Primary Submission Category: Causal Inference in Networks
Quasi-randomization tests for network interference
Authors: Supriya Tiwari, Pallavi Basu,
Presenting Author: Supriya Tiwari*
Many classical inferential approaches fail to hold when interference exists among the population units. This amounts to the treatment status of one unit affecting the potential outcome of other units in the population. Testing for such spillover effects in this setting makes the null hypothesis non-sharp. There is a growing body of literature that considers this problem in an experimental setup with the network structure amongst the population to be fixed and assumed to be given. An interesting approach to tackling the non-sharp nature of the null hypothesis in this setup is constructing conditional randomization tests such that the null is sharp on the restricted population. In randomized experiments, conditional randomized tests hold finite sample validity. However, such approaches are computationally intensive as finding these appropriate sub-populations can involve solving an NP-hard problem. In this paper, we view the network amongst the population as a random variable instead of fixed and treat the given network as the observed outcome of the network random variable. We propose a new approach that builds a conditional quasi-randomization test. We highlight that the approach is easier to implement than the current state-of-the-art methods. We conduct a simulation study to verify the finite-sample validity of our approach and illustrate our methodology to test for interference in a weather insurance adoption experiment run in rural China.