Statistics for Robust Methods and Model Selection for Causal Inference under Missingness of the Study Exposure