Call +1 352-376-0505

OZEAN MEDIA | Strategic Public Affairs

We win the outside game with data-driven advocacy.

The Anatomy of Political Hatred: What 135 Studies Reveal About Our Toxic Divisions

TL;DR

Political animosity is not random; political animosity is a predictable outcome of human psychology, media consumption, and social group identities.

Introduction

Political disagreement is normal. It is a core feature of a functioning democratic society. Animosity is different. The transformation of policy debate into personal contempt defines modern politics. This shift from issue disagreement to personal dislike harms social trust, personal relationships, and democratic norms.

I have written extensively on polarization and affective polarization. I find the topic fascinating.

Some examples of related writings:

You may also remember, that I love me a good meta-study. And this is a research paper I came across this week. It is a comprehensive review that synthesizes 135 peer reviewed studies to map the evidence in an attempt to identify the mechanisms that turn political opponents into personal enemies.

Asking, what do we know about the causes of affective polarization?

Citation and Links

Methodology

The authors conducted a systematic review of the scientific literature on affective polarization, following the PRISMA framework for transparent reporting. They searched Google Scholar and Scopus for articles published between 2012 and 2025, an initial haul of 761 records. After screening titles and abstracts for empirical, causal claims about affective polarization, and filtering out anything not published in a peer reviewed, SCImago indexed journal, 135 articles remained for analysis.

The corpus skews heavily toward the United States, which accounts for 58 of the 135 studies, with 22 more cross-national comparisons and a long-tail of single country cases from Spain, South Korea, Sweden, and elsewhere. Most of the evidence is also methodologically modest: 44.4 percent of studies are cross-sectional and associative, meaning they show a relationship without establishing causation. Experimental designs make up 28.9 percent, panel studies 20.0 percent, and true quasi-experimental designs, the kind that exploit real-world natural experiments to isolate cause and effect, just 3.0 percent.

To make sense of this sprawling literature, the authors coded each study along several dimensions: research design, country, outcome measured, and cause. For cause, they sorted every study into one of five driver families. One author recoded all 135 studies fifteen days after the initial pass to check the coding scheme’s stability, a step worth flagging now because it shapes how confidently the five categories below should be read.

Results and Findings

The review reveals that research on affective polarization has accelerated sharply since 2020.

The authors organized the evidence into five distinct driver families:

  1. Psychological factors, 32.6 percent of studies
  2. Media and information environment, 25.2 percent
  3. Identity and group processes, 23.0 percent
  4. Institutional factors, 13.3 percent
  5. Socioeconomic factors, 5.9 percent
  • Psychological Factors: Many studies attribute polarization to cognitive biases. People consistently overestimate the ideological extremity of their opponents. They assume others have the same information they do, so they conclude opponents are not merely uninformed but malicious or biased. This turns disagreement into a judgment of character.
  • Media and Information Environment: The way people consume information matters immensely. Active engagement with political content on social media is linked to higher polarization. Consuming news that confirms one’s existing beliefs, or pro-attitudinal media, is strongly associated with increased animosity toward the other side.
  • Identity and Group Processes: Partisan identity itself is a primary driver. For many, party affiliation is a stronger source of division than actual policy disagreement. Hostility is driven more by which team a person is on than by their specific beliefs. Social environments reinforce this: interacting exclusively with like-minded people in partisan “echo chambers” measurably increases affective polarization.
  • Institutional Factors: Polarization peaks before elections and eases afterward. Elections drive these peaks by surging warmth toward allies rather than immediate hatred toward opponents. Credible signals of interparty cooperation reduce polarization.
    • Interesting aside: short-run studies (elections) attribute the spike mostly to in-group warmth, but long-run trend data in the same paper shows the multi-decade rise is driven more by growing dislike of the out-party.
  • Socioeconomic Factors: Personal financial hardship does not predict partisan hostility. Economic inequality drives polarization when citizens believe social mobility is blocked and opportunity is unfair.

Frankly, as I write in Two Paths to End Political Polarization (But Only One Actually Works), I believe institutional factors such as first-past-the-post voting and Duverger’s law deserve much more attention, but my thinking is that the low odds of changing institutional factors are why they get researched less, but I digress.

Deep Dive: The Emergence Mode

Affective polarization does not have a single master cause. The evidence suggests it is an emergent property of a complex system. Think of it like a wildfire. Psychological biases, like the tendency to favor one’s own group, are the dry underbrush. A partisan media environment and curated social media feeds provide the sparks. Identity-based politics acts as the wind, fanning the flames and making the fire spread quickly.

Affective polarization is being driven at both a micro, meso, and macro levels:

  • Micro Level: Individual cognitive biases, personality traits, and moral convictions create the initial propensity to polarize.
  • Meso Level: Partisan media, separated social networks, and group identities supply the triggers and channels.
  • Macro Level: Elections and political cycles supply the timing and stakes.

