On the bias of complete- and available-case meta-regressions with missing covariates

Abstract: