Predicting heterosexual reactions to coming out in China

For many Chinese sexual minority men, coming out-disclosing their sexual orientation-is stressful because of concerns about how heterosexual recipients may react. This study examined predictors of those reactions and applied machine learning to model them. Between August 2022 and January 2023, we screened 54,902 individuals on a Chinese crowdsourcing platform and recruited 650 Chinese cisgender heterosexual participants who had experienced a coming-out encounter with a sexual minority man. Participants reported on 85 personal, contextual, and behavioral factors surrounding the encounter. Bivariate analyses and machine learning models revealed that background characteristics-such as the recipient's sex, age, and relationship to the discloser-were the strongest predictors of reactions. Smaller contributions emerged from select contextual parameters and coming-out strategies: negative reactions were linked to disclosures involving a love confession, assault, or substance use by disclosers; positive reactions were associated with recipients anticipating the disclosure and disclosers referencing a same-sex romantic relationship. The machine learning models captured complex interactions among predictors. These models achieved moderate predictive accuracy (ACC = 0.62-0.65, R² = 0.20-0.31), comparable to that reported in prior machine learning studies in social and personality psychology. While recipients' reactions are predictable, improving coming-out experiences in high-stigma contexts requires attention to the recipient-side processes that sustain or challenge heteronormative norms.