Assessing effects of confounding variables.