(e.g., additive versus dominant or recessive modes of inheritance), minor allele frequency (MAF), type I error rate, and the hypothesis under consideration (i.e., genetic main effect, gene–environment interaction, or even gene–gene interaction). To illustrate, let us consider how several of these factors might influence the sample size required to replicate the effects of a putative SNP with association with major depressive disorder, a phenotype of interest to many social scientists which affects ~6.7% of adults in the U.S. population in a given year (Kessler et al., 2005). We will begin by setting our minimum acceptable thresholds for power at 0.80 and type I error rate at 0.05. Given the dichotomous nature of the phenotype, we will further assume a 1:1 matched case–control design, and start with a baseline model for a genotypic main effect that reflects an additive mode of inheritance, a MAF of 0.1, and an OR of 1.15. To detect this effect under the circumstances described above would require 4261 participants, half cases and half controls. Holding all else constant, hypothesizing a dominant mode of inheritance would require 5037 participants, whereas reducing the frequency of the risk allele by half (i.e., 0.005) would require 8021 participants. Similarly, to
