- Wühr, P., Lange, B. P., & Schwarz, S. (2017). Tears or fears? Comparing gender stereotypes about movie preferences to actual preferences. Frontiers in psychology, 8, 428.
This study investigated the accuracy of gender-specific stereotypes about movie-genre preferences for 17 genres. In Study 1, female and male participants rated the extent to which 17 movie genres are preferred by women or men. In Study 2, another sample of female and male participants rated their own preference for each genre. There were three notable results. First, Study 1 revealed the existence of gender stereotypes for the majority of genres (i.e., for 15 of 17 genres). Second, Study 2 revealed the existence of actual gender differences in preferences for the majority of genres (i.e., for 11 of 17 genres). Third, in order to assess the accuracy of gender stereotypes on movie preferences, we compared the results of both studies and found that the majority of gender stereotypes were accurate in direction, but inaccurate in size. In particular, the stereotypes overestimated actual gender differences for the majority of movie genres (i.e., 10 of 17). Practical and theoretical implications of these findings are discussed.
The study didn’t really properly showcase the amazing accuracy of sex stereotypes in their study, but they published their summary stats in a table, so I will help them.
r = .85! The correlation size is not even reported in the paper because they relied on a weird ANOVA approach. While this value might seem very high to you, it is in fact quite common in the study of sex stereotype accuracy. The median of 14 studies recently meta-analyzed was r = 0 .79 (Jussim et al 2018).
The authors conclusion about the stereotypes being larger than the real differences is questionable because they didn’t use the same scale for measuring them. If one wants to investigation the matter of exaggeration — generally not found — one should use ratio scale data which avoids these problems. We did this in our previous stereotype study with immigrant groups in Denmark and found estimation of differences not overestimation (Kirkegaard & Bjerrekær 2016).