Citizens’ commitment to democratic norms is often conditional and instrumental. Especially under conditions of heightened polarization, many are willing to bypass norms for preferred political outcomes. I study this dynamic in a crucial, ideologically and affectively polarized, yet understudied case: climate change. Using original survey data from Germany (N = 2,101), I show that stronger preferences – whether for or against climate action – are robustly correlated with support for abstract and specific non-democratic governance aiding those preferences. This includes, for example, restricting climate activism, ignoring pro-climate action court rulings, and censoring climate change denial. Addressing concerns about ‘staged’ democrats, I additionally use two original list experiments to uncover support for non-democratic governance that respondents might hide due to strong social norms for democracy. The experiments reveal ‘true’ support for undemocratic climate action (prevention) by about 35 (39)% of citizens. Especially climate action skeptics tend not to disclose their non-democratic attitudes openly. This study thus contributes empirical insights to three closely related debates: on democratic resilience under polarization, on the political conflict over climate action, and on the concealment of undemocratic preferences. I show that policy attitudes, irrespective of their direction, can motivate instrumental norm trade-offs on both sides of a strongly polarized cleavage, offer the first joint account of both sides of the climate divide’s democratic commitment, and provide new evidence for ‘staged democrats’ who conceal authoritarian preferences behind public support for democracy. Recognizing these dynamics is crucial for protecting democratic commitments amid intensifying polarized conflict.

In this paper we demonstrate that political parties in Germany systematically alter their issue emphasis in easy-read manifestos (ERMs) compared to regular manifestos. We argue parties face incentives to target ERMs’ distinct audience – citizens with disabilities and those struggling with complex language – by emphasizing issues most salient to this electorate. Applying natural language processing on 74 ERMs between 2009 and 2025 from five parties we find substantial distortions in communicated political priorities. Welfare-related issues are, on average, overemphasized by 9 percentage points, while economic issues are de-emphasized. These differences are generally stronger among left-leaning parties. Furthermore, parties address ERMs’ target audience with group-based appeals nearly four times more frequently in ERMs than in their regular manifestos. We corroborate parties’ contribution to this mismatch through qualitative questionnaires sent to easy-language translation agencies, which confirm that parties are directly and extensively involved in translating ERMs. Our study adds to the literature on strategic issue emphasis by identifying targeting behavior in a previously unstudied form of communication. We also show that NLP methods can be applied to less prominent text types like easy-read language. Finally, our findings raise important questions about political inclusivity, as targeting may compromise ERMs’ role as neutral empowerment tools.

Can a financial scandal in another country shake citizens’ political support at home? We examine the 2014 LuxLeaks, which exposed secret tax deals between Luxembourg and more than 340 multinational firms. The leak broke during fieldwork for Round 7 of the European Social Survey, creating a natural experiment in Germany, the Netherlands, Belgium, and Switzerland, which we exploit using an Unexpected Event during Survey Design. Its peculiarities make this an instructive case: the deals were legal, and only directly implicated Luxembourgish politicians. However, the involvement of firms from 82 countries and Jean-Claude Juncker, who had become European Commission President days earlier, linked the scandal to politics across Europe. Pooled results show that the disclosure significantly lowered satisfaction with democracy (SWD). Our finding holds across control strategies, estimators, and bandwidths. Robustness checks speak against time trends, cyclical patterns, and a general shift in public mood. Respondents in Germany, whose major firms featured prominently in the leaks, drive the decline in SWD and simultaneously became less satisfied with their economy. Dutch respondents show a smaller, non-significant drop in SWD but a sharper loss of trust in national political institutions. Belgian and Swiss respondents barely reacted. Our findings overall support dysfunctionalist accounts of scandal: When a scandal is salient and can be tied to domestic institutions or actors, even legal practices exposed abroad can causally erode political support. The magnitude and shape of the backlash depend on context, such as involvement of domestic firms, media reporting, and how national governments respond.

