The psychological underpinnings of democracy and democratic backsliding
The Social Psychology of Democracies in Transition
Overview of the lecture
The lecture asks the cumulative question that the course has been building toward: are there psychological pre-requisites for democracy itself, and do CEE populations possess them? The first half surveys the literature on democratic values and applied political tolerance, identifies the determinants of intolerance (threat perception, internalised democratic values, personality), and examines the “least-liked group” method that reveals how much narrower applied tolerance is than abstract support for democratic values. The second half turns to cross-national comparative work, with particular attention to Marquart-Pyatt and Paxton’s 2007 analysis of tolerance learning in Eastern and Western Europe, which tests the democratic-learning hypothesis using World Values Survey data from the 1990s. The third and final section applies the apparatus to contemporary democratic backsliding in Hungary, Poland, and Slovakia, drawing on recent research by Gessler and Wunsch (2025) and Sadowski (2024) to show that the psychology of backsliding is the mirror image of the psychology of democratisation. The mechanisms that built democratic engagement in the 1990s can, under different conditions, be run in reverse.
Democratic values and civil liberties
What democracy requires of citizens
Democracies make demands of citizens that other forms of government do not. Citizens must tolerate their fellow citizens’ participation in politics, even when those citizens’ views diverge sharply from the norm. They must participate in collective self-governance rather than leaving it to others. And they must recognise a loyal opposition — they must accept that political opponents are not enemies, that opposition is legitimate, and that the peaceful transfer of power is a routine feature of democratic life rather than a crisis. These demands are psychologically unusual, and they do not follow automatically from the institutional arrangement of democratic government. A society can have elections, parliaments, courts, and constitutions without having citizens who make the psychological demands of democracy work in practice.
This lecture focuses on the psychological underpinnings of the restraint that creates space for a wide range of political groups to participate, and on the attitudes, values, and orientations that lead citizens to participate themselves. Together, these constitute what the literature calls the civic culture of democracy: not the institutions, but the dispositions that make the institutions work.
Democratic values as consensual agreement
For democracy to exist in a functioning form, certain attitudes must be widespread across the population. Procedural norms governing how substantive disputes are negotiated must be accepted; fundamental values such as liberty, equality, and individualism must be endorsed; political tolerance and interpersonal trust must be sufficiently distributed. The development of the sample survey in the mid-twentieth century made it possible to test the theory that these psychological and cultural prerequisites for democracy actually existed in democratic populations.
Early research in the United States produced mixed results. On the one hand, many surveys suggested limited support for and understanding of democratic rules among the broad citizenry: respondents frequently endorsed abstract democratic values in one breath and rejected specific applications of those values in the next. On the other hand, support for democratic rules and principles was stronger among the better-educated and more politically active segments of the population, suggesting an “activist class” with the proper psychological attitudes, values, and predispositions to sustain democracy. Over time, with rising education levels in most democracies, support for democratic values in the mass public has risen.
Civil liberties versus political tolerance
The distinction between abstract support for democratic values and applied political tolerance has become the central empirical finding of the field, and it has its historical roots in religious tolerance. The concept of tolerance evolved from efforts to moderate the harmful and often violent effects of religious conflict. Religious tolerance was promoted as a mechanism to allow religions to “put up with” each other, even when they disliked or hated each other vehemently. The concept of political tolerance evolved as a way to live with one’s ideological and political enemies: political opponents need not be eliminated physically or even politically; one need not like or support one’s opponents and their ideas, but one ought at least to tolerate them, to respect their right to participate.
Applied tolerance in this sense means supporting the political rights of groups and ideologies that one actively opposes and dislikes. It is a demanding standard, and research has consistently shown that abstract support for democratic values is not a reliable predictor of applied tolerance in specific cases. The “least-liked group” method, developed by Sullivan, Piereson, and Marcus in the late 1970s, operationalises this insight: respondents are first asked to identify the group in politics that they like the least, and then asked to say how far they would support the civil liberties of that group. Using this method, researchers have consistently found that the public is far less tolerant of other groups than their self-declared democratic values might suggest.
