Groups and group dynamics
The Social Psychology of Democracies in Transition
Overview of the lecture
The lecture moves from individual-level social psychology, which has dominated the course so far, to group-level dynamics, where the anti-social and pro-social mechanisms traced in earlier lectures are enacted by parties, parliaments, courts, NGOs, street protests, and other collectivities that populate post-1989 political life. The first half of the lecture sets out the social-psychological theory of groups: how groups are defined, why they form, and how they make decisions well and badly. The second half turns to Natalia Letki’s 2004 study of political involvement in East-Central Europe, which used 1993–1994 World Values Survey data to test what actually predicts democratic participation in the first post-transition decade, and follows the question forward through two decades of subsequent research. Letki’s counterintuitive finding — that former communist-party membership is a positive predictor of post-1989 democratic engagement — becomes the pivot for the lecture’s argument that participation in non-democratic organisations can produce positive social capital for democracy, a claim that inverts the naïve expectation that communist-era institutional legacies would unambiguously obstruct post-communist democratic consolidation.
What is a group?
Definitions
A group is a collection of people who are perceived to belong together and are interdependent. Two criteria for “groupness” co-exist in the social-psychological literature. The first, social integration, emphasises the frequency and depth of interaction among members. The second, entitativity (Campbell 1958), emphasises the extent to which a collection is perceived as a coherent entity on the basis of shared fate, similarity, proximity, or common cause. Entitativity is largely a perceptual property and varies sharply across contexts: a queue at a bus stop has low entitativity even if social integration is technically present, while a platoon of soldiers has high entitativity whether or not the soldiers know each other personally.
Groups are not simply aggregations of individuals. They display emergent dynamics — behaviours that no individual member would display alone. This is why group psychology cannot be reduced to individual psychology, and why the shift from individual to collective analysis is indispensable for understanding post-transition politics.
Types of groups
| Type | Features | Need served | Examples |
|---|---|---|---|
| Intimacy groups | Small, frequent interaction, high similarity | Affection | Family, close friends, cohorts |
| Task groups | Small, high interaction around a task | Achievement | Jury, study group, cabinet meeting |
| Social categories | Large, low interaction, shared identity | Affiliation | Nation, religion, gender, ethnicity |
The same person belongs to multiple groups in all three categories simultaneously. Which group is salient at any moment depends on context, and much of the political mobilisation the course has covered — Solidarity, the 2022 Ukrainian refugee response, anti-PiS civic protests, Fidesz rallies — works by raising the salience of one group identity over others.
Size, composition, and the diversity trade-off
As group size rises, individual participation and commitment decline while tardiness, absenteeism, and turnover rise. Larger groups tend to see more conflict, less cooperation, and poorer coordination. Diversity of composition (along lines of sex, ethnicity, education, background) carries costs and benefits that are not easily separable. Diverse groups experience more interpersonal friction and communication overhead but are typically more flexible, more innovative, and better at external relationships. The CEE transition produced a rapid formation of new organisations whose members were unusually diverse (former dissidents alongside former party members, returnees alongside stayers), and the costs of that diversity were visible quickly while the benefits took a decade or more to materialise.
Why do groups form?
Groups form for a mixture of reasons that usually operate simultaneously. They serve functional purposes (firms, courts, parliaments); they confer survival benefits (kin, clan, nation); they satisfy psychological needs for inclusion, control, and affection (Schutz’s three needs); they provide informational inputs through social comparison; they offer interpersonal support; and they enable collective efficacy — outcomes that can be achieved together but not alone. The informational function is particularly salient in CEE transition contexts, because the 1990s were periods of pervasive information scarcity and opacity, and populations relied on in-group social networks as epistemic anchors when formal institutions could not yet be trusted. The same mechanism, operating in the algorithmically curated information environments of the 2020s, has only intensified (Wojcieszak et al. 2021).
