Improving Randomization Tests under Interference Based on Power Analysis

Mizuho Yanagi, Tomonari Sei

In causal inference, we can consider a situation in which treatment on one unit affects others, i.e., interference exists. In the presence of interference, we cannot perform a clas…