The determinants of political intolerance
The tolerance of elites
Successive studies have shown that political elites tend to be more supportive of civil liberties than ordinary citizens. One cause of this differential is political expertise and participation: repeated engagement with the rules of democratic politics tends to socialise respect for those rules. But other factors matter too, and research has identified a cluster of variables that predict applied tolerance across many contexts.
Threat perception
Perceived threat has one of the strongest direct relationships to levels of political tolerance. Two kinds of threat matter. Dispositional threat refers to a chronic tendency to feel threatened, which some people hold as a stable personality characteristic (associated with low openness to experience and high neuroticism). Situational threat refers to short-term variations in how threatening the current information environment represents specific groups to be. Some people have a predisposition to be easily threatened and are therefore very sensitive and responsive to potential threats in the political environment; they are less tolerant than individuals who are more calm and more easily reassured.
The information environment is malleable and politically consequential. For many citizens, information describing the nature and activities of potentially unpopular groups has a profound impact on their level of tolerance toward those groups and their ideas. If the information environment portrays such groups as violating normative expectations with regard to orderly behaviour and proper procedures, many citizens — even those not particularly predisposed toward intolerance — refuse to tolerate the group and its activities. If the information environment portrays the same groups as behaving properly and in an orderly fashion, far more people — often a majority — will tolerate the group and its activities, despite the group’s unpopular or extremist image. Aggregate levels of intolerance are therefore somewhat malleable, depending on how political elites and the media portray those with less popular ideas. Threat perceptions, both dispositional and environmental, play a central role in determining whether a set of citizens will internalise and apply the democratic principles of restraint and tolerance, or whether they will set those principles aside in particularly difficult situations.
Internalisation of democratic values
Early analyses of tolerance emphasised the finding that most ordinary citizens were inconsistent and perhaps even hypocritical. Most citizens expressed a strong belief in democratic values but coupled that belief with an unwillingness to apply those values to groups and ideologies they found most objectionable. It would be easy, and even logical, to conclude that the ideas and ideals of democracy operate at a level disconnected from actual practices — in other words, that ideas do not guide behaviour.
More recent work has qualified this conclusion. The internalisation of democratic norms can make a considerable difference in determining how tolerant an individual will be when their tolerance is put to the test. The more completely an individual has internalised and believes strongly in democratic norms, the more likely the individual is to tolerate groups and ideas they find obnoxious. Threat perceptions and strength of belief in the more abstract norms of democracy are both very strong predictors of applied tolerance judgements. Democratic values, in other words, do matter; they are simply not the only thing that matters, and they are weaker against strong situational threat than enthusiasts of democratic education have sometimes assumed.
Personality
In addition to threat perceptions and democratic norms, personality characteristics predict applied political tolerance. Studies have shown that misanthropy, anomie, self-esteem, flexibility, dogmatism, and trust all predict applied tolerance scores. On the Big Five, openness to experience is the robust predictor, consistently associated with higher tolerance, while neuroticism — through its link to threat-sensitivity — is associated with lower tolerance; other traits, including extraversion, show weak or inconsistent effects. These personality effects are relatively context-independent: highly dogmatic, authoritarian personalities are intolerant regardless of where they live, while highly open personalities are tolerant regardless of context. This gives personality a distinctive place in the overall architecture of tolerance: it is not the product of situational factors alone but reflects stable individual differences that interact with situational ones.
Summary of the empirical record
Empirical research over the last four decades has demonstrated that there are strong psychological underpinnings to what political scientists have called “support for the rules of the political game”. At the abstract level, there is consensual support for democratic values across most contemporary democracies. When these values are applied to difficult cases, however, there is far less consensus. Applied tolerance has increased somewhat over time in many countries, but the increases are modest at best. In general, political experts exhibit higher levels of applied tolerance than political novices. Strongly internalised beliefs in democratic values “constrain” citizens to be more tolerant in practical situations, though the constraint is not absolute. Individual differences in both perceptions of threat and personality predispositions influence applied tolerance judgements considerably.
