Modeling Migration in Warblers

Benjamin Gochanour, Andrea Contina

Orange Crowned Warbler Species Distribution Modeling

A Multiply Robust Multiple Imputation Method for Causal Inference

Benjamin Gochanour, Sixia Chen, Laura Beebe, David Haziza

Abstract: Evaluating the impact of non-randomized treatment on cancer-related health outcomes is difficult in observational studies because of confounding variables that may affect both the behavior and the outcome of interest. In the present study, we develop a non-parametric multiply robust multiple imputation method for estimating mean treatment effects in such observational studies, treating the challenge as a missing data problem. Our method relies on multiple propensity score models and outcome regression models and is multiply robust in that it performs well as long as at least one of the models is correctly specified. We develop the asymptotic properties of our method and test it in a simulation study, evaluating its performance in terms of bias, efficiency, and coverage probability. Rubin’s variance estimator formula can be used safely for estimating the variance of our proposed estimators. Finally, we apply our method to an Oklahoma Helpline intervention study to evaluate the effect of various smoking cessation intervention types on smoking status.