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

Abstract

We 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.

Motivated Reasoning in the Field: Polarization of Prose, Precedent, and Policy in U.S. Circuit Courts, 1891-2013

“…This is evidence of relatively low partisanship in court opinion text. This is different from the main result in GST [22], where the language patterns of congressmen exhibited high partisanship. Now we examine how partisanship in phrases varies across topics.…”

“…Following Gentzkow, Shapiro, and Taddy (2018) [22] (GST), we adopt an intuitive approach to demonstrate how political considerations may lead to partisanship of language and citations in the written opinions of U.S. Circuit Courts.…”

“…We assume the following model, based on GST (2018) [22]. Let c it be the vector of phrase counts for judge i in year t, with each member c ijt representing the frequency of a certain phrase j.…”

“…The partisanship measure is based on Gentzkow, Shapiro, and Taddy (2018, [23], henceforth GST). It is defined as the accuracy with which we can predict judge political affiliation based on the content of their decisions.…”

“…To implement these methods, we use the R packages distrom and textir, developed by Taddy (2015)[37]. In the latest version of Gentzkow, Shapiro and Taddy (2018)[22], the authors add a constant penalty on intercepts and other covariates. This feature is not currently available in the current versions of distrom and textir.…”