Measuring Group Differences in High-Dimensional Choices: Method and Application to Congressional Speech

Abstract

for their comments and suggestions. We thank Frances Lee for sharing her data on congressional communications staff. We also thank numerous seminar audiences and our many dedicated research assistants for their contributions to this project. This work was completed in part with resources provided by the University of Chicago Research Computing Center and the Stanford Research Computing Center. The data providers and funding agencies bear no responsibility for use of the data or for interpretations or inferences based upon such uses. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.At least one co-author has disclosed a financial relationship of potential relevance for this research. Further information is available online at http://www.nber.org/papers/w22423.ack NBER working papers are circulated for discussion and comment purposes. They have not been peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.© 2016 by Matthew Gentzkow, Jesse M. Shapiro, and Matt Taddy. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given to the source. ABSTRACTWe study the problem of measuring group differences in choices when the dimensionality of the choice set is large. We show that standard approaches suffer from a severe finite-sample bias, and we propose an estimator that applies recent advances in machine learning to address this bias. We apply this method to measure trends in the partisanship of congressional speech from 1873 to 2016, defining partisanship to be the ease with which an observer could infer a congressperson's party from a single utterance. Our estimates imply that partisanship is far greater in recent years than in the past, and that it increased sharply in the early 1990s after remaining low and relatively constant over the preceding century.

Measuring Group Differences in High‐Dimensional Choices: Method and Application to Congressional Speech

“…Instead, our approach is to specify a generative model of the data and to measure group differences using objects that have a well‐defined meaning in the context of the model . In the body of the paper, we note some formal connections to the literature on residential segregation, and in an earlier draft, we pursue a detailed application to trends in residential segregation by political affiliation (Gentzkow, Shapiro, and Taddy ()).…”

Random Projection Estimation of Discrete-Choice Models with Large Choice Sets

“…The rank-order 2 See also Davis et al (2016) and Keane and Wasi (2012) for other applications of sampled logit-type discrete choice models. On a related note, Gentzkow et al (2016) use a Poisson approximation to enable parallel computation of a multinomial logit model of legislators' choices among hundreds of thousands of phrases.…”

The Imperfect Beliefs Voting Model

“…Evidence has increasingly shown that such differences have an impact on the policies implemented (e.g., Levitt (1996)). Indeed, the language of the two parties is increasingly becoming more differentiated (e.g., Gentzkow, Shapiro and Taddy (2016)). Therefore, in a world of increasing social-cultural polarization, it makes sense to begin by allowing for correlation between policy preferences and sociocultural dis-13 tance.…”

“…This provides a new avenue for empirical exploration, as social sorting may be driving recent ideological divergence within the parties. In particular, the stark results of Gentzkow, Shapiro and Taddy (2016), finding that parties increasingly use particularistic, polarized language which is only intended to reach base voters, should indicate a potential social mechanism for political polarization.…”

“…Similarly, parties are simultaneously becoming more polarized both along cultural and socioeconomic grounds, as well as in policy (e.g., Brewer and Stonecash (2007); Mason (2015); Gentzkow, Shapiro and Taddy (2016)). For example, the Pew Research Center has shown that white, working class voters, a historically split demographic, have strongly and rapidly sorted into the Republican Party over the past decade, with nearly 60% now identifying as strong Republicans (Pew (2012)).…”