Notwithstanding the increasingly addressed effect from multilingualism on psychological variables in additive context (e.g. Dewaele & Van Oudenhoven, 2009; Wei & Hu, 2018), few studies explore the influence of English on homophobia, a psychology-rooted item (Weinberg, 1973) in an EFL context, China. Based on the newest wave of Chinese General Social Survey (CGSS) 2015, the current study attempts to present a most recent depiction of homophobia in China, and estimate the effect yielded by English proficiency and several other variables. According to the established hierarchical regression models, after controlling sociodemographic factors, the maximum of English proficiency's unique contribution can reach 1% (p < .0005), which is a small but typical effect size. Internet use and support of free speech can maximally explain 1.4% (p < .0005) and 1.3% (p < .0005) of variance in attitude toward homosexuality respectively, and the unique contribution of gender role belief remains marginal, ranging from 0.2% to 0.3% (p < .0005). The implications concerning sexual minorities are discussed, and suggestions for data representation such as avoiding a dichotomous interpretation of p and reporting useful effect sizes are provided.