Group decision-making
The promise
Groups are often assumed to make better decisions than individuals do, because they can pool cognitive, informational, and skill resources; because divergent perspectives catch individual errors; and because collective decisions carry more legitimacy than individual ones. When these mechanisms work, groups outperform individuals: juries, well-designed peer review, adversarial litigation, and deliberative mini-publics all realise the promise. When the mechanisms fail, the opposite occurs, and cohesive groups under time pressure with directive leaders can produce decisions that are substantially worse than any of their individual members would have made alone.
Groupthink
Irving Janis introduced the concept of groupthink in 1972 to describe a deteriorated mode of thinking that arises in cohesive in-groups when the members’ pursuit of concurrence overrides realistic appraisal of alternatives. The syndrome has several recognisable symptoms.
| Symptom | What it looks like |
|---|---|
| Pressure to conform | Sanctions for dissent |
| Self-censorship | Members suppress private disagreement |
| Mindguards | Gatekeepers protect the group from disconfirming information |
| Illusion of unanimity | Surface agreement masks private doubt |
| Illusion of invulnerability | Overconfidence in the group’s decisions |
| Illusion of morality | The group’s decisions must be right because we made them |
| Biased out-group perception | Enemies are stupid, weak, or evil |
The empirical fate of Janis’s theory has been mixed. Subsequent research has found only partial support for the specific antecedent-outcome links he proposed (Esser 1998; Rose 2011), particularly the centrality of cohesiveness as a necessary condition. The syndrome remains diagnostically useful, however, and groupthink-style dynamics are visible in several high-profile post-1989 CEE political decisions. The point is not that cohesive groups are always pathological but that cohesion combined with a directive leader and an insulated information environment reliably degrades decision quality.
Group polarisation
A related phenomenon is group polarisation: after discussion, group-average opinion becomes more extreme in the direction of the pre-discussion mean, not more moderate. Three mechanisms account for the effect. Biased information sampling means that members contribute arguments consistent with the group’s leaning, so disconfirming information is systematically under-voiced. Social approval means that members shift toward positions they believe the group values. Perceived group norm means that members overshoot to align with an imagined consensus, which is typically more extreme than the actual average.
Affective polarisation: the contemporary face of the problem
Recent work has shifted the centre of gravity from polarisation over issues to affective polarisation: dislike, distrust, and dehumanisation of partisan out-groups regardless of issue agreement (Iyengar, Lelkes, Levendusky, Malhotra & Westwood 2019). Originally diagnosed in the United States, the pattern is now well documented across European democracies, including in CEE (Reiljan 2020; Wagner 2021; Garzia, Ferreira da Silva & Maye 2023). The mechanisms that drove issue polarisation are amplified in digital information environments, but the empirical picture is more complicated than the popular “echo chamber” framing suggests. Algorithmic curation amplifies in-group content; cross-cutting exposure does not reliably reduce hostility and may increase it (Bail et al. 2018; Wojcieszak et al. 2021). The strongest polarising signal comes not from exposure as such but from exposure framed as conflict by political elites. PiS / anti-PiS, Fidesz / opposition, and SMER / progressive publics are by now classic affective-polarisation cases.
Causes of group conflict
Conflict between groups escalates through several recurring mechanisms. Attribution errors lead members to explain other groups’ behaviour by disposition rather than situation: “they voted that way because they are evil/stupid” rather than “because the incentives their region faced made it rational”. Communication problems (harsh criticism, personality clashes, naysaying) produce revenge cycles that outlive their original causes. Escalation of commitment means that once group members have publicly backed a position, they defend it more strongly under attack, even when evidence accumulates against it. Sunk-cost reasoning, face-saving, and ego-involvement all contribute.
Trust, membership, and democratisation: Letki (2004)
The big question
Letki’s 2004 paper in Political Research Quarterly asks what predicts ordinary citizens’ willingness to engage in politics in the first post-1989 decade. The question matters because democracy requires participants: without citizens who vote, discuss politics, join parties, or join associations, the institutional architecture of democracy floats free of its society. Post-1989 observers feared two outcomes. The pessimistic fear was that CEE publics would be too passive to sustain democracy, because the anomie and disillusionment of transition would suppress participation. The opposing fear was that the habits of pre-1989 politics — protest, street action, confrontation — would destabilise the new democracies by over-supplying participation of the wrong type.
