What is propensity score matching and when would you use it?

• A. Test to compare means across groups without adjusting for covariates.
• B. A method to balance observed covariates between treatment and control groups to approximate randomization; used when random assignment is not possible.
• C. A method to randomly assign participants to groups after measuring outcomes.
• D. A way to impute missing data in longitudinal studies.

Answer: (B)

Propensity score matching focuses on creating comparable groups in observational data by balancing the likelihood of receiving the treatment across observed covariates. You estimate each unit’s probability of getting the treatment given its observed characteristics, then pair or group treated and untreated units with similar scores. This lining up of propensity scores makes the treated and control groups more alike with respect to those covariates, so differences in outcomes are more plausibly due to the treatment itself rather than to confounding factors. You’d use this approach when random assignment isn’t possible, such as in many real-world studies where you can’t or shouldn’t assign treatment arbitrarily. It’s important to remember that it only accounts for observed covariates; hidden or unmeasured confounders can still bias results. After matching, you compare outcomes within the matched sets and check covariate balance to ensure the groups are well-aligned.