Interventions to Reduce Sexual Prejudice: A Study-Space Analysis and Meta-Analytic Review

Abstract

Sexual prejudice is an important threat to the physical and mental well-being of lesbians, gay men, and bisexual people. Therefore, we reviewed the effectiveness of interventions designed to reduce such prejudice. A study-space analysis was performed on published and unpublished papers from all over the world to identify well-studied and underexplored issues. Most studies were conducted with North American undergraduates and were educational in nature. Dissertations were often innovative and well designed but were rarely published. We then performed meta-analyses on sets of comparable studies. Education, contact with gay people, and combining contact with education had a medium-size effect on several measures of sexual prejudice. The manipulation of social norms was effective in reducing antigay behavior. Other promising interventions, such as the use of entertainment media to promote tolerance, need further investigation. More research is also needed on populations other than American students, particularly groups who may have higher levels of sexual prejudice.

Acknowledgement

We wish to thank Ellie Brodie for her patient help in testing the reliability of the inclusion criteria and Robert Nash, Ilka Gleibs, and Y. Gavriel Ansara for their helpful feedback and methodological suggestions.

Notes

¹The term homophobia is usually credited to George Weinberg, who used it in his 1972 book Society and the Healthy Homosexual. However, the term was in use earlier, a fact readily acknowledged by Weinberg himself (personal communication cited in Herek, 2004). The earliest academic paper using this word seems to be Kenneth Smith's (1971) “Homophobia: A Tentative Personality Profile.”

²Note that homosexuality had previously enabled a similar focus on the individual psyche at the expense of other issues (Sell, 1997).

³Only the variables used in the final version of this paper are reported here. The initial coding scheme, which is slightly more detailed, is available from the authors upon request.

Note. This classification is based on Paluck and Green (2009).

⁴We used the same symbol (d) both for the effect sizes of individual studies and for summary effect sizes; we appreciated that the context would always be clear enough to avoid confusion.

⁵Begg and Mazumdar's τ should be interpreted with caution when the number of studies is small; see Field and Gillett (2010) for details.

⁶Rosenthal's fail-safe number is meaningless in this case: Because the mean effect size is zero, there is no need to consider the possibility of unpublished studies with nonsignificant results.

⁷One might question whether accountability induces a change in people's attitudes or merely a socially desirable behavior. See Crandall, Eshleman, and O'Brien (2002) for a more sophisticated view on the matter.

k = the number of studies on which d is based.

References marked with an asterisk indicate studies included in the study space-analysis and meta-analysis.

This study was conducted in partial fulfillment of the requirements for a doctoral degree by the first author, under the supervision of the third, and with the assistance of the second.