Survey designs for producing population prevalence estimates, such as abortion rates, need to consider multiple sources of error. Abortion prevalence has been found to suffer from underreporting. Abortion rates can also be underestimated due to unit nonresponse. However, it is when these two...