Iyengar, S., Lelkes, Y., Levendusky, M., Malhotra, N., & Westwood, S. J. (2019). The origins and consequences of affective polarization in the United States. Annual Review of Political Science, 22, 129–146. https://doi.org/10.1146/annurev-polisci-051117-073034
Summary
This review article synthesizes an interdisciplinary literature on affective polarization in the United States — the growing tendency of Democrats and Republicans to dislike, distrust, and avoid members of the opposing party. Drawing on social identity theory, the authors argue that partisanship functions as a powerful social identity that sorts the world into a liked in-group and a disliked out-group. Crucially, they contend that affective polarization is theoretically and empirically distinct from ideological (issue-based) polarization: extreme policy positions are not a necessary condition for partisan animosity. The paper traces measurement approaches, causes, and consequences that spill well beyond politics into economic and social life, before evaluating mitigation strategies and laying out a future research agenda.
Key Contributions
- Provides an integrative framework grounding affective polarization in social identity theory and separating it from ideological polarization (reframing the Fiorina–Abramowitz debate).
- Systematically catalogs three measurement traditions — survey self-reports, implicit measures, and behavioral measures — with their respective strengths and limitations.
- Assembles evidence on the nonpolitical spillover of partisan animus into social relationships, marriage, hiring, and economic behavior.
- Clarifies which candidate causes (echo chambers, residential sorting) remain empirically contested.
- Sets a detailed research agenda, including in-group-love versus out-group-hate, the causal role of sorting, and comparative extensions.
Methods
A narrative synthesis of political science and social psychology research. The authors review three measurement approaches: (1) survey self-reports (ANES feeling thermometers, trait ratings, social distance/marriage items), (2) implicit measures (partisan Implicit Association Test), and (3) behavioral measures (trust and dictator economic games, scholarship-selection and hiring experiments). Evidence is drawn from longitudinal survey time-series (ANES ~1978–2016, Pew), audit and field experiments, online dating data, voter files, and consumer behavioral data.
Findings
- ANES feeling thermometer polarization rose from 22.64° in 1978 to 40.87° in 2016, driven mainly by declining out-party warmth (with a slight dip 2012–2016).
- Opposition to a child marrying across party lines rose from ~4–5% in 1960 to roughly a third of Democrats and half of Republicans by 2010.
- Implicit partisan bias is widespread (~70% show in-party favoritism) and larger than implicit racial bias (D-score 0.50 vs. 0.18).
- In a scholarship experiment, ~80% of partisans chose the co-partisan even when the out-party applicant was more qualified.
- Partisan matching increased dating message exchange by 9.5%; ~80.5% of married couples share a party identification.
- Sorting explains only ~5% of the variance in affective polarization, and animus rose comparably among ideologically consistent and inconsistent partisans.
- Correcting misperceptions about party composition (people overestimate stereotypical group prevalence by ~342%) and priming a common American identity both reduce out-party animus.
Connections
This paper is a foundational statement on affective polarization and the identity-based (rather than issue-based) account of partisan division; its skepticism about echo chambers as a driver connects directly to work on selective exposure and cross-cutting information such as Bail2018-fk, Bakshy2015-rn, Barbera2015-fw, Guess2023-ai, and Nyhan2023-gb. Its treatment of partisan media and elite messaging as intensifiers of animus links to Benkler2018-lw and Osmundsen2021-et, while its call to extend the framework comparatively speaks to cross-national work like Humprecht2020-gd.