The article discusses how the SAS PROC TTEST procedure can be used to test for differences in population means, even when assumptions of normality and variance homogeneity are not met. Previous research has shown that the t-test is robust to deviations from normality, but questions remain about the performance of conditional testing in these cases. The paper extends this research to evaluate conditional testing under non-normal, heteroscedastic distributions and provides guidance on its proper use.