Abstract:

Step regression and AIC are applied to estimate the propensity score. Based on the weighting propensity score estimator(WPSE) and propensity score matching estimator(PSME), the weighting propensity score matching estimator(WPSME)is proposed. Then the three methods of WPSE、PSME、WPSME are compared by simulating using the bootstrap method. It is shown that WPSME is more precise and more efficient than the other methods in terms of bias and standard deviation.  Finally, the average treatment effect of the job training program is estimated with WPSME.

Key words: weighting propensity score; matching method; average treatmen t effect