Parent–child sexual-health communication is critical. Religious involvement is important in many African-American families, but can be a barrier to sexual-health communication. We tested a theory-based, culturally tailored intervention to increase sexual-abstinence communication among church-attending African-American parent–child dyads. In a randomized controlled trial, 613 parent–child dyads were randomly assigned to one of three 3-session interventions: (i) faith-based abstinence-only; (ii) non-faith-based abstinence-only; or (iii) attention-matched health-promotion control. Data were collected pre- and post-intervention, and 3-, 6-, 12- and 18-months post-intervention. Generalized-estimating-equations Poisson-regression models revealed no differences in communication by intervention arm. However, three-way condition × sex-of-child × sex-of-parent interactions on children’s reports of parent–child communication about puberty [IRR=0.065, 95% CI: (0.010, 0.414)], menstruation or wet dreams [IRR=0.103, 95% CI: (0.013, 0.825)] and dating [IRR=0.102, 95% CI: (0.016, 0.668)] indicated that the non-faith-based abstinence intervention’s effect on increasing communication was greater with daughters than with sons, when the parent was the father. This study highlights the importance of considering parent and child gender in the efficacy of parent–child interventions and the need to tailor interventions to increase fathers’ comfort with communication.