Data on individuals of immigrant origin are used in the epidemiological approach in comparative development for understanding cultural persistence, the determinants of cultural norms, and the effects of genetic traits. A widespread presumption is that this approach is exposed to attenuation bias. We describe how the increasing reliance on foreign ancestries to identify respondents' origin can invalidate this presumption. Self-selection into reporting a foreign ancestry and unobserved heterogeneity in the time elapsed since ancestral migration can overestimate the effect of interest. A simple theoretical framework describes the joint influence of these two factors on the estimates obtained from a canonical specification. We provide illustrative examples of the empirical relevance of our concerns drawing on two influential papers in the literature: Fernández and Fogli (2006) and Giuliano and Nunn (2021).