Cross-national studies of political tolerance
The comparative question
It is difficult to make broader claims about psychological prerequisites for democratic attitudes and systems without conducting cross-national, comparative studies. The central question is whether the underlying relationships between psychological characteristics and institutional democracy are sufficiently sensitive to historical and political context that generalisation is virtually impossible, or whether certain psychological predispositions and attitudes invariably sustain a democratic political culture and institutions, while other predispositions and attitudes invariably subvert them. The evidence suggests that the answer is intermediate: personality-based intolerance (dogmatism, authoritarianism, closed-mindedness) is relatively context-independent, while institutional and cultural effects on tolerance vary substantially across contexts.
Gibson (1992) on the former Soviet Union
A classic study of the former Soviet Union in the early 1990s by James Gibson found that Russians were quite supportive of the general norms of democracy, but when asked to apply these norms to disliked groups they were much less tolerant than those in more long-standing democracies. By far the most significant predictor of intolerance among Russian citizens was perceived threat. A subsequent study by Gibson and Duch (1993) found that dogmatism also increased intolerance in the former Soviet Union. The Gibson findings established two patterns that have been replicated across much of the post-communist literature: that abstract support for democracy is higher in new democracies than the applied tolerance literature would predict, and that threat perception (rather than principled rejection of democracy) is the core mechanism that produces the applied-tolerance deficit.
Attitudes, economic development, and democratic participation
Scholars continue to debate whether the attitudes associated with the civic culture cause transitions to democratic regimes, ensure that democratic regimes persist over time, or are actually a result of the pre-existence of democratic regimes. It is difficult to demonstrate conclusively that psychological orientations cause democracy rather than result from it, and even time-series data do not resolve the question definitively. Ronald Inglehart (1977, 1990, 1997) argues that economic development can lead to changes in people’s psychological orientations and preferences, and that this in turn creates attitudes and expectations that favour democracy over less participatory forms of government, and sustain democracy once it begins to develop. The growth of a service sector is particularly critical in the Inglehart framework, because careers in this sector of the economy demand skills (problem-solving, communication, abstract reasoning) that are particularly conducive to democratic participation.
Interpersonal trust and subjective well-being significantly affect the duration and level of democracy. Interpersonal trust lends credibility to the concept of a loyal opposition: when people do not trust their fellow citizens, elections and transitions of power appear far more dangerous than when they do, and the peaceful alternation of governing parties becomes harder to sustain. Cognitive mobilisation — the increase in skills and motivation required to engage in decision-making, including political decision-making — combines with widespread material and physical security to produce a shift from material to post-material values. Post-material values (individual self-expression, quality of life, tolerance, environmental concern, gender equality) in turn support democracy and civil liberties. The overall pattern in the Inglehart framework is a developmental sequence: economic development → cognitive mobilisation → post-material values → democratic consolidation.
Marquart-Pyatt and Paxton (2007)
The study
Sandra Marquart-Pyatt and Pamela Paxton’s 2007 Political Behavior article, “In Principle and in Practice: Learning Political Tolerance in Eastern and Western Europe”, applies the civic-culture framework to the specific case of comparative post-communist tolerance. The study covers 14 countries: 5 Western European, 8 Eastern European, and the United States. The core finding: while levels of espoused support for democracy in Eastern Europe are similar to those in the West, levels of political tolerance are lower. The authors test the democratic-learning hypothesis using an analysis of age cohorts in Eastern Europe.