The actual trajectory lay between the two. The years immediately after 1989 saw high turnout, flourishing parties, and widespread civic activity, followed by a rapid decline that by the mid-1990s had produced a heterogeneous landscape: Czechia, Hungary, and early-1990s Bulgaria approached Western European levels of political involvement; Poland, Russia, Ukraine, and several other countries lagged. Letki’s project was to explain the variation.
Hypotheses
| H | Claim |
|---|---|
| 1 | Voluntary-association membership predicts political involvement; types of association vary in effect |
| 2 | Higher interpersonal trust predicts higher political involvement |
| 3 | The combination of trust and membership produces an additional boost |
| 4 | Former communist-party members are more politically involved than the rest of the population |
| 5 | Country-level democratic experience has a positive effect on political involvement |
The fourth hypothesis is the provocative one. The expectation most Western observers would have held in 1993 was the opposite: former communist-party members would be reluctant to participate in post-1989 democracy because they had been pillars of the old regime and were now displaced, resentful, or compromised. Letki’s hypothesis inverts this and predicts that CP-membership socialisation transfers positively to democratic participation.
Data and design
Letki uses World Values Survey data from 1993–1994, covering Belarus, Bulgaria, Czech Republic, Estonia, Hungary, Poland, Romania, Russia, Slovak Republic, and Ukraine. The dependent variable is a political-involvement index (scaled 0–3) that combines three items: political discussion, support for a political party, and party membership. The key independent variables are generalised social trust (a four-item Rosenberg-style scale), voluntary-association membership (grouped into community, professional/lifestyle, and labour categories), self-reported communist-party membership in the past, and country-level democracy score.
Country-level results
Bulgaria (1.11), Hungary (1.05), Slovakia (1.02), and the Czech Republic (1.01) cluster at the top of the index. Romania (0.90), Estonia (0.88), and Belarus (0.84) sit in the middle. Ukraine (0.76), Russia (0.72), and Poland (0.68) anchor the bottom. The pattern does not map neatly onto GDP or onto measured democracy. Poland, which has the lowest involvement score in the sample, also has the strongest post-communist civil-society legacy (Solidarity). The rank order suggests that political involvement is not a straightforward function of material conditions or institutional maturity; it depends on mechanisms Letki’s regression analysis is designed to isolate.
Main findings
Letki’s regression analysis produces five findings. First, community-association membership has the largest effect on political involvement — almost twice as large as the effect of professional or labour-association membership. Second, social trust has a statistically significant but relatively weak effect, much weaker than a simple Putnam framework would predict. Third, former communist-party members are significantly more politically involved than non-members. Fourth, democratic experience (country-level exposure to free elections and competitive parties) is a strong positive predictor. Fifth, the effect of membership and social trust on political involvement is independent of the level of democracy in a country: membership mechanisms work similarly across contexts; what democracy varies is the base rate of involvement, not the returns to involvement.
The substantive effects are large. Letki offers an illustrative comparison: a Czech Republic citizen who had been a communist-party member in the past and was a community-association member at the time of the survey scored 27.4 percent higher on the involvement index than a Russian citizen who had never been involved in either.
Why former communist-party members participate more
Two mechanisms together account for the finding. The first is skills transfer: CP membership taught organisational, persuasive, and meeting-running skills — even if the content of those skills was different. These skills transfer to democratic politics, where the formal machinery of parties, unions, and parliaments operates recognisably similarly. The second is socialisation for participation: membership in any political organisation cultivates the habit of participation, which is in part a habitual behaviour rather than a deliberate choice. The communist party inadvertently built these habits in large segments of CEE populations.
The theoretical implication is that participation in non-democratic organisations can produce positive social capital for democracy. This inverts the pessimistic framing of communist-era institutions as wholly destructive to post-communist democratic prospects. Some legacies of communism are antithetical to democracy (political learned helplessness, low institutional trust, tolerance of corruption); others, unexpectedly, are supportive (skills, habits of organisation, a baseline of political engagement).
