Figure 1.
Directed acyclic graph for a general hierarchical causal model (equation (1)) with U as unmeasured factors, E as cluster-level covariates, (W1., … , WN.) as individual-level covariates, A as the cluster-level exposure, and (Y1., … , YN.) as individual-level outcomes. Identifiability (Section 2.3) will require additional assumptions on the unmeasured factors U (Supplementary Figure S1).