This paper examines the religion-human capital link, examining a recent household survey for Ghana. Insights from the recent anthropological literature leads to a prediction of Islam being associated with lower human capital levels than Christianity, since Islam, perhaps surprisingly, may be clustered together with Traditional/Animist religion within the group of orally based religions for the case of Ghana. While previous studies typically have only considered the main religions, thereby not allowing for heterogeneous associations in the links at the sub-group level, and also have not allowed religious affiliation to be endogenously determined, these possibilities are explored here, as well. I find a strong association between individual religious affiliation and human capital as measured by years of schooling, with Christians as a group being more literate and having completed more years of schooling than Muslims and Animists / Traditionalists, thus confirming the predictions from the conceptual framework. At the same time, there is a great deal of heterogeneity in the strength of this relationship within different types of Christianity. The instrumental variables estimation strategy proves to be preferable to OLS, while at the same yielding higher associations in the religion-human capital relations ship. In turn, this indicates that previous studies, which have typically used OLS, may have systematically underestimated the strength of the religion-human capital link. Directions for future research are also presented.