What Contiguity?: Using an ERGM to Understand Perceived State Similarity

• In Event: Leveraging the States: Improving Inferences with US State Data

Abstract

In this paper we explore what factors contribute to citizen perceptions of state similarity. Traditionally, policy diffusion research has relied on contiguity as a proxy for state similarity. However, research in political culture has long argued that similarity is more complex than geographic proximity (Lieske 1993, Chinni and Gimple 2003), and many bordering states have dramatically different characteristics, both demographically and culturally. The similarity scores we analyze are a strong predictor of policy diffusion and policy similarity among the states, and a better predictor of adoption than contiguity (Bricker and LaCombe 2018).
But what factors influence people’s perceptions of states? To investigate this, we use a unique survey of citizens' perceived ties between states combined with state demographic information to estimate an Exponential Random Graph Model (ERGM) to predict perceived similarity ties in the network of states. This approach will help us isolate what factors contribute to citizens perceiving other states as similar, and understand important differences between our proposed measure and the traditional contiguity measures.

Authors

• Christine Bricker, Warren Wilson College

• Scott James LaCombe, University of Iowa