Moral elevation reduces prejudice against gay men

Abstract

Disgust is linked to social evaluation. People with higher disgust sensitivity exhibit more sexual prejudice, and inducing disgust increases sexual prejudice. We tested whether inducing moral elevation, the theoretical opposite of disgust, would reduce sexual prejudice. In four studies (N = 3622), we induced elevation with inspiring videos and then measured sexual prejudice with implicit and explicit measures. Compared to control videos that elicited no particular affective state, we found that elevation reduced implicit and explicit sexual prejudice, albeit very slightly. No effect was observed when the target of social evaluation was changed to race (Black–White). Inducing amusement, another positive emotion, did not significantly affect sexual prejudice. We conclude that elevation weakly but reliably reduces prejudice towards gay men.

Keywords:

• Attitudes
• Prejudice
• Sexual orientation
• Elevation
• Implicit Association Test

This project was supported by a gift from Project Implicit. Lai is a consultant and Nosek is an officer of Project Implicit, Inc., a non-profit organisation that includes in its mission ‘To develop and deliver methods for investigating and applying phenomena of implicit social cognition, including especially phenomena of implicit bias based on age, race, gender or other factors'. Lai was supported by a National Science Foundation Graduate Fellowship. Contributions: Conceived research: all authors; Designed research: C.K.L., J.H.; Performed research: C.K.L.; Analysed data: C.K.L.; Wrote paper: C.K.L., B.A.N.; Revised paper: all authors.

This project was supported by a gift from Project Implicit. Lai is a consultant and Nosek is an officer of Project Implicit, Inc., a non-profit organisation that includes in its mission ‘To develop and deliver methods for investigating and applying phenomena of implicit social cognition, including especially phenomena of implicit bias based on age, race, gender or other factors'. Lai was supported by a National Science Foundation Graduate Fellowship. Contributions: Conceived research: all authors; Designed research: C.K.L., J.H.; Performed research: C.K.L.; Analysed data: C.K.L.; Wrote paper: C.K.L., B.A.N.; Revised paper: all authors.

Notes

¹ We planned to collect 300 participants (150/condition) in Study 1, 800 participants (100/condition) in Study 2, 448 participants in Study 3 (112/condition) and 1868 participants (467/condition) in Study 4. The planned sample size in Study 3 was for 80% power to detect an effect of the same size as Study 1 (d = .36), and the planned sample size in Study 4 was for 95% power to detect effects of d = .24. We stopped the automated data collection when we observed that our planned sample size was exceeded. In Study 3, data collection ended early due to a miscalculation.

² See http://openscienceframework.org/project/fG5xB/ for the results of all analyses we conducted on attrition. We found evidence for differential attrition by condition in Experiments 3 and 4 (ps = .03, .02). In those two experiments, the control condition elicited greater attrition than the elevation conditions. However, there was no evidence for experimental condition leading to differential attrition by demographics (i.e., age, religiosity, gender, ideology), suggesting that the results were not attributable to differences in sample.