Results of the Two-Part Random Effects Regression Modeling the Sexual Risk Score, Circumcision Status, and Time With and Without Covariates

In the logistic part of the model, exponentiated betas represent population averaged odds ratios.
In the lognormal portion of the model represent, exponentiated betas represent % change in Y (risk score), per change in unit X.
In the adjusted model, the following covariates were adjusted for: age, marital status, education, income, and belief that circumcision reduces the risk of acquiring HIV.