In randomized designs, how does randomization affect confounding?

• A. It ensures confounding variables are completely eliminated
• B. It minimizes the influence of known and unknown confounders by balancing groups
• C. It increases confounding by selection bias
• D. It has no effect on confounding

Answer: (B)

Randomization is a design feature that helps prevent confounding by making the groups comparable with respect to both measured and unmeasured characteristics. By assigning participants to treatment or control purely by chance, any variables that influence the outcome and are related to the exposure tend to be distributed evenly between groups, so differences in outcomes are more likely due to the treatment itself. This minimizes the influence of known and unknown confounders, though perfect balance isn’t guaranteed—especially in smaller studies or with noncompliance—so some residual confounding can remain. It does not increase confounding, nor does it have no effect; instead, it reduces confounding by balancing groups.