Disentangling spatial interference and spatial confounding biases in causal inference

arXiv preprint arXiv:2602.02777 (2026)

Preprint
Authors

Ifedayo Ogunsola

Olatunji Johnson

Published

February 1, 2026

Links: arXiv

Summary. Develops methods to separately identify and correct for spatial interference (treatment effects spilling over to neighbours) and spatial confounding (unmeasured spatially structured covariates) in causal inference with geo-referenced data, settings where these two biases commonly co-occur.