Political polarization threatens cooperation, democratic stability, and human flourishing across many societies (Bramson et al., 2017; Finkel et al., 2020; Graham & Svolik, 2020). Yet the field still lacks globally validated measures and shared infrastructure for mapping polarization between and within cultures, especially measures that capture how people actually justify their views and how they evaluate the reasoning of political opponents (Bramson et al., 2017). Most cross-national work relies on imported survey batteries and focuses primarily on ideological distance (Bramson et al., 2017). By contrast, our program uses cross-cultural measures and treats people's explicit reasons for political beliefs as structured psychological data: reasons reflect values, perceived threats, standards of evidence, and moral priorities (Mercier & Sperber, 2011). Polarization is sustained not only by disagreement over outcomes, but also by disagreement over what constitutes a good reason, and by the tendency to discount opponents' reasoning as irrational or illegitimate (Dorst, 2023; Iyengar & Westwood, 2015; Kahan, 2017).

This planning grant will build the partnerships, translation/validation protocols, measurement instruments, and multilingual web infrastructure needed to support a competitive mapping proposal and a durable cross-cultural consortium. We will focus on Latin America (Brazil, Colombia, Argentina), where political systems, party structures, and salient conflicts differ from the U.S. and Europe (Bergman & Fernández, 2025; McCoy, 2024). We have assembled a multidisciplinary team from the U.S., Brazil, and Colombia with cross-cultural research experience to bring local expertise to the project.

A core practical lesson from our ongoing U.S./Brazil work concerns the need for careful cross-cultural measurement development. For example, many studies treat psychological warmth as a default metric of affective polarization (Iyengar et al., 2012), but "warmth/coldness" is not a straightforward or valid concept in Brazilian Portuguese (there is no natural everyday translation that preserves the intended English interpersonal meaning). In our project, we ran validation studies of close translations to identify measures that were comprehensible and psychometrically adequate. Similar issues will arise across Latin America (and across Spanish dialects), and our proposal formalizes a translation-and-validation pipeline designed to prevent precisely these hidden failures of construct validity (Beaton et al., 2000).

Definition and model

Definition. For this project, we treat affective polarization as the primary target phenomenon: the tendency to disrespect, denigrate, and withdraw goodwill from political opponents (Iyengar et al., 2012; Iyengar & Westwood, 2015; Finkel et al., 2020). This is expressed as negative feelings, moralized contempt, diminished trust, preferences for social distance, and exclusionary or punitive attitudes toward political outgroups (Graham & Svolik, 2020). This is distinct from ideological polarization, or distance between policy preferences. Our definition emphasizes that affective polarization reflects an attitudinal shift in how opponents are socially and epistemically regarded: opponents are seen as less reasonable, less sincere, and less entitled to equal standing in shared civic deliberation.

Model. We model affective polarization as partly driven by how citizens represent opponents' minds and reasoning, especially (i) perceived extremity, (ii) perceived motivations and reasons, and (iii) perceived standards of justification. These representations shape emotions (anger, disgust), interpersonal judgments (trust, respect), and downstream civic behaviors (avoidance, exclusion, willingness to cooperate). The mapping phase will test this model cross-culturally using validated measures of affective polarization outcomes and targeted measures of representational accuracy and epistemic evaluation.

Core mechanisms. Affective polarization is sustained in part by metaperceptual ignorance: systematic errors in beliefs about what opponents believe, why they believe it, and what standards they use to judge reasons (Ahler & Sood, 2018; Levendusky & Malhotra, 2016; Moore-Berg et al., 2020). When individuals misperceive opponents' positions and reasons, opponents seem to be irrational, immoral, or acting in bad faith. These distorted models make contempt and social distance feel warranted, even when opponents' actual reasoning is more diverse, constrained, and intelligible than assumed.

This mechanism has three components:

Position misperception: People overestimate how extreme opponents' views are and how far away they are from their own, inflating perceived distance and threat.

Reason distortion: People misrepresent opponents' justificatory repertoires‚ imagining motives and values that are more simplistic, selfish, or morally objectionable than opponents' own stated reasons. This "reason gap" is not just ignorance; it is often a structured caricature of the opponent's explanatory world.

Credibility filtering: People infer that opponents do not care about evidence, are easily manipulated, or reason in a corrupted way (Iyengar & Westwood, 2015; Kahan, 2017; Dorst, 2023). This shifts disagreement from "they're wrong" to "they're unreasonable."

Psychological pathway to affective polarization. Metaperceptual ignorance undermines cognitive empathy. When opponents are seen as extreme, morally suspect, and epistemically illegitimate, people become more prone to moral condemnation, disrespect, emotional hostility, and exclusionary attitudes. Inaccurate contrapartisan models make denigration feel rational, converting political disagreement into a justification for social and epistemic exclusion. This pathway offers a concrete bridge to depolarization research: if we can improve accuracy about opponents' reasons and evaluative standards without demanding agreement, we may reduce contempt and restore some willingness to engage.

