Effects of Correlated Covariates on the Efficiency of Propensity Score--Based Estimators Using Estimated Propensity Score

• Activity Number: 552
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Social Statistics Section
• Author(s): Ronnie Pingel*+ and Ingeborg Waernbaum
• Companies: Uppsala University and Umeå University
• Keywords: Efficiency bound ; propensity score ; variable selection ; observational study

Abstract

Propensity score based estimators are commonly used in evaluation research to estimate causal effects. To sufficiently reduce bias in observational studies researchers might be tempted to include many, perhaps correlated, covariates when estimating the propensity score model. Taking into account that the propensity score is estimated, this study investigates how the efficiency of matching, inverse probability weighting and doubly robust estimators change when the covariates are correlated. Analytical propositions regarding the large sample variances under certain assumptions on the DGP are given. This is supplemented by several numerical large sample and finite sample results from a wide range of models. The results show that the correlation may increase or decrease the variances of the estimators. Moreover, several aspects influence how correlation affects the estimators both with respect to magnitude and direction, such as the choice of estimator, the strength of the confounding towards the outcome and the treatment, and whether constant or non-constant treatment effect is present. On the other hand, the distribution of the covariates per se seems to be of no importance.