What to do about polarization? Evaluating the evidence base of recommendations for depolarization

van Eck, C. W., Meyer, H. M. A., & Brüggemann, M. (2026). What to do about polarization? Evaluating the evidence base of recommendations for depolarization. Annals of the International Communication Association. https://doi.org/10.1093/anncom/wlag012

View paper

Summary

This paper is a meta-level systematic review of how social science scholarship translates empirical findings on polarization into recommendations for depolarization. Following PRISMA guidelines, the authors analyzed 89 peer-reviewed articles (2013–2023), paying special attention to the discussion sections where practical implications are typically stated. Their central argument is that while polarization research has accumulated substantial empirical evidence about the drivers of polarization, the recommendations scholars offer for reducing it are largely disconnected from that evidence — resting instead on implicit normative ideals, above all those of deliberative democracy. The paper calls for greater transparency about both the evidence base and the value commitments guiding scholarly recommendations.

Key Contributions

• First systematic mapping distinguishing empirically supported factors, explicit recommendations, and underlying normative ideals in the (de)polarization literature.
• Documents a critical gap between evidence and recommendations, exposing the role of implicit normative commitments (especially deliberative democratic theory).
• Identifies two unresolved structuring debates: knowledge/education vs. culturally resonant communication, and reducing echo chambers vs. cautioning against cross-cutting exposure.
• Proposes a research agenda urging separation of facts from value judgments, interdisciplinary collaboration, a shift toward studying depolarization, and methodological diversification.
• Offers conceptual distinctions such as “mitigative preventive” versus “progressive” approaches to depolarization.

Methods

Systematic literature review following PRISMA 2020 guidelines, preregistered on OSF. A Web of Science search on polarization/depolarization terms (2013–2023, English-language, peer-reviewed) yielded 7,994 results, screened by citation count to retain up to ten most-cited relevant articles per year, narrowing to 89 after full-text review. A mixed deductive-inductive coding scheme captured conceptualizations (ideological/affective; state/process), methods, identified factors, recommendations, and normative ideals. Inter-coder reliability averaged Krippendorff’s Alpha of 0.87, with grounded-theory three-stage coding for inductive categories and descriptive statistics for deductive codes.

Findings

• The field is heavily skewed toward polarization: 82 of 89 articles studied polarization alone, only 1 depolarization alone; 62 addressed ideological polarization, 6 affective, 21 both.
• Quantitative methods dominate (38 experiments, 30 surveys, 40 content analyses); only one study examined direct face-to-face communication.
• Best-supported drivers of polarization: confirmation bias/motivated reasoning (N=21), selective exposure/echo chambers (N=13), one-sided media framing (N=12), and exposure to dissonant views (N=8).
• Depolarization factors appeared in only 26 studies; most supported were knowledge/critical thinking/education (N=5) and culturally/emotionally resonant communication (N=4).
• Evidence supports contradictory claims — both avoiding dissonant exposure and facilitating cross-cutting communication can depolarize, conditional on context.
• Only 1 study grounded its depolarization recommendation in its own evidence; 41 made recommendations unsupported by their analyses.
• Most common recommendations: culturally resonant communication (N=7), knowledge/media literacy (N=5), misinformation interventions (N=4, emerging post-2020); deliberative democracy (often via Habermas) was the dominant normative ideal (N=7).

Connections

This review sits above much of the empirical work on partisan news and cross-cutting exposure, and its debate over echo chambers vs. cross-cutting exposure directly bears on findings like Bakshy2015-rn on ideological exposure via social media and Mosleh2024-op on cross-partisan interaction. Its emphasis on misinformation interventions as an emerging depolarization strategy connects to work such as Allen2025-ot, while its skepticism about whether cross-cutting exposure reliably depolarizes resonates with debates addressed in Tornberg2025-ir.

Podcast

A research-radio episode discusses this paper: 🎧 MP3 · Spotify · Apple Podcasts