The democratic-learning hypothesis
Learning theories of tolerance argue that political tolerance must be learned, because it is a cognitively difficult concept. The argument has two channels. Tolerance can be learned through increased contact with diverse groups, which helps people understand the other and leads to increased tolerance. And tolerance can be learned through exposure to democratic institutions, which increases exposure to the dominant democratic norms and thereby encourages political tolerance. Marquart-Pyatt and Paxton use Eastern Europe as the test case because, within EE, the institutional learning environment has varied substantially across cohorts: the post-1989 cohort grew up under democratic institutions, the “thawed” post-1956 cohort grew up under relatively less repressive communism, and the earliest cohort grew up under Stalinist-type arrangements. If democratic learning is operative, the youngest cohort should be more tolerant than the middle cohort, which in turn should be more tolerant than the oldest. Western Europe and the US function as controls, because within these countries all cohorts were educated under democracy and no within-country learning effect should be visible.
Data and design
Marquart-Pyatt and Paxton use 1995–1997 World Values Survey data, a period when communism was still recent but respondents had some time to adjust to transition. The dependent variable is applied tolerance: respondents were asked to pick the group they liked least from a list (Jews, capitalists, communists, immigrants, homosexuals, criminals, neo-Nazis), and then asked whether their most-disliked group should be allowed to hold public office, teach in schools, and hold public demonstrations. The three tolerance questions were combined into a single index. The independent variables include measures of contact with diverse ideas and people (education, urban residence, association memberships), political orientations (interest in politics, views of democracy, political ideology), threat perception (trust, national pride, protectionism), religiosity, democratic activism, and age cohorts (educated under democracy, educated under “thawed” period, educated with no democratic or thawed period of education).
Descriptive findings
The descriptive findings confirm the applied-tolerance gap. The share of respondents willing to allow their most-disliked group to hold public office is 21.6% in the United States, much lower in Western-industrialised countries (Finland 13.6%, Spain 2.7%, Sweden 9.8%, Switzerland 4.9%, West Germany 6.5%), and lower still in Eastern Europe (Bosnia 2.1%, Bulgaria 5.1%, Croatia 3.4%, East Germany 3.2%, Macedonia 4.7%, Poland 2.8%, Serbia/Montenegro 5.0%, Slovenia 5.5%). Similar patterns appear for the other two tolerance items and for the composite index. Espoused support for democracy is remarkably consistent across the three country groups — 3.4 in the United States, approximately 3.3–3.6 in Western Europe, approximately 3.0–3.6 in Eastern Europe — confirming that the gap is in application rather than in abstract endorsement.
Regression findings
The key predictors of tolerance are strongly associated with it in the United States but less so elsewhere. The effect sizes of many variables are weaker in Western Europe and much weaker in Central and Eastern Europe. Across all three groups, positive predictors include education, urban residence, interest in politics, espoused support for democracy, and democratic activism. Negative predictors include national pride and protectionism. The age-cohort effect is the specific test of the democratic-learning hypothesis. In the United States and Western Europe, there is no significant difference in tolerance across the three age cohorts, suggesting no difference in the democratic socialisation of these age cohorts. In Eastern Europe, compared to the reference group of 18–24-year-olds, 25–44-year-olds are less tolerant, and those older than 45 are even less tolerant. Age cohorts in Eastern Europe are distinguished in their levels of tolerance by their experience with democratic, or less totalitarian, governments. The democratic-learning hypothesis is therefore supported for CEE: exposure to democracy during one’s formative years increases applied tolerance.
Why the weaker effects in CEE?
One speculative explanation for the smaller effect sizes of individual-level predictors in CEE is the institutional uncertainty present in the region. No matter what individual characteristics might lead an individual in CEE to political tolerance, the greater institutional uncertainty in CEE may cause citizens to be unwilling to translate democratic values into practice, for fear that others will not fairly play the game. Tolerance, in this reading, is an equilibrium outcome: if I suspect you will not tolerate me, I am less willing to tolerate you, and the psychological dispositions that would support tolerance in a more predictable environment are less effective in an uncertain one. This is the “tolerance-coordination problem”, and it suggests that the remedy for low applied tolerance in CEE may be institutional stability as much as individual-level attitude change.
