Political Polarisation: Why Societies Divide and When It Becomes Dangerous
Political polarisation is often described as a society becoming divided, but disagreement alone is not evidence that democracy is failing. Citizens in a free political system are expected to disagree about taxation, welfare, migration, religion, climate policy, education, national identity and the proper limits of government. Elections exist partly because societies do not share one view of how they should be governed.
The important question is what disagreement becomes.
Political polarisation occurs when political positions, identities or feelings become increasingly organised around separated camps. Sometimes this means parties or citizens move farther apart ideologically. Sometimes policy differences change relatively little while supporters develop stronger dislike and distrust of one another. In other cases, political identity becomes aligned with religion, geography, education, culture or social class until one political divide begins to organise many parts of social life.
Researchers therefore distinguish several forms of polarisation rather than treating the term as one measurement. Markus Wagner's 2024 review of European research, for example, defines affective polarisation as the gap between positive feelings toward one's preferred political group and negative feelings toward supporters of competing groups. It also stresses that measuring polarisation becomes more complicated in multiparty democracies than in two-party systems. (cambridge.org)
This distinction is essential because a society can disagree intensely without becoming incapable of democratic coexistence. The danger increases when political opponents cease to be understood as legitimate citizens with different views and begin to be imagined as enemies who cannot safely be allowed to participate or govern.
Ideological Polarisation and Affective Polarisation Are Different
Ideological polarisation concerns distance in political beliefs.
Imagine citizens distributed along a policy spectrum. If large groups increasingly cluster near opposing ends while the middle shrinks, ideological polarisation has increased. The same idea can be applied to political parties, legislators, activists or voters.
But those groups do not necessarily polarise in the same way or at the same time.
Political elites may move farther apart while ordinary voters remain comparatively moderate. Citizens may divide sharply over one question but agree on several others. A country's parties can polarise economically while converging on aspects of foreign policy, or divide culturally while retaining similar positions on some public services.
This is why measuring ideological polarisation requires examining the actual distribution of opinions rather than treating every heated political argument as evidence that society has split into two ideological extremes.
Affective polarisation, by contrast, is about feelings toward political groups. Someone may hold relatively moderate policy views while strongly disliking the people who support another party. Politics becomes less about “I disagree with your policy” and more about “people like you are untrustworthy, immoral or dangerous.”
Research in Europe shows that affective polarisation is not merely an American two-party phenomenon. Wagner's review finds substantial partisan hostility across European multiparty systems while also emphasising that the concept must be measured differently when citizens choose among several parties rather than two opposing blocs. (cambridge.org)
The two forms can reinforce each other, but they should not be treated as identical. Citizens can become more hostile without moving equally far apart on policy, and ideological disagreement does not inevitably produce hatred.
That distinction changes how polarisation should be addressed. If the problem is genuine policy disagreement, asking everyone simply to “get along” does not resolve the dispute. If the problem is exaggerated hostility toward people whose actual policy preferences are less extreme than imagined, better information and cross-group contact may help.
Political Identity Becomes More Powerful When Other Identities Line Up With It
Political disagreement becomes especially potent when party identity overlaps with other social identities.
Suppose political preference increasingly predicts where someone lives, what media they consume, which religion they practise, their educational background, social networks and cultural tastes. A disagreement over government policy can then become a signal of an entire way of life.
Political scientists often describe this process as sorting.
Sorting does not necessarily mean every citizen becomes ideologically extreme. It means identities that once crossed political boundaries increasingly line up on the same divide. Someone may encounter fewer neighbours, colleagues, relatives or friends who belong to the opposing political camp.
That makes stereotypes easier to maintain.
If people personally know many political opponents, they encounter variation inside the group. One opponent may be generous, another difficult, one religious, another secular, one wealthy and another poor. When cross-cutting relationships decline, people can begin imagining the out-group through politicians, activists and highly visible online examples instead.
The most provocative members of a political camp may then become psychologically representative of millions of quieter supporters.
This creates a gap between actual polarisation and perceived polarisation. Citizens can believe the opposing side is more internally uniform, radical or hostile than it really is.
But correcting such perceptions is not a universal solution. A 2024 PNAS study combining panel data, a meta-analysis and preregistered experiments found that correcting Americans' misperceptions about opposing partisans did not consistently reduce partisan animosity, support for political violence or support for undemocratic practices. The authors concluded that misperception correction can help under some circumstances but should not be treated as a general cure for democratic division. (pubmed.ncbi.nlm.nih.gov)
The lesson is important: polarisation is maintained by more than factual error.
