Using data for a large cross-country sample, a reasonable model is estimated to judge the effect of adult educational attainment on prevalence of HIV. Three main points are noted. First, there is an indication of a significantly negative effect of educational attainment on HIV prevalence. Second, magnitude of the impact appears sizable. Third, a simple test suggests that the model does not have any major specification problem. Three additional aspects may also be interesting. First, the impact of per capita income is negative but statistically weak. Second, the effect of health expenditures lacks significance and does not carry the expected sign. Third, the proportion of population that is Muslim has a significantly negative association with HIV prevalence, perhaps reflecting the widely-shared view that male circumcision lowers acquisition of the disease.