Which covariates should be controlled in propensity score matching? Evidence from a simulation study
Version of Record online: 27 NOV 2012
© 2012 The Authors. Statistica Neerlandica © 2012 VVS
Volume 67, Issue 2, pages 169–180, May 2013
How to Cite
Cuong, N. V. (2013), Which covariates should be controlled in propensity score matching? Evidence from a simulation study. Statistica Neerlandica, 67: 169–180. doi: 10.1111/stan.12000
- Issue online: 25 APR 2013
- Version of Record online: 27 NOV 2012
- Received: 30 Janauary 2012. Revised: 17 October 2012.
- impact evaluation;
- treatment effect;
- propensity score matching;
- covariate selection;
- Monte Carlo
Propensity score matching is a widely-used method to measure the effect of a treatment in social as well as medicine sciences. An important issue in propensity score matching is how to select conditioning variables in estimation of the propensity scores. It is commonly mentioned that variables which affect both program participation and outcomes are selected. Using Monte Carlo simulation, this paper shows that efficiency in estimation of the Average Treatment Effect on the Treated can be gained if all the available observed variables in the outcome equation are included in the estimation of propensity scores. This result still holds in the presence of non-sampling errors in the observed control variables.