Leaders or Laggards? A New Dataset on State and Federal Policy Adoptions

• In Event: Laboratories of Democracy: Advances in Policy Diffusion Research

Abstract

This paper uses text from the legislative, judicial, and executive branch categorize hundreds of thousands of state bills, state supreme court decisions, and state bureaucracy rules by topic area from 2000-2010. We use a Recurrent Neural Network (RNN) to categorize over 400,000 adoptions using the major topic codings from the Comparative Agendas Project using policy text and a training set of 10,000 categorized state policy adoptions. These data and method provide us with unprecedented detail into the policy decisions taken by all three branches of government in all 50 states from 2000-2010. We then compare these data to federal policy adoption data over the same period to evaluate the extent to which states set the policy agenda or follow the federal government’s lead by examining the distribution of topic areas over time in both the states and Congress. This new dataset of categorized state policies will be a resource for state politics scholars to evaluate what sets and drives agendas across branches of government.

Authors

• Scott James LaCombe, Smith College

• Frederick J. Boehmke, University of Iowa

• Jeffrey J. Harden, University of Notre Dame

• Bruce Desmarais, Pennsylvania State University

• Jason H. Windett, University of North Carolina, Charlotte

• Robert J. McGrath, George Mason University

• William W. Franko, West Virginia University

• Yu-Ru Lin, University of Pittsburgh