Imposing nonlinear constraints when estimating genetic and cultural transmission under assortative mating: a simulation study using Mx and BUGS

Modeling both genetic and cultural transmission in parent-offspring data in the presence of phenotypic assortment requires the imposition of nonlinear constraints. This article reports a simulation study that determined how well the structural equation modeling software package Mx and the Bayesian-oriented BUGS software package can handle such nonlinear constraints under various conditions. Results generally showed good and comparable results for Mx and BUGS, although BUGS was much slower than Mx. However, since BUGS uses Markov-chain Monte Carlo estimation it could be used for parent-offspring models with non-normal data and/or item-response theory models