From tolerance to backsliding
The backsliding reversal
Marquart-Pyatt and Paxton’s data are from 1995–97. The prediction of the democratic-learning framework would have been that continued exposure to democracy would raise applied tolerance in CEE over subsequent decades. That prediction has been partly confirmed: generational replacement has produced measurable shifts in values (Lecture 13) and support for democracy has remained high or risen in most CEE countries. But the prediction has been partly inverted by the phenomenon of democratic backsliding. Since 2010, several CEE countries have experienced deterioration of democratic institutions, led by governing parties that have rolled back civil liberties and undermined opposition rights. Hungary under Fidesz from 2010 to 2026, Poland under PiS from 2015 to 2023, and Slovakia under Fico from 2024 all exemplify the pattern. Serbia and, on different trajectories, Russia and Belarus fit the broader regional story of democratic erosion.
The backsliding phenomenon returns salience to the question of democratic psychological foundations, because if governments can dismantle institutional democracy relatively rapidly, the stability of citizens’ democratic dispositions matters for whether the dismantling succeeds or fails. The global context remains adverse: V-Dem’s 2026 report counts 44 countries becoming more autocratic, with six of the ten newest cases in Europe and North America — which makes the CEE reversals of 2023–2026 all the more notable as counter-examples.
Backsliding and applied tolerance
Marquart-Pyatt and Paxton’s finding that applied tolerance depends on the information environment predicts precisely what backsliding regimes do to undermine tolerance. Governing parties that repeatedly portray opposition as illegitimate — opposition figures as “traitors” or “agents of foreign powers”, civil society as “enemies of the nation”, independent media as “fake news” — reduce citizens’ applied tolerance for these groups. The information environment, when used by the governing party against democratic norms rather than in support of them, moves applied tolerance in the opposite direction. Backsliding is not therefore a case where the Marquart-Pyatt-Paxton framework fails; it is a case where the framework correctly predicts the outcome, given the information environment that has been produced.
Affective polarisation and backsliding
Recent work has added the concept of affective polarisation to the backsliding literature. Affective polarisation — emotional dislike of out-parties and their supporters — is increasingly central to explanations of backsliding. Partisan-based affective polarisation is posited as a key explanation for citizens’ tolerance of democratic backsliding: voters are more willing to overlook democratic violations when those violations come from their in-party. Gessler and Wunsch (2025, European Journal of Political Research 64(4): 1593–1617) reverse the usual causal arrow: rather than polarisation simply driving backsliding, sustained backsliding itself generates affective polarisation organised around the defence of democracy. Prolonged erosion opens a “new regime divide” — a democracy cleavage that cuts across the old left–right axis. In Hungary this has let an ideologically diverse opposition coalesce around defence of liberal democracy itself; in Poland (analysed in their appendix) the same dynamic is visible in weaker form, but the 2023 parliamentary election shows it can be activated as a mobilising resource.
Incumbent actions that degrade democracy worsen polarisation, and the effect is amplified when party elites endorse rather than condemn the leader’s behaviour. Polarisation is endogenous to democratic backsliding, which means that backsliding and polarisation reinforce each other, producing a feedback loop in which each round of backsliding produces a more polarised response, which facilitates the next round of backsliding.
Rising support for democracy despite backsliding
Despite the feedback loop, diffuse support for democracy has held up in several CEE countries. Ireneusz Sadowski’s 2024 analysis of Polish attitudes toward democracy 1995–2021 documents a rising wave of support for democracy across the period, even as institutional trust has fluctuated. One interpretation is that visible backsliding has raised the value citizens place on democracy: they see what is being taken away, and they value it more as a consequence. Generational replacement compounds the effect, because younger Polish cohorts support democracy at higher rates than older cohorts. The 2023 Polish electoral reversal was the political expression of these attitudinal shifts — though the 2025 presidential election, won by the PiS-backed Karol Nawrocki, checked them: the president’s veto, with no government supermajority to override it, has largely stalled re-democratisation.
