Define confounding and describe two strategies to control confounding in evaluation studies.

• A. A third variable related to both exposure and outcome that distorts effect; control via randomization and stratification or multivariable adjustment.
• B. A random variation in measurement; control via increasing sample size.
• C. An error in data entry; control via double data entry.
• D. An interaction term in a model.

Answer: (A)

Confounding occurs when a third variable is related to both the exposure and the outcome, distorting the observed exposure–outcome relationship and potentially creating a false or obscured association. Two solid ways to control confounding in evaluation studies are randomization and stratification or multivariable adjustment. Randomization helps by assigning participants to exposure groups by chance, which tends to balance both known and unknown confounders across groups, so differences in outcomes are more likely due to the exposure itself. Stratification involves analyzing the data within subgroups defined by the confounder (or using multivariable adjustment in a regression model) to account for the confounder’s influence and estimate the exposure effect after holding the confounder constant. The other options describe measurement error or data-entry issues, or refer to effect modification, not confounding.