Why this matters. This work addresses a central bottleneck in polarization research: cross-cultural datasets often rely on measures that have not been validated for language and cultural meaning, and they rarely capture the justificatory structure of political disagreement (Beaton et al., 2000). By focusing on reasons and credibility judgments, we can identify when hostility is driven by misunderstanding and dismissal rather than ideological disagreement. This also yields an actionable depolarization lever. Reducing ignorance about opponents' reasons and standards should reduce contempt even when disagreement remains. Cross-cultural mapping is essential because the sources and consequences of ignorance likely differ across political systems, media environments, histories of conflict, and languages.

Why Latin America. We target Brazil, Colombia, and Argentina because they provide theoretically meaningful variation while remaining feasible for coordinated comparative work. Across these countries we observe differences in party system structure (two-party vs multi-party dynamics, coalition politics), institutional trust (Carlin & Love, 2018), salient conflict domains (security and crime, inequality, corruption, religion and rights), and recent democratic stressors (McCoy, 2024; Bergman & Fernández, 2025). These conditions are likely to shape both affective polarization and the specific form that reasoning gaps take. At the same time, the region offers a tractable family of comparisons (Portuguese and Spanish contexts with related but non-identical political vocabularies), allowing us to test translation/measurement equivalence under realistic conditions. Practically, we also have established experience and infrastructure for high-quality recruitment in these sites through a regional data collection partner (Netquest), enabling efficient pilot work during the planning period and scalable implementation in the mapping phase.

Measurement battery

The mapping phase will measure both (i) the extent of affective polarization in each country and (ii) the mechanistic drivers emphasized in our model. The planning period will translate, culturally adapt, and validate these measures, explicitly anticipating "warmth"-style construct failures and documenting adaptation decisions.

Affective polarization (primary outcomes). We will use convergent indicators of affective polarization, including: outparty evaluation (validated alternatives to warmth/coldness where needed), anger, disgust, distrust, social distance, and denigrating trait attributions (e.g., irrationality, immorality). We will also include behavioral intentions relevant to civic life (e.g., willingness to engage, cooperate, or support exclusionary actions).

Understanding opposing reasons (two-step process). We quantify understanding using paired comparisons that separate what contrapartisans actually think from what people believe contrapartisans think. First, we collect first-person reasons for political stances and have co- and contrapartisans rate the quality of the same reasons to estimate cross-partisan differences in what counts as a good reason. Second, we elicit metaperceptions (e.g., "How good would the typical [contrapartisan] find this reason?") and compare them to opponents' actual ratings to estimate ignorance. We also measure epistemic standards directly using the partisan-penalty design ((Iyengar & Westwood, 2015; Nguyen et al., 2026; see below), estimating how source cues shift perceived reason strength and legitimacy.

In the mapping phase, these validated measures will enable preregistered tests of scalable interventions (e.g., structured exposure to opponents' best reasons) designed to improve representational accuracy and reduce denigration.

Research activities

Planning-period aims. During the planning period, we will: (Aim 1) translate, culturally adapt, and psychometrically validate the core measurement battery in Portuguese and regional Spanish; and (Aim 2) build and test multilingual web infrastructure (version-controlled translations, modular study deployment, and public-facing dissemination pages) to support implementation at scale in the mapping phase. The planning period will produce a submission-ready mapping blueprint (sites, governance, validated measures, and multilingual deployment), supported by targeted pilot studies in each country. Activities are organized as follows:

A. Reason collection
We will field open-ended prompts on controversial country-specific issues to collect reason statements. We will develop and document translation/cultural adaptation rules for prompts to ensure neutrality and comparable elicitation.

B. Reason ratings by co- and contrapartisans.
Using a curated subset of reasons, we will collect reason-quality evaluations from co- and contrapartisans to build cross-partisan reason-quality profiles and quantify evaluation differences.

C. Metaperceptual judgment studies
Participants will estimate how typical opponents would evaluate the same reasons. Comparing these metaperceptions to opponents' actual evaluations yields a reason-gap miscalibration index.

D. Partisan penalty studies (including label manipulations)
To quantify source-based credibility filtering and its contribution to affective polarization, we will replicate the core experimental procedure from our prior work in each country, adapting labels to locally meaningful political identities (e.g., government/opposition blocs or salient partisan groupings when party labels are unstable).

