Gender stereotypes are not accurate: A replication of Martin (1987) using diagnostic vs. self-report and behavioral criteria

Women and men (58 females and 42 males) reported whether they possess each of several stereotypic and neutral traits (self-report criterion) and recorded the percentage of each gender they believed display each trait (diagnostic). Observations of subjects (57 males and 49 females) discussing controversial topics yielded the behavioral criterion. (Approximately 81% of the total sample was European-American, 8% African-American, 2% Hispanic, 1% Asian, and the remainder of “other” ethnicity.) Gender differences were larger and more inaccurate in stereotypes (diagnostic) than in “reality” (self-report criterion), but behavioral criterion results were mixed. Difficulties in specifying criteria for “actual differences,” including behavioral criteria, and recent work related to C. L. Martin's [(1987) “A Ratio Measure of Gender Stereotyping,” Journal of Personality and Social Psychology, Vol. 52, pp. 489–499] paradigm (and results) are considered. Implications of results for the importance of replication and for research on accuracy of stereotypes are also discussed.