Association for Public Policy Analysis & Management

Bills and floor votes are popular metrics in legislative studies. At the same time, every legislative scholar appreciates that each has important limitations. Bills are containers for policy ideas, and a bill’s content frequently changes as it moves through the legislative process. As a result, the bill that becomes law can be completely different from what the bill’s sponsor originally proposed. Floor votes are widely used to study legislator preferences. However, because floor votes involve constrained choices in a highly politicized context, voting positions may not be representative of legislative behavior in other less visible and less constrained legislative contexts.

Computational text reuse methods provide new opportunities to investigate lawmaking. (A familiar application is plagiarism detection software.) We begin with an overview of text reuse methods and the largely untapped opportunities they present for public policy researchers. We then turn to the more specific subject of lawmaking to investigate several related questions: How much do bills change as they move through the process? What kinds of changes are made? Where are these changes made? What inspires them? This ability to investigate the evolution of bills quantitatively leads to new insights. Among other things, we discover a more inclusive legislative process, varying degrees of committee deference, and a clear pattern of Senate dominance as the chambers resolve their differences.