Participants will evaluate a series of brief, issue-specific reason statements for agreeing or disagreeing with a focal position. On each trial, participants will see a prompt of the form: "Please consider the following reason for [agreeing/disagreeing] with the statement: [Position]." They will then rate the reason on a common battery (e.g., strength/quality, legitimacy, sincerity/good faith, and perceived reasonableness). This design yields a comparable "reason evaluation" profile across countries and issues.

We will implement three randomized attribution conditions:
No-label baseline: reasons are presented without any source information, providing a baseline estimate of perceived reason quality absent identity cues.

Correct-label condition: reasons are presented with a true source label indicating the political group of the person who (purportedly) provided the reason (e.g., "provided by a [ingroup/outgroup] supporter"). This tests whether the same content is discounted when attributed to an outgroup source.

Flipped-label condition: reasons are presented with a mixture of true and switched labels (i.e., some reasons are attributed to the opposite group), allowing us to isolate the causal effect of the label while holding content constant and to test whether credibility filtering is symmetric across groups.

In addition, for ideologically aligned participants, we will include an anchoring item that asks respondents to compare each displayed reason to their own main reason for holding that stance (-2 = Much worse to 2 = Much better"). This provides a within-participant benchmark for reason quality and helps distinguish generalized hostility from more specific epistemic discounting.

Primary outputs from this activity include: (i) the labeling effect on perceived reason quality/legitimacy (the partisan penalty), (ii) whether penalties are stronger for particular issue domains or reason types, and (iii) whether individual differences in penalty magnitude predict affective polarization outcomes (e.g., outgroup denigration, distrust, social distance). We will also use the no-label baseline to quantify the extent to which cross-group differences reflect content-based disagreement versus source-based discounting.

Translation/validation workflow. For each construct we will implement: forward translation, reconciliation by bilingual experts, pretesting/cognitive checks, psychometric validation, and documentation of adaptation decisions. The U.S./Brazil "warmth/coldness" issue is our canonical example of why this is non-negotiable: without validation, cross-national comparisons can be driven by linguistic artifacts rather than real psychological differences.

Collaboration and capacity building

The team includes: Dr. Samuel Murray (PI; Providence College), Dr. Walter Sinnott-Armstrong (Co-PI; Duke University), Dr. Paolo Boggio (Co-PI; Presbyterian Mackenzie University), Dr. Gino Marttelo Carmona Díaz (Personnel; Universidad de los Andes, Bogotá), and Beatriz B. de Souza (Personnel; Mackenzie Presbyterian University).

Collaboration and capacity building are integrated across all phases of the planning grant. Work will be coordinated through rotating leads. One team member will serve as point person for each core area ( Measures & Theory, Translation & Adaptation, Issues & Stakeholders, and Data Governance) so decisions are coordinated and documented. We will hold one biweekly all-hands meeting to coordinate progress and resolve cross-cutting decisions, and we will maintain a shared implementation log (decisions, versions, and action items) to ensure transparency and continuity across countries.

In each country, we will convene meetings with local practitioners and educators. These sessions will serve three functions: first, to vet the salience and framing of target issues so that the study captures the real contours of political disagreement; second, to identify wording sensitivities and culturally specific connotations that may undermine neutrality or measurement equivalence; and third, to shape dissemination products so that outputs are interpretable and useful beyond academia, including for education, civic engagement, and public communication.

Capacity building is a core aim of the group and will be treated as an explicit deliverable. We will develop and share training materials covering translation and cultural adaptation protocols, coding and annotation procedures for open-ended reasons, reproducible analysis workflows, and web-based study deployment. To support equitable collaboration over the long term, we will establish shared authorship norms and transparent contribution tracking. This ensures that local researchers are not only implementers but also co-designers of the measures, protocols, and research agenda that will define the subsequent mapping phase.

Capacity for success

Our team has an established track record managing international, interdisciplinary research, including ongoing cross-cultural polarization work, and we already have reliable recruitment pipelines in the target countries (Netquest; BeSample). We will use Netquest for Colombia and Argentina because it provides strong, reliable panel coverage and proven fielding capacity in those Spanish-language contexts, while BeSample will be used for Brazil to maintain continuity with our existing Brazil recruitment pipeline and to support efficient implementation of the partisan-penalty design in Portuguese. Brazil partnership capacity is confirmed through Paulo Boggio (Presbyterian Mackenzie University). In Colombia, we have established ties to the School of Social Sciences at Universidad de los Andes (Bogotá) through Dr. Murray's prior postdoctoral appointment and Gino Carmona's alumni network, which will support rapid identification of a local academic partner and stakeholder engagement. Argentina partner identification is underway and will be finalized via the planning workshop and targeted outreach during the grant period.

