Evaluating recent methods to overcome spatial confounding

Abstract The concept of spatial confounding is closely connected to regression, although no general definition has been established. A generally accepted idea in regression models the change fixed effects estimates that may occur when spatially correlated random collinear with covariate are included model. Different methods have proposed alleviate linear models, but it not clear if they provide correct estimates. In this article, we consider some those proposals such as restricted spatial+ model, and transformed Gaussian Markov fields. objective determine which one provides best effects. Dowry death data Uttar Pradesh 2001, stomach cancer incidence Slovenia period 1995–2001 lip Scotland between years 1975–1980 analyzed. Several simulation studies conducted evaluate performance different scenarios confounding. Results reflect method seems closest true value standard errors could be inflated.