Identity, status, institutions, emotions and political incentives also matter.
Leaders, Media and Digital Platforms Can Raise the Stakes of Conflict
Citizens do not form political interpretations independently of political institutions.
Parties, politicians, campaigners, journalists, influencers and public intellectuals provide cues about what issues matter, which events should produce anger and whether political opponents should be treated as legitimate competitors.
Elite language can therefore influence the boundaries of political conflict.
A leader can say that an opposing policy is dangerous while still recognising the opposition's right to compete democratically. Or the leader can describe opponents themselves as illegitimate, corrupt by definition or enemies whose political participation threatens national survival.
Those frames do not automatically determine what citizens believe, but repeated signals can change how supporters interpret political competition.
Media incentives add another layer. Conflict is legitimately newsworthy because democracy requires scrutiny of disagreements and competing interests. Yet political communication systems also reward drama, novelty and emotional clarity. The most inflammatory statement can become more visible than a complicated compromise supported by a quieter majority.
Digital platforms accelerate this process by making political approval and hostility measurable through likes, shares, reposts and comments.
A large 2021 PNAS study examining millions of social-media posts found that references attacking or criticising political out-groups were especially strongly associated with engagement. The authors concluded that out-group language was the strongest engagement predictor among the factors they measured, creating an incentive structure in which hostility toward opponents can become unusually visible. (pnas.org)
But the claim that social media alone created contemporary polarisation goes much further than the evidence permits.
Political division existed before modern platforms, differs substantially among countries using similar technologies and is shaped by party systems, television, economic change, institutions and historical conflicts.
Even the assumption that exposing people to opposing political views automatically makes them more moderate has performed poorly in some experiments. A widely cited 2018 study that paid American Twitter users to follow bots retweeting politicians and commentators from the opposing side found no depolarising effect and found evidence of increased ideological polarisation among some participants. (pnas.org)
The quality and context of political exposure therefore matter.
Showing someone the most hostile arguments from the opposing camp may reinforce rather than weaken stereotypes.
Economic Insecurity, Inequality and Status Can Feed Political Division
Political polarisation cannot be reduced to economics, but material conditions can influence which political conflicts become powerful.
Unemployment, regional decline, housing pressure, inequality, insecure work and rapid technological change can create genuine grievances. Political movements then compete to explain why those problems exist and who is responsible.
Objective hardship is only part of the story.
People evaluate their situation relative to expectations, other communities and earlier generations. A national economy can grow while a particular town loses industries that once provided stable work and status. A graduate can earn more than their parents did while feeling less economically secure because housing costs have increased faster.
Such experiences do not mechanically produce one political ideology. Different movements can organise the same economic frustration around taxation, class, migration, regional identity, globalisation, corruption or cultural status.
Inequality can similarly interact with polarisation without creating it automatically.
Large economic inequalities may intensify disputes over taxation, redistribution and access to political influence. If those economic divisions overlap with ethnicity, religion, geography or party affiliation, material conflict can become an identity conflict as well.
Yet some unequal societies remain politically stable while more economically equal societies develop severe cultural divisions.
Institutions and political narratives mediate the relationship.
Trust becomes particularly important here. The OECD's 2024 survey of 30 member countries described political polarisation as one of the pressures confronting democratic governments and found major differences in institutional trust according to whether citizens believed government was responsive, accountable, reliable and open to participation. (oecd.org)
People are more likely to tolerate losing a political contest when they believe institutions will still treat them fairly afterward.
When every institution is interpreted as permanently captured by an enemy camp, political competition becomes more zero-sum.
Polarisation Is Not Automatically Bad for Democracy
Calls to “end polarisation” can themselves become politically misleading.
Democracy needs meaningful disagreement.
A party system in which every major organisation offers essentially the same policies may produce frustration rather than harmony. Social movements often emerge precisely because an existing political consensus excludes a serious injustice or ignores an affected group.
Strong disagreement over slavery, civil rights, authoritarianism or equal political participation cannot sensibly be criticised simply because it produces polarisation.
Nor should “moderation” be treated as an independent moral virtue.
Sometimes one political position really is more compatible with evidence, rights or constitutional rules than another. A neutral analysis should not force artificial symmetry by assuming that two political camps have necessarily moved equally far from democratic norms.
Symmetry is an empirical question.
Researchers should ask who moved, on what dimension, over what period and according to which evidence.
Political disagreement becomes dangerous for democracy for a different reason: when people abandon the procedural commitment that opponents who follow democratic rules retain the right to compete, speak, vote and potentially govern.
This is why the distinction between an opponent and an enemy is central.
