Causal Inference with Interference and Transmission

We review previous results for a two-stage randomization procedure for unbiased estimates of direct, indirect, total and overall effects of an intervention program, such as vaccination. We extend these results to the situation in which one or both of the stages is not randomized, which is quite often the case in vaccine studies. We present a couple concrete examples of such vaccine studies. We then extend our previous research on causal inference for binary post-infection outcomes using principal stratification to define causal estimates for the effects of vaccination on person-to-person transmission. This research is joint with Michael Hudgens.