Concern about climate change has risen sharply across Europe, yet its democratic implications remain understudied. Drawing on the idea of a state–citizen contract that extends beyond material prosperity to citizens’ well-being, we argue that rising climate concern could negatively impact citizens’ satisfaction with democracy (SWD). Cross-sectional data, however, may mask this relationship, as climate concern is most prevalent among groups that are more satisfied with democracy to begin with. Using four waves of the German Socio-Economic Panel (2010–2021; 65,409 observations) and random-effects-within-between models, we uncover exactly this divergence. Between individuals, higher average climate concern is associated with higher SWD. Within individuals, by contrast, increases in climate concern are robustly associated with declines in SWD. By 2020/21, this within-person association exceeds that of financial concern and approaches those of concern about immigration and crime. It is most pronounced where climate change is arguably most salient: in the most recent data, among post-materialists, and among stable Green partisans. Our analysis thus recasts climate concern as a previously overlooked correlate of democratic dissatisfaction, hidden in cross-sectional data and concentrated among citizens whom democracies could otherwise count among their strongest supporters. As climate impacts intensify, unaddressed concern may become a growing challenge for democratic support, making credible government climate action not only an environmental but also a democratic imperative.

Democratic backsliding has reignited debates over militant democracy, i.e. pre-emptive measures to keep undemocratic actors out of power. While existing research examines the design and effectiveness of measures such as proscription and employment bans, we know little about citizens’ support for such measures. Yet democratic resilience depends not only on institutions and elites but also on citizens’ willingness to defend democracy. We make two arguments. First, when state authorities officially designate political organisations as extremist, this acts as an institutional cue that increases support for pre-emptive measures, particularly among citizens with high levels of institutional trust. Second, support is not always principled but may instead reflect instrumental reasoning, with citizens more willing to endorse restrictions against ideological opponents than against ideologically proximate actors. We evaluate these claims with data from an original survey of ≈ 4,000 respondents in Germany. We also introduce and validate the first measure of citizens’ support for militant democracy. We show that support for pre-emptive measures reflects both institutional legitimacy and ideological bias: official extremist designations correlate with support, particularly among high-trust citizens, while individuals remain more willing to target ideological opponents than allied groups. Our findings suggest that the legitimacy of militant democracy ultimately depends on the credibility of the institutions implementing it.

Many rural areas in developed democracies face a surplus of unattached men, as women of reproductive age have disproportionately moved away. We show that this demographic imbalance is positively associated with radical right vote shares in Europe and in Germany and provide individual-level evidence for a mechanism beyond composition. We argue that difficulties finding a partner during family-formation years shift men’s political behavior toward the radical right. Survey data from Germany show that a higher municipal male surplus predicts greater support for the radical right AfD among men but not women. Moreover, men respond specifically to gender imbalance within their age group. These effects concentrate among those aged 18–44. Further, those who report greater difficulty finding a partner are more likely to hold xenophobic attitudes and support the AfD. Gender imbalances may thus reshape politics by altering men’s political preferences, not just their share of the electorate.

Social class is central to contemporary debates on political identity and attitude formation in Germany. But objective indicators of class often diverge from subjective perceptions of class membership. We examine this divergence and possible consequences by modeling differences in error variance when predicting subjective class perceptions and left-right self-placement from three standard objective class indicators – education, income, and occupational prestige – through a heteroskedastic maximum likelihood estimator. Using cross-sectional survey data from Germany (2021/2022; N = 8,328), we first show that these markers predict subjective class more precisely among individuals with a higher education, those with higher life satisfaction and a sense of receiving their fair share in life. This pattern further extends to political attitudes: objective class is similarly less precisely predictive of left-right self-placement among citizens with lower incomes and those dissatisfied with what they receive in life. Together, these results speak towards obstacles of mobilizing dissatisfied lower-class citizens around shared material interests: It is exactly those citizens for whom objective material circumstances translate least coherently into shared political attitudes and class identities. Additionally, citizens embedded in homogeneous networks show a noisier link between class position and political stance. Those with more diverse social contacts instead possibly benefit from cross-cutting exposure that keeps political views more closely calibrated to material circumstance. More broadly, our analysis further suggests that treating subjective and objective class as interchangeable introduces systematic modeling error, and illustrates the usefulness of treating heteroskedasticity as informative rather than a nuisance for political science research in general.