Comparison of Inverse Probability of Treatment Weighting and Optimal Full Matching Methods to Estimate the Average Treatment Effect: A Monte Carlo Simulation Study

• In Event: Alternative Methods for Data Analysis

Abstract

We investigated the performance of two different methods of reduce selection bias in estimates of the Average Treatment Effect (ATE) in observational studies: inverse probability of treatment weighting (IPTW) and optimal full matching. We conducted a Monte Carlo Simulation study manipulating number of subjects, the ratio of untreated and treated subjects, and the propensity score distributions of treated and untreated subjects. Results indicated that each algorithm removed almost all of the bias under all investigated conditions. As the sample size increases, each algorithm removed more bias but the changes in the percent relative bias were minimal.

Authors

• Sungur Gurel, University of Florida

• Walter L. Leite, University of Florida