Data were simulated as described in Simulation study 1, with moderation on the cross paths being absent. We now, however, introduced moderation on the paths unique to T, with moderation parameters being either β a = .08, or β c = .10, or β e = .035 (As shown in Table 5, the power to detect moderation on E variance is much greater than the power to detect moderation of A or C variance, which is why β e was chosen much smaller than both β a and β c). Note that the effect size of the chosen moderation parameters depends on the nature of the correlation between T and M. For example, β a was set at .08. If the correlation ran via C or E, then the genetic variance of trait T was calculated as (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \sqrt {.4} $$\end{document} + .08 * M)2. If the correlation between T and M ran via A, however, then the genetic variance of T was calculated as .15 + (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts}
