Network Cluster-Robust Inference

Network data commonly consists of observations on a single large network. Accordingly, researchers often partition the network into clusters in order to apply cluster-robust inference methods. All existing such methods require be asymptotically independent. We show that for this requirement hold, under certain conditions, it is necessary and sufficient have small conductance, which ratio edge boundary size volume. This yields quantitative measure cluster quality. Unfortunately, there are important classes networks small-conductance appear not exist. Our simulation results networks, can exhibit substantial distortion. Based well-known spectral graph theory, we suggest using eigenvalues Laplacian determine existence number clusters. also discuss use clustering constructing practice.