Model-free causal inference of binary experimental data

Peng Ding, Luke W. Miratrix

For binary experimental data, we discuss randomization-based inferential procedures that do not need to invoke any modeling assumptions. We also introduce methods for likelihood an…