This document discusses the differences between effect modification, bias, and confounding in epidemiological studies. It provides an example where age modifies the effect of sex on hospitalization after a motor vehicle collision. Effect modification exists when the effect of an exposure on an outcome differs across levels of another variable. This is not a bias, but rather an interaction that should be reported. Bias occurs when systematic errors distort the exposure-outcome association. Confounding happens when a third variable is associated with both the exposure and outcome but does not cause the outcome.