For m phenotypes and k common factors, data were simulated according to the model:(3)where Σ denotes the m×m variance-covariance matrix between the phenotypes, Λ is the m×k matrix of factor loadings (t denotes matrix transpose), Ψ is the k×k variance-covariance matrix between the common factors, and Θ is the m×m diagonal matrix of residual variances (i.e., the part of the phenotypic variance that is not explained by latent factors). In simulations with multiple factors (k>1), we maintained simple structure, i.e., each phenotype is related to only one factor.
