks, we investigated the relative empirical adequacy of these four hypotheses. Dependent variables included amount of time speaking and frequencies of gesturing, speech initiations, looking while speaking, looking while listening, chin thrusts, smiling, laughing, and self-touching. Dovidio and his colleagues reported that for some measures, but not for others, statistical interactions exist between the actor's gender and the gender linkage of the task. Using a more precise theoretical analysis, we found that the expectation-states hypothesis describes the first category of variables extremely well, while the normative-influence hypothesis describes the second category extremely well. The weighted-averaging hypothesis deviates systematically from the empirical data collected by Dovidio and his colleagues. We conjecture that ''gender effects'' in putatively task-oriented interaction emanate from gender-as-worth and from gender-as-personal-identity. Some implications for further research are discussed.