Journal
SCANDINAVIAN JOURNAL OF STATISTICS
Volume 48, Issue 3, Pages 790-816Publisher
WILEY
DOI: 10.1111/sjos.12457
Keywords
average treatment effect; Bregman divergence; causal inference; covariate balance
Categories
Funding
- Data Science to Patient Value (D2V) initiative from the University of Colorado
- NSF [DMS-1420451]
Ask authors/readers for more resources
In observational research, estimating marginal causal effects in the presence of confounding variables can be achieved by using propensity score weights to balance the data, making it appear randomized. While propensity score weights have desirable asymptotic properties, they may not achieve sufficient balance in finite samples. Empirical covariate balancing methods offer an appealing alternative by exactly balancing the sample moments of the covariate distribution.
A common goal in observational research is to estimate marginal causal effects in the presence of confounding variables. One solution to this problem is to use the covariate distribution to weight the outcomes such that the data appear randomized. The propensity score is a natural quantity that arises in this setting. Propensity score weights have desirable asymptotic properties, but they often fail to adequately balance covariate data in finite samples. Empirical covariate balancing methods pose as an appealing alternative by exactly balancing the sample moments of the covariate distribution. With this objective in mind, we propose a framework for estimating balancing weights by solving a constrained convex program, where the criterion function to be optimized is a Bregman distance. We then show that the different distances in this class render identical weights to those of other covariate balancing methods. A series of numerical studies are presented to demonstrate these similarities.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available