Causal Mediation Analysis with Spatial Interference

In the causal inference framework, estimating causal effects requires specific assumptions, among which is the assumption of independence of each unit’s outcome from the treatment assigned to other units. Although it can be reasonable in some settings, this is not the case when dealing with spatial data. In this paper, we address causal mediation analysis in the presence of spatial interference: we discuss assumptions for the estimation of direct and indirect effects and provide an applied example.