A Bayesian Parametric Approach to Measuring Legislative Productivity

Abstract

We develop and present a new measure of legislative effectiveness that acknowledges the challenge of assigning scores to individual members who are engaged in a collective legislative process. To produce our effectiveness scores, we opt for a value-added measurement model that is fit on bill-level data with varying intercepts by sponsor. We include majority party membership, committee chair status, and other institutional variables in our models to ensure that our estimates of effectiveness are purged of the effects of organizational status, instead of being driven by them. Our models produce member-level measurements of the predicted marginal change in the probability of bill advancement attributable to the bill's sponsor. We compare our scores to existing legislative effectiveness scores, and include a discussion of the theoretical and methodological differences between our approach and existing approaches.