This study compares self-categorization, expectation states, and role congruity theories' explanations for female influence. Male and female participants (N = 267) listened to a recording of a female speaker who used either tentative or assertive language under conditions that led participants to categorize her as a woman or as college-educated. There was no evidence that women were differentially influenced by the speaker's linguistic style or by the categorization. Men, however, were more influenced by the tentative speaker when she was categorized as a woman than as a college student. Men were more influenced by an assertive than tentative speaker when the speaker was categorized as a college student. Mediational findings provided evidence for self-categorization and expectation states, but not role congruity, processes.According to Lakoff (1973Lakoff ( , 1975, women are caught between two forms of male prejudice. When women use tentative language-hedges (e.g., sort of, you know), disclaimers (e.g., I'm not sure, I suppose), and tag questions (e.g., isn't it? right?)-men like them but perceive them as relatively unintelligent and incompetent, thus justifying their exclusion from serious discussion and power. When women eschew tentative language, men see them as intelligent and competent, but they are ostracized for being unfeminine; epithets reserved for such women include bully broad, iron maiden, and bitch.Research has verified Lakoff's contention. Wiley and Eskilson (1985) found that women who used assertive language were seen as successful managers, despite being seen as less likeable, than women who used tentative language. Juodvalkis, Grefe, Hogue, Svyantek, and DeLamarter (