The Legislative Influence Detector: Finding Text Reuse in State Legislation
Matthew Burgess, University of Michigan; Eugenia Giraudy, YouGov; Julian Katz-Samuels, University of Michigan; Joe Walsh*, University of Chicago; Derek Willis, ProPublica; Lauren Haynes, University of Chicago; Rayid Ghani, University of Chicago
State legislatures introduce at least 45,000 bills each year. However, we lack a clear understanding of who is actually writing those bills. As legislators often lack the time and staﬀ to draft each bill, they frequently copy text written by other states or interest groups.
However, existing approaches to detect text reuse are slow, biased, and incomplete. Journalists or researchers who want to know where a particular bill originated must perform a largely manual search. Watchdog organizations even hire armies of volunteers to monitor legislation for matches. Given the time-consuming nature of the analysis, journalists and researchers tend to limit their analysis to a subset of topics (e.g. abortion or gun control) or a few interest groups.
This paper presents the Legislative Inﬂuence Detector (LID). LID uses the Smith-Waterman local alignment algorithm to detect sequences of text that occur in model legislation and state bills. As it is computationally too expensive to run this algorithm on a large corpus of data, we use a search engine built using Elasticsearch to limit the number of comparisons. We show how LID has found 45,405 instances of bill-to-bill text reuse and 14,137 instances of model-legislation-to-bill text reuse. LID reduces the time it takes to manually ﬁnd text reuse from days to seconds.
Filed under: Classification