Doubly robust elimination: Best practices for comparative effectiveness research

Project Description: The proposed research has six aims:

Quantify the degree to which doubly robust (DR) estimation leads to bias reduction relative to standard methods.

Evaluate estimated standard errors under a broad range of conditions.

Determine the optimal covariate types for inclusion in the component models.

Validate a new approach to variable and model-form selection.

Develop and test strategies specific to DR estimation for the identification of non-uniform treatment effects.

Identify conditions under which DR methods perform worse than standard methods.