Implications for the Putnam framework
Letki’s findings partially validate and partially complicate Robert Putnam’s social-capital framework. They validate the importance of associational membership as a predictor of democratic participation, but they downgrade the importance of generalised interpersonal trust, which Putnam makes central. They complicate the framework further by showing that non-democratic associational membership can positively predict democratic engagement, which the Putnam framework does not easily accommodate. The practical implication is that if a society wants to build democratic engagement, it should invest in associations rather than in diffuse trust-building campaigns, and it should not discount the transferable-skill value of prior organisational experience even in non-democratic contexts.
Updates from twenty years of subsequent research
Howard and the “weak civil society” thesis
Marc Morjé Howard’s The Weakness of Civil Society in Post-Communist Europe (Cambridge UP, 2003) places Letki’s results inside a broader diagnosis: post-communist publics are systematically less associationally engaged than either established democracies or younger Latin American or Asian democracies. Howard identifies three drivers — mistrust of formal organisations as Soviet-era residue, persistence of informal networks that substitute for formal joining, and disappointment with post-1989 outcomes. Letki’s results — modest absolute levels of engagement, with associations doing the heavy lifting within a low ceiling — are consistent with Howard’s diagnosis and refine it by identifying which kinds of organisational experience produce the engagement that does occur.
Pop-Eleches and Tucker, Communism’s Shadow (2017)
Grigore Pop-Eleches and Joshua Tucker’s Communism’s Shadow (Princeton UP, 2017) provides the most systematic test of communist-legacy effects to date. Three findings are particularly relevant to Letki. First, living under communism leaves a persistent negative imprint on interpersonal trust and on participation in formal civic organisations, holding socioeconomic and institutional variables constant. Second, the effects are strongest for those who came of age under late communism and weaker for cohorts socialised before or after, which suggests that socialisation effects fade as the relevant cohorts age out. Third, and most importantly for our reading of Letki, communist exposure and former-regime membership operate differently: passive exposure tends to depress participation, while active CP membership (Letki’s variable) elevates it. Pop-Eleches and Tucker’s results refine Letki rather than overturning her: the legacy is real, persistent, and heterogeneous, and the type and intensity of regime contact matters.
Letki and Mieriņa (2015) on networks and inequality
Letki has continued the line of work, most directly in her co-authored 2015 paper with Inta Mieriņa in Social Science Research, “Getting support in polarized societies: income, social networks, and socioeconomic context”. The paper reframes the social-capital question: it asks not whether networks exist but who benefits from them. The central finding is that in economically polarised societies, the poor maintain larger informal support networks than the wealthy, yet extract less support from them, because polarisation traps low-income people in networks of similarly resource-poor others. The wealthy use smaller, more selective networks to reinforce their advantages.
This complicates the 2004 picture. The ealier paper showed that associational membership predicts political engagement and that interpersonal trust is weaker than Putnam predicts. The 2015 paper adds that informal networks — the alternative to formal associational participation — are not a neutral substitute. Under inequality, informal networks become regressive: the people who most need them to compensate for institutional weakness or for limited access to formal organisations gain the least from them. The CEE relevance is direct, because post-communist societies combine relatively low associational membership with high economic inequality, and the regressive-network effect bites hardest precisely where Letki (2004) located the engagement deficit.
Protest and networked mobilisation
The classic Putnam/Letki measures (party membership, association membership) under-count the protest repertoire that has dominated CEE politics in the 2010s and 2020s. Several mobilisations exemplify the channel. The Polish Czarny Protest (2016) and Women’s Strike (2020) brought hundreds of thousands of women into public political life around reproductive-rights issues (Korolczuk & Graff 2018). The Polish KOD mobilisation of 2015–2017 organised middle-class anti-PiS protest at scale (Ekiert & Kubik 2017). The Slovak For a Decent Slovakia movement, following the Kuciak murder in 2018, brought down a Smer government. Recurrent Hungarian opposition mobilisations have continued through the 2010s and into the 2020s, even if without comparable institutional results (Greskovits 2020). Romanian anti-corruption protests of 2017–2019 were among the largest sustained mobilisations in the country since 1989. These movements pulled in many first-time participants — people without prior associational membership — which a strict reading of Letki’s design would mark as “low engagement” although the participation was substantial.