Open science and dissemination

We will preregister the mapping-phase design template and analysis plan skeleton, maintain a transparent translation/adaptation log, and create reproducible pipelines for data processing. Dissemination will be built into the multilingual website: public-facing summaries of findings, educational explainers on reasoning gaps and credibility filtering, and documentation that supports reuse by researchers and practitioners (subject to privacy-sensitive data governance). We will ensure publications are made openly available and will share de-identified data, code, and materials when ethically permissible, consistent with TWCF's open research/open access policies.

Planning-phase deliverables

Partner site network and governance plan (including data governance and authorship norms)
Validated multilingual measures for affective polarization and mechanism indices
Core issue module + country-specific modules (pilot-tested prompts)
Multilingual web deployment capability (beta) with translation version control
Full mapping proposal blueprint: sampling plan, preregistration template, analysis plan skeleton, and an implementation-ready workflow for January 2027 submission.

References

Ahler, D. J., & Sood, G. (2018). The parties in our heads: Misperceptions about party composition and their consequences. The Journal of Politics, 80(3), 964-981. https://doi.org/10.1086/697253

Bergman, M., & Fernández, P. (2025). Affective polarization in Latin America: A research note. Latin American Politics and Society, 67(1), 114-132. https://doi.org/10.1017/lap.2024.51

Beaton, D. E., Bombardier, C., Guillemin, F., & Ferraz, M. B. (2000). Guidelines for the process of cross-cultural adaptation of self-report measures. Spine, 25(24), 3186-3191. https://doi.org/10.1097/00007632-200012150-00014

Bramson, A., Grim, P., Singer, D. J., Berger, W. J., Sack, G., Fisher, S., Flocken, C., & Holman, B. (2017). Understanding polarization: Meanings, measures, and model evaluation. Philosophy of Science, 84(1), 115-159. https://doi.org/10.1086/688938

Broockman, D. E., Kalla, J. L., & Westwood, S. J. (2023). Does affective polarization undermine democratic norms or accountability? Maybe not. American Journal of Political Science, 67(4), 808-828. https://doi.org/10.1111/ajps.12719

Carlin, R. E., & Love, G. J. (2018). Political competition, partisanship and interpersonal trust in electoral democracies. British Journal of Political Science, 48(1), 115-139. https://doi.org/10.1017/S0007123415000526

Dorst, K. (2023). Rational polarization. The Philosophical Review, 132(3), 355-458. https://doi.org/10.1215/00318108-10469499

Finkel, E. J., Bail, C. A., Cikara, M., Ditto, P. H., Iyengar, S., Klar, S., Mason, L., McGrath, M. C., Nyhan, B., Rand, D. G., Skitka, L. J., Tucker, J. A., Van Bavel, J. J., Wang, C. S., & Druckman, J. N. (2020). Political sectarianism in America. Science, 370(6516), 533-536. https://doi.org/10.1126/science.abe1715

Graham, M. H., & Svolik, M. W. (2020). Democracy in America? Partisanship, polarization, and the robustness of support for democracy in the United States. American Political Science Review, 114(2), 392-409. https://doi.org/10.1017/S0003055420000052

Iyengar, S., Sood, G., & Lelkes, Y. (2012). Affect, not ideology: A social identity perspective on polarization. Public Opinion Quarterly, 76(3), 405–431. https://doi.org/10.1093/poq/nfs038

Iyengar, S., & Westwood, S. J. (2015). Fear and loathing across party lines: New evidence on group polarization. American Journal of Political Science, 59(3), 690–707. https://doi.org/10.1111/ajps.12152

Kahan, D. M. (2017). The expressive rationality of inaccurate perceptions. Behavioral and Brain Sciences, 40, e6. https://doi.org/10.1017/S0140525X15002332

Levendusky, M. S., & Malhotra, N. (2016). (Mis)perceptions of partisan polarization in the American public. Public Opinion Quarterly, 80(S1), 378–391. https://doi.org/10.1093/poq/nfv045

McCoy, J. (2024). Latin America’s polarization in comparative perspective. Latin American Politics and Society, 66(S2), 161–178. https://doi.org/10.1017/lap.2024.17

Mercier, H., & Sperber, D. (2011). Why do humans reason? Arguments for an argumentative theory. Behavioral and Brain Sciences, 34(2), 57–74. https://doi.org/10.1017/S0140525X10000968

Moore-Berg, S. L., Ankori-Karlinsky, L.-O., Hameiri, B., & Bruneau, E. (2020). Exaggerated meta-perceptions predict intergroup hostility between American political partisans. Proceedings of the National Academy of Sciences, 117(26), 14864–14872. https://doi.org/10.1073/pnas.2001263117

Nguyen, B. K., Fuller, M., Sinnott-Armstrong, W., & Murray, S. (2026). The partisan penalty: The effect of partisanship on evaluating reasons for political belief (Preprint). OSF Preprints. https://doi.org/10.31234/osf.io/8pk7z_v1