Winsorizing and Trimming with Subgroups

Winsorizing and trimming are used to minimize the effects of outliers on estimated treatment effects. The typical approach winsorizes/trims the tails of the whole sample, even if there are heterogeneous subgroups within the sample-– like a treatment and control group in Randomized Controlled Trials. An alternative approach – Stratified Winsorizing/Trimming – winsorizes subgroups separately, ensuring that an equal proportion of observations are winsorized/trimmed per subgroup. Monte Carlo simulations of an RCT illustrate that Stratified Winsorizing/Trimming reduces the treatment effect bias and risk of Type II errors compared to the traditional approach, although at the cost of a greater likelihood of Type I errors. Applications to Angelucci et al. (2023) and Jack et al. (2023) illustrate that the chosen winsorizing/trimming technique can affect the magnitude and statistical significance of treatment effects. Practical guidelines for researchers wanting to winsorize/trim a sample that consists of heterogeneous subgroups are discussed.