Please rotate your device to landscape mode to view the charts.

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

Journal: Econometrica

Date: 20190701

Author: Gentzkow, Matthew; Shapiro, Jesse M.; Taddy, Matt

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.

Link: Google Scholar

Key Findings

Unprecedented Rise in Partisan Language

Congressional speech partisanship remained relatively constant from 1873 to early 1990s, then exploded to unprecedented levels after 1994, with the probability of correctly guessing a speaker's party from one minute of speech rising from 54% to 73%.

Innovation in Political Messaging

The sharp increase coincided with the 1994 Republican takeover of Congress and innovations in political messaging, including new focus group techniques and media strategies.

Topic-Specific Partisanship

The rise in partisanship is driven more by how parties discuss topics rather than which topics they choose to discuss, with particularly large increases in partisan language around immigration, crime, and religion.

Partisan Speech Over Time

  • Probability of correctly guessing party from speech remained around 54-57% from 1873-1990
  • Sharp increase begins in 1994, reaching 73% by 2008
  • Unprecedented levels of partisan language in modern era

Most Partisan Topics (2007-2008)

  • Immigration, crime and religion show highest partisan language divide
  • Economic topics like taxes and labor show moderate partisanship
  • Traditional topics like alcohol show decreased partisanship over time

Party-Specific Phrases (2015-2016)

  • Republicans emphasize phrases like "radical islam" and "taxpayer dollars"
  • Democrats focus on phrases like "climate change" and "gun violence"
  • Shows clear linguistic differentiation between parties

Contribution and Implications

  • Developed novel statistical methods to measure partisan language while controlling for bias in high-dimensional data
  • Demonstrated that partisan language divide is a recent phenomenon, challenging previous assumptions about historical polarization
  • Highlighted the role of strategic political communication and media changes in shaping congressional discourse

Data Sources

  • Partisanship over time chart based on data from main results in Figure 2 and associated text
  • Topic partisanship visualization derived from Figure 6 topic-specific analysis
  • Party-specific phrases data from Table I, Session 114 (2015-2016) frequency counts