An opponent can defeat you in an election and still remain a legitimate citizen.
An enemy must be defeated by whatever means are necessary.
The movement from the first idea toward the second is where polarisation becomes especially dangerous.
Hostility Does Not Automatically Become Political Violence
Affective polarisation often attracts concern because of its possible relationship with intolerance and violence. But the connection should not be simplified.
A 2024 study using nationally representative panels in the United Kingdom and Norway examined political avoidance, intolerance and support for violence. It found that simple dislike of political opponents predicted some willingness to avoid them but performed increasingly poorly as an explanation of more serious norm-breaking behaviour. Support for violence was more closely associated with particular emotions and characteristics such as anger, fear, Manichean worldviews and relative deprivation than with ordinary out-group dislike alone. (onlinelibrary.wiley.com)
This matters because the statement “polarisation causes violence” can obscure several steps in between.
Millions of people can dislike political opponents without supporting violence.
More serious warning signs include dehumanisation, willingness to deny opponents basic political rights, acceptance of violence by one's own side, claims that every election loss is inherently illegitimate and willingness to suspend constitutional rules to prevent opponents from governing.
At that point, the issue is no longer simply strong political disagreement.
It is democratic tolerance.
Research also suggests that support for explicitly antidemocratic behaviour may be less widespread among ordinary citizens than political rhetoric sometimes implies. A 2024 PNAS study found that Democrats and Republicans in the United States overwhelmingly rejected many forms of partisan violence and democratic norm violation, even though elite behaviour and public perceptions sometimes suggested much greater hostility. (pmc.ncbi.nlm.nih.gov)
The implication cuts in both directions.
There is no reason for complacency about antidemocratic rhetoric or violence.
But exaggerating how much ordinary citizens support those behaviours can itself create a false picture of a society already beyond democratic repair.
Reducing Polarisation Is More Complicated Than Making Everyone Agree
If disagreement is normal and sometimes desirable, depolarisation cannot mean eliminating political differences.
A more useful objective is reducing unnecessary hostility while strengthening the rules that allow disagreement to continue peacefully.
Research offers some promising approaches, but there is no universal intervention.
A major 2024 Science megastudy involving more than 32,000 American participants tested 25 interventions designed to reduce partisan animosity and antidemocratic attitudes. Several reduced hostility, particularly approaches that presented sympathetic people holding different political views or emphasised identities shared across partisan groups. Other interventions reduced support for undemocratic practices or partisan violence. Crucially, however, the study found that partisan animosity, support for undemocratic behaviour and support for political violence are distinct attitudes rather than one single problem. (science.org)
An intervention that makes someone like political opponents more does not necessarily strengthen their commitment to democratic institutions.
Cross-partisan conversation presents a similar mixed picture. Experiments by Erik Santoro and David Broockman found that structured conversations can produce short-term improvements under some conditions, particularly when people establish positive social connections, but the effects were limited, conditional and often did not persist. (science.org)
This makes intuitive sense.
A ten-minute conversation cannot permanently repair political divisions generated by years of institutional incentives, media environments, social sorting and historical conflict.
Durable depolarisation is therefore partly an institutional project.
Elections must allow people to lose without losing basic rights. Courts and administrative institutions must apply rules credibly. Public services have to demonstrate sufficient competence that citizens do not interpret every failure through partisan identity. Political leaders have to retain incentives for cooperation rather than receiving only rewards for humiliation and obstruction.
The OECD's work on trust makes a similar institutional point: citizens' confidence depends substantially on perceptions of government reliability, responsiveness, fairness, integrity and openness. (oecd.org)
Trust does not require agreement with whichever government holds office.
It requires enough confidence in the system that political defeat does not feel equivalent to exclusion from the political community.
Individuals Can Resist the Social Logic of Polarisation
Citizens cannot individually redesign electoral systems or platform algorithms, but everyday behaviour can either reinforce or weaken political sorting.
One useful discipline is to stop treating the most extreme viral example as representative of an entire political group.
Millions of people rarely share one motive.
When possible, ask people to describe their own position instead of inferring it from a party label. Distinguish factual disputes from moral disagreements. Criticise specific policies, statements or actions rather than treating every controversy as evidence about the character of everyone who votes a particular way.
Cross-cutting relationships also matter.
A neighbour can be a political opponent and still help during an emergency. Colleagues who vote differently can cooperate professionally. Parents may disagree nationally while working together to improve a school. Sports clubs, workplaces, neighbourhood associations, religious organisations and professional groups create identities that do not map perfectly onto political camps.