The research points to a critical distinction: political animosity is often more expressive than instrumental. It is less about achieving a policy outcome and more about signaling loyalty to a social tribe.

This review also separates animosity into two types: horizontal and vertical. Horizontal polarization is citizen versus citizen, the dislike you may feel for a neighbor with a different yard sign. Vertical polarization is citizen versus elite, the dislike for politicians and parties on the other side.

Scholars study both forms in equal measure, but they rarely examine how the two interact. This is a critical gap. Dislike of opposing leaders and dislike of their voters are almost certainly connected, but the mechanics of that relationship are not well understood.

Why It Matters

The rise of affective polarization has serious consequences. When citizens view opponents as immoral or untrustworthy or now “evil”, the potential for political cooperation evaporates.

This gridlock prevents progress on critical national issues.

Socially, it erodes trust between neighbors, coworkers, and even family members. This breakdown in social cohesion weakens the foundations of civil society.

At its extreme, affective polarization can lead to the rejection of democratic norms and an increased tolerance for political violence. It changes the nature of politics from a contest over ideas to a conflict between enemies and J6 types of events. (Note: this is my claim, not the researchers.)

Critiques and Future Research

This review provides a comprehensive map, but it has boundaries.

  • Limits of the Study: The search strategy focused on the term “affective polarization. ” It may have missed studies on “partisan animosity” or “negative partisanship. ” The focus on English-language publications also creates a potential blind spot.
  • Potential Biases: The body of research is overwhelmingly focused on the United States. For my purposes, this is okay, but the dynamics of a two-party system are unique. Findings from the U.S. may not apply to multi-party democracies in Europe or elsewhere. In fact, would being in a multi-party democracy lower affective polarization? (This would be a fantastic topic for a grad student out there!!!!!)
  • What is Missing: There is a clear lack of research on how institutional factors, like electoral systems, and socioeconomic factors, like economic inequality, drive polarization. The link between citizen-to-citizen hostility and citizen-to-elite hostility also remains a major unanswered question.
  • What Should Be Studied Next: Future work should prioritize comparative research outside the U.S. to see if these findings hold in different political systems. Longitudinal studies that track the same individuals over many years are needed to separate short term political shocks from long term trends.
  • Measurement Issues: These studies use different ways of measuring affective polarization. There is no consensus in the field on how best to measure affective polarization.
  • Coding Reliability: In the author’s own re-coding exercise, they found only moderate stability, with the most disagreement clustered in the psychological / indentity boundary. For me this isnt as critical, because the major point is the paper’s main point of affective polarization being emergent and the 5 categories all work together and the lines are blurred in the real world.

Practical Implications for Policymakers

  • Recognize that election campaigns are predictable drivers of polarization. Develop communication strategies that emphasize national unity and de-escalate rhetoric in the post-election period.
  • Support and fund civic programs that encourage cross-partisan contact and dialogue. The evidence shows interaction can reduce animosity, especially among moderate citizens.
  • Incorporate education on cognitive bias and media literacy into public school curricula. Making citizens aware of their own mental shortcuts is a first step toward mitigating them. (Hey, read this blog!)

Practical Implications for Public Affairs Professionals

  • Craft messages that focus on policy substance, not the character or identity of the opposition. Avoid language that frames political opponents as an existential threat. If talking to opposition, you must do it in a non-identity threatening manner, by…..
  • Identify and leverage shared, non-political identities in messaging. Appealing to “higher” identities like parent, veteran, or local sports fan can create bonds that cut across partisan lines. Let’s not forget “American.”
  • Anticipate that polarization will spike during campaigns. Plan for this environment by preparing messages that can break through the noise with calls for common purpose or focus on widely accepted goals.

Conclusion

Political disagreement is not the problem. The transformation of that disagreement into personal contempt is.

The research is clear: political hatred is not an accident. It grows out of predictable interactions between cognitive bias, media diet, and group identity.

A wildfire needs dry brush, a spark, and wind. Affective polarization needs the same three ingredients, and the 135 studies in this review show exactly where each one comes from. Cognitive bias supplies the brush. Media environments supply the spark. Group identity supplies the wind. Take away any one ingredient, and the fire does not spread the same way.

The field has done a fairly good job in mapping the fire, but it hasn’t built a fire break. Only 3% of the 135 studies are quasi-experimental, and 44% are cross-sectional associative. The causal evidence could be stronger, because real interventions, the kind policymakers and public affairs professionals can actually run, need stronger tests: field experiments, panel studies, natural experiments that track the same people across an election cycle instead of taking a single snapshot.

Political disagreement will not disappear, and it should not. What can disappear is the machinery that turns a disagreement over economic policy into a belief that your neighbor is evil and not worthy of discussion or understanding.

Navigate Your Next Move.

We help organizations win contested political and regulatory fights using campaign-style strategy.

Don’t leave your outcome to chance.

Let’s Talk Strategy

20-minute introductory call.  No obligation.

RELATED INTELLIGENCE