It's all relative: Spatial positioning of parties and ideological shifts

The usefulness of the general class of spatial econometric models, which relaxes the assumption that the observations are independent, has only recently been realised. One particularly fruitful application includes models of parties' ideological change as well as the electoral consequences of party competition. In these studies, scholars can explicitly model the spatial interconnectedness of political parties in theoretically pleasing ways, producing inferences that are consistent with formal models of party competition, but are beyond the grasp of traditional ordinary least squares (OLS) regression models. To illustrate these benefits, this article replicates Adams and Somer‐Topcu's 2009 study of parties' responses to ideological shifts by rival parties to show that appropriately modeling patterns of interconnectivity between parties via weights matrices provides more realistic inferences that are more consistent with formal models of party competition.