Which method is commonly used to handle data missing at random (MAR)?

• A. Complete-case analysis
• B. Multiple imputation
• C. Listwise deletion
• D. Mean imputation

Answer: (B)

Handling data missing at random means the reason a value is missing can be explained by information we have observed, so we can use that information to estimate plausible values. The method that’s commonly used in this scenario is multiple imputation. It works by generating several complete datasets where the missing values are replaced with values drawn from the predictive distribution of the missing data given the observed data. Each of these datasets is analyzed separately, and the results are then combined to produce overall estimates that reflect the uncertainty about what the true values might be. This approach helps preserve relationships among variables and provides valid standard errors and confidence intervals under MAR.

Other approaches fall short under MAR. Complete-case analysis, which only uses cases with no missing values, often leads to biased results if the missingness is related to observed data. Mean imputation fills in missing values with a variable’s mean, but it ignores relationships with other variables and underestimates variability, giving overconfident and biased conclusions.