On the Assumption of Bivariate Normality in Selection Models: A Copula Approach Applied to Estimating HIV Prevalence

Heckman-type selection models have been used to adjust HIV prevalence estimates for selection bias, which arises when participation in HIV testing and HIV status are correlated after controlling for observed variables. These models typically rely on the assumption that the error terms in the participation and outcome equations are distributed as bivariate normal. We introduce a novel approach for relaxing this parametric assumption using copulae. Here we describe our simulation study and provide the R code for evaluating the performance of copula based selection models for binary outcomes.