Gender threat was induced by providing participants with false feedback on the Bem Sex Role Inventory (BSRI; Bem, 1974 (link)). The BSRI has participants rate themselves on 60 personality traits: 20 that are stereotypically masculine (e.g., assertive, self-reliant, analytical), 20 that are stereotypically feminine (e.g., affectionate, gentle, cheerful), and 20 that are gender-neutral (e.g., reliable, sincere, conscientious). All traits are positive in valence, and their order of presentation was randomized. Participants recorded their responses on a scale from 1 (never true of you) to 7 (always true of you).
Once participants completed the BSRI, a score from 0 to 100 was calculated, such that higher scores indicated more agreement with the masculine traits and lower scores indicated more agreement with the feminine traits; neutral traits were excluded from scoring. For male participants, manhood threat was induced by subtracting 30 points from their actual score, thereby placing them closer in personality to a stereotypical woman. For female participants, womanhood threat was induced by adding 30 points to their actual score, thereby placing them closer in personality to a stereotypical man. In this way, feedback provided to participants was anchored on their actual levels of (stereotypical) masculinity and femininity, helping to ensure that no participant received scores vastly—and thus unrealistically—discrepant from their actual responses. Scores could be no lower than 3 or higher than 97.
Participants in the threat condition saw their adjusted score juxtaposed with the putative score of the average person of their gender (80 for men and 32 for women). Even the most masculine man in the threat condition would receive a score of 70 (100 minus 30), which is below the average man’s score. In light of research showing that women tend to be higher in androgyny than men (Donnelly & Twenge, 2017 (link)), we adjusted the average women’s score to be further away from 0 (i.e., 32) than the average man’s score was from 100 (i.e., 80). Participants in the no-threat condition received no BSRI feedback, but all participants received the same instructions prior to taking the test. In a pilot test, participants were asked after the manipulation whether they suspected the true purpose of the research. None correctly guessed the hypothesis or purpose of the manipulation.
Once participants completed the BSRI, a score from 0 to 100 was calculated, such that higher scores indicated more agreement with the masculine traits and lower scores indicated more agreement with the feminine traits; neutral traits were excluded from scoring. For male participants, manhood threat was induced by subtracting 30 points from their actual score, thereby placing them closer in personality to a stereotypical woman. For female participants, womanhood threat was induced by adding 30 points to their actual score, thereby placing them closer in personality to a stereotypical man. In this way, feedback provided to participants was anchored on their actual levels of (stereotypical) masculinity and femininity, helping to ensure that no participant received scores vastly—and thus unrealistically—discrepant from their actual responses. Scores could be no lower than 3 or higher than 97.
Participants in the threat condition saw their adjusted score juxtaposed with the putative score of the average person of their gender (80 for men and 32 for women). Even the most masculine man in the threat condition would receive a score of 70 (100 minus 30), which is below the average man’s score. In light of research showing that women tend to be higher in androgyny than men (Donnelly & Twenge, 2017 (link)), we adjusted the average women’s score to be further away from 0 (i.e., 32) than the average man’s score was from 100 (i.e., 80). Participants in the no-threat condition received no BSRI feedback, but all participants received the same instructions prior to taking the test. In a pilot test, participants were asked after the manipulation whether they suspected the true purpose of the research. None correctly guessed the hypothesis or purpose of the manipulation.
