The Performance of Largest Caliper Matching: A Monte Carlo Simulation Approach
The paper presents an investigation of estimating treatment effect using different matching methods. The study proposed a new method which is computationally efficient and convenient in implication-'largest caliper matching' and compared the performance with other five popular matching methods by simulation. The bias, empirical standard deviation and the mean square error of the estimates in the simulation are checked under different treatment prevalence and different distributions of covariates. A Monte Carlo simulation study and a real data example are employed to measure the performance of these methods. It is shown that matched samples improve estimation of the population treatment effect in a wide range of settings. It reduces the bias if the data contains the selection on observables and treatment imbalances. Also, findings about the relative performance of the different matching methods are provided to help practitioners determine which method should be used under certain situations.
READ FULL TEXT