Women faculty in predominantly male departments at a large university were interviewed and responded to paper-and-pencil instruments in a test of Laws' analysis of the necessity for the “token women” adaptation to their marginality for success in academia. By multiple criteria, three clusters were identified: (a) token women, who accepted academia as a meritocracy, were aware of little sex discrimination and belonged to no feminist group; (b) non-token women, who disagreed with academia as a meritocracy and were aware of sex discrimination; and (c) women with mixed or moderate orientations. (Membership in feminist groups was found in the latter two clusters.) Women in the three clusters did not differ significantly by academic rank or marital status and only marginally by age and longevity in academia. As predicted, however, they did differ by tenure status. Contrary to Laws' analysis, token women were not more likely to have had a sponsor, which was significantly related only to rank. Women in the three clusters were equally accurate in recognizing male-female status discrepancies. Their differing definitions of sex discrimination were revealed in differential bias when choosing among alternatives of indeterminate correctness. Token women minimized such discrepancies, in line with beliefs attributing them to women, rather than the system. Others maximized such discrepancies.