What the 2023–2026 cases show
Poland in 2023 saw record turnout and an opposition coalition victory, and the Gessler-Sadowski dynamic — backsliding producing counter-mobilisation — operated as predicted. The reversal has since stalled, however: the PiS-backed Karol Nawrocki won the 2025 presidential election, and his veto power, combined with the government’s lack of a supermajority, has produced gridlock rather than full re-democratisation. Slovakia under Fico has abolished the NAKA anti-corruption police, passed a public-broadcasting media law, and moved against NGOs and the right to protest; the counter-mobilisation has been substantial, culminating in the “Chalk Revolution” of November 2025 (around 40,000 protesters in Bratislava, peaking on the anniversary of the Velvet Revolution, 17 November). Hungary — the case where the feedback loop was most deeply entrenched — produced the most dramatic reversal of all: on 12 April 2026, Péter Magyar’s Tisza party won a two-thirds majority (138 of 199 seats on 53.6% of the vote) and Viktor Orbán conceded after sixteen years in power. This is the strongest available confirmation of the Gessler-Wunsch thesis: sustained backsliding built exactly the cross-ideological, pro-democracy coalition the model predicts, and it won at the ballot box.
The psychological underpinnings of democracy that the course has analysed remain present in CEE populations. What varies is whether political entrepreneurs exploit those mechanisms to undermine democracy (Fidesz’s nostalgia-deprivation mobilisation, PiS’s anti-elite framing, Fico’s post-traumatic appeal) or to defend it (Poland’s 2023 pro-EU mobilisation, Slovak civil society’s response to Fico’s reforms, and Hungary’s 2026 opposition victory). The same psychology produces different outcomes depending on framing and institutional channelling, which is the signature claim of the course. Comparative evidence reinforces the more optimistic reading: Guasti and Michal (2025) find that in Central Europe polarisation has mobilised citizens for democracy more strongly than for autocracy — as in the 2023 Czech presidential election — confirming that the counter-mechanisms are not merely hypothetical.
Conclusions
Democracy makes demands on citizens that other forms of government do not: tolerance, participation, and recognition of legitimate opposition. These demands are met only partially in most democracies, and the gap between abstract support for democratic values and applied tolerance is a universal feature that is especially wide in CEE. The determinants of applied tolerance include threat perception, internalised democratic values, personality (openness, dogmatism), education, and experience of democratic institutions. Democratic learning operates: CEE citizens educated under democracy are more tolerant than those educated under less democratic regimes. But backsliding operates in reverse: information environments that portray opposition as illegitimate reduce applied tolerance.
The lecture has argued, consistent with the course’s broader claim, that the same psychological mechanisms drive democratic consolidation and democratic backsliding, and that what differs is framing, institutional channelling, and political leadership. The contest between these outcomes is live in contemporary CEE, and its resolution depends not on any deep psychological deficit in CEE populations but on the institutional and narrative choices of political actors. The Polish 2023 and Hungarian 2026 elections show that the contest can be won by pro-democratic actors, even in the region’s most deeply entrenched case of backsliding. The Polish 2025 presidential result shows that such reversals can themselves stall, and the Slovak and Serbian cases show that the contest can also be lost or remain contested indefinitely.
Social capital
Robert Putnam’s social-capital framework defines social capital as features of social organisation — such as trust, norms, and networks — that can improve the efficiency of society by facilitating coordinated actions. When people are embedded in dense social networks, possess norms of generalised reciprocity, and have high levels of interpersonal trust, they are more likely to be able to organise and act collectively. Places that have higher levels of interpersonal trust and more voluntary associations also have more citizens who concern themselves with public affairs and who believe that others will act fairly and obey the law. In places with lower levels of social capital, citizens are more likely to believe that public affairs are not part of their lives, and corruption is common and expected. Social capital is therefore one of the main mechanisms by which the civic-culture dispositions are transmitted and sustained.