Digital information and group polarisation
Wojcieszak and colleagues’ work on selective exposure (2021) shows that echo-chamber effects are real but smaller than popular discourse assumes, and that the strongest polarising effect comes not from exposure but from exposure framed as conflict by political elites. For CEE, this means elite framing — Kaczyński’s “they / us”, Orbán’s nation-vs-Brussels rhetoric, Fico’s media populism — is doing more polarising work than algorithms alone, even if the algorithms are amplifying it.
What this implies for democratic backsliding today
Letki’s findings were from 1993–1994 data, when democratic institutions were new and still being built. The mechanism she identifies — that exposure to democratic processes produces higher individual-level engagement — can run in reverse. When democratic institutions are systematically undermined, as they have been in Hungary since 2010, in Poland between 2015 and 2023, and more recently in Slovakia, we should expect engagement through formal channels to decline. But the post-Letki literature shows that the same mechanism can be reactivated under specific conditions: when backsliding becomes an organising political cleavage, opposition mobilisation can re-forge participation, often through the protest and networked-mobilisation channels above rather than through associational membership in the original Letki sense. The mechanism that built engagement in the 1990s can therefore unwind it under sustained backsliding — but it can also rebuild it under sustained counter-mobilisation.
The theories in parallel
| Theory | What Letki’s evidence says |
|---|---|
| Social capital (Putnam) | Partially supported. Associational membership works; interpersonal trust works weakly |
| Social learning (Bandura) | Supported. Skills from CP membership transfer to democratic contexts |
| System justification | Partially relevant. Former CP members could have defended the old system but instead channelled skills into the new one |
| Relative deprivation | Partially relevant. Disappointment suppresses engagement in some countries, not others |
| Collective memory | Supported. Solidarity, Charter 77 and similar prior mobilisations created participation templates |
Conclusions
Groups are not aggregated individuals. They display emergent dynamics — groupthink, polarisation, escalation, in-group bias — that are potential pathologies, alongside pooled resources, shared identity, and mutual learning that are potential assets. The same social psychology that drives pathological group dynamics also drives pro-social collective action, and the institutional and narrative choices that frame group activity determine which outcome predominates.
Letki’s specific empirical findings on post-1989 political participation reshape the standard story. Civic engagement in CEE is predicted most strongly by associational membership (community associations above all), by country-level democratic experience, and — counter-intuitively — by prior communist-party membership. Generalised interpersonal trust matters less than the Putnam framework predicts. Twenty years of subsequent research deepen rather than overturn these findings: legacies are persistent but heterogeneous (Pop-Eleches & Tucker); informal-network substitution becomes important when institutions weaken (Letki & Mieriņa); participation has migrated toward protest and networked mobilisation that the original design did not capture; and affective polarisation has replaced apathy as the central pathology. The problem of post-communist democratic consolidation was not that citizens lacked participation skills; it was that the institutions channelling those skills were thin, unevenly developed, and, over subsequent decades, in some countries actively eroded.
Questions for discussion
- If former communist-party members transfer their organisational skills into democratic participation, what does this imply about the moral framing of “collaboration” versus “cooperation” in the transition period? Is the distinction between regime complicity and civic competence analytically clean, or does Letki’s finding suggest that it blurs?
- Affective polarisation has replaced apathy as the central pathology of CEE democracy. Can democratic CEE institutions be designed to attenuate it, or is within-camp polarisation now a structural feature of mediated mass politics that institutions can only blunt?
- Letki’s findings are from 1993–1994 data. Thirty years on, how much of the picture still holds, and where should we expect the biggest departures? In particular, what does the ageing-out of the CP-membership cohort imply for the continuing applicability of her mechanism?
- Groupthink warns against cohesive decision-making, but some CEE challenges (defending independent media, coordinating judicial reform, EU-level bargaining, sustaining protest mobilisations) reward cohesion. When is cohesion an asset and when is it a liability, and how should democratic institutions distinguish between the two cases?