These relationships do not eliminate disagreement.
They make political identity less total.
Local institutions can be particularly useful because they force cooperation around concrete problems. Water systems, rubbish collection, school safety and disaster response do not disappear because the people solving them voted differently.
The objective is not friendship with every political opponent.
It is retaining the ability to recognise that political categories describe only part of a person.
How Do Researchers Know Whether a Society Is Polarised?
Because political polarisation has several dimensions, there is no single universally sufficient measure.
Researchers may examine ideological distance by asking where citizens or parties stand on policy questions. They may measure affective polarisation by comparing people's feelings toward preferred and opposing political groups. Social-distance questions ask whether people would accept opponents as neighbours, friends, colleagues or family members.
Other measures examine party-system distance, discrimination against political out-groups, willingness to cooperate, attitudes toward political violence or support for undemocratic actions.
Each answers a different question.
This is why a headline announcing that a country has become “more polarised” should prompt an immediate follow-up:
More polarised in what sense?
Parties may be farther apart while citizens are not.
Citizens may dislike one another more while holding similar policy positions.
One issue may become sharply divided while the rest of the policy landscape changes little.
A multiparty democracy can contain several competing blocs rather than one binary divide.
Wagner's comparative review emphasises precisely this measurement problem, particularly in European systems where partisan hostility cannot simply be represented as one gap between two parties. (cambridge.org)
Political polarisation is therefore better understood as a family of related phenomena than as one universal score.
Frequently Asked Questions
What is political polarisation?
Political polarisation is the increasing separation of political groups in ideology, identity or emotional attitudes. It can refer to greater distance in policy beliefs, stronger hostility toward political opponents or the alignment of multiple social identities along political lines.
What is affective polarisation?
Affective polarisation describes the emotional divide between political groups—typically greater warmth toward one's preferred group combined with stronger dislike of political opponents. It can increase even when policy differences do not increase by the same amount. (cambridge.org)
Is political polarisation always harmful?
No. Clear political alternatives, strong advocacy and disagreement over injustice can be healthy components of democratic politics. Polarisation becomes especially concerning when opponents are treated as illegitimate citizens, political violence becomes acceptable or democratic rules are abandoned to prevent the other side from governing.
Does social media cause political polarisation?
Social media can amplify hostile political content and change how quickly conflict spreads, but it is not the sole cause of polarisation. Research shows that out-group attacks can attract high engagement, while experiments also show that simply exposing users to opposing views does not necessarily reduce polarisation. (pnas.org) (pnas.org)
Are echo chambers the main cause of polarisation?
No single mechanism explains political polarisation. Like-minded networks can reinforce identity, but political leaders, party systems, economic conditions, traditional media, institutions and historical conflicts also matter. Exposure to opposing views can sometimes reduce hostility and sometimes intensify defensiveness.
Can political polarisation be reduced?
Yes, although there is no simple cure. Large experiments have found that shared identities, relatable members of opposing groups and some forms of corrective information can reduce particular forms of partisan hostility or antidemocratic attitudes. The effects differ according to what type of polarisation is being targeted. (science.org)
Democracy Does Not Require Agreement—It Requires Legitimate Opposition
The most important lesson about political polarisation is that democracy was never designed to eliminate disagreement.
It was designed to organise it.
Citizens can disagree profoundly over economic policy, religion, immigration, social norms or the country's future and still remain members of the same political community.
The system becomes fragile when that shared membership disappears.
When every election is portrayed as the final battle between good and evil, losing becomes intolerable. When opponents are assumed to be inherently illegitimate, compromise becomes betrayal. When institutions are trusted only when one's own side controls them, each group has an incentive to capture those institutions before the other one does.
That spiral is more dangerous than disagreement itself.
Research also cautions against reducing the problem to one cause or one cure. Ideological distance, affective hostility, social sorting, mistrust, antidemocratic attitudes and political violence are related but distinct. Social media can amplify conflict but did not invent it. Misperceptions can matter but correcting them does not always solve the problem. Cross-party conversation can help under some conditions without repairing structural incentives that continue rewarding polarisation. (science.org) (pubmed.ncbi.nlm.nih.gov)
Nor should the goal be compulsory centrism.
A free society needs citizens willing to defend strong principles, organise movements and challenge established power. It also needs them to recognise that people with opposing views retain political rights.
The boundary that matters is therefore not between agreement and disagreement.
It is between opposition and illegitimacy.
A democratic society can survive citizens believing that the other side is wrong.
It becomes much harder to sustain when citizens conclude that the other side has no right to exist as a political opponent at all.
