Infringement proceedings serve as a key legal mechanism through which the European Union (EU) enforces compliance with common laws among its member states. This article, for the first time, examines how the nationality and partisanship of EU officials influence outcomes of these proceedings. We argue that Commissioners and Director Generals face a principal–agent dilemma, as they simultaneously act as EU officials and as nationals of their home countries.

We expect that infringement proceedings are less likely to escalate when they are overseen by EU of- ficials who share the nationality of the member state under investigation. We further posit that officials’ partisan affiliations and ideological leaning may shape the outcomes of such cases. Using a novel dataset combining 19,352 infringement proceedings between 2002-2020 with information about the responsible officials, we find evidence of a nationality bias, however, only among Commissioners. By contrast, we find no indication of bias linked to ideology or partisanship.

Understanding how individuals perceive their social class is central to political behavior research, yet subjective and objective indicators often diverge. Using German survey data, we examine the structure of that divergence by explicitly modeling differences in the error variance when predicting subjective class perceptions with three standard objective class indicators – education, income, and occupational prestige. Employing a heteroskedastic maximum likelihood estimator, we show that these objective class markers predict subjective class more precisely among individuals scoring highly on them and those satisfied with their position in life. Conversely, citizens from objectively lower classes have more dispersed subjective perceptions. This carries broad implications for research on, for example, class awareness, class mobility, and class-based appeals. Moreover, the analysis cautions against treating subjective and objective class as interchangeable and illustrates the usefulness of heteroskedastic regression, which can be easily adopted to other empirical settings.

Does growing climate concern endanger support for democracy? Climate change is a highly salient issue and concern about its consequences is growing. Therefore, I expect climate concern to increasingly factor into citizens’ evaluation of their democratic system. Building on prominent conceptualizations by Easton and Norris, these evaluations can range from being very specific (e.g. trust in the current government or satisfaction with democracy) to very diffuse (e.g. support for democracy and democratic principles).

Using original German panel data (2021-2023; N > 10.000) and random-effects-within-between models, I provide a novel empirical investigation into the causal impact of growing climate concern on system evaluations on different levels. Moreover, through mediation analyses, I am also able to evaluate whether negative impacts indeed follow the causal chain from specific to diffuse evaluations. All hypotheses and analysis steps are pre-registered.

Finally, I test multiple possible conditionalities: The described mechanism should apply more strongly for citizens with a higher climate change salience, because saliency moderates whether an issue factors into citizens’ overall system evaluation. Contrary, feeling that one’s climate position is well represented by parties and higher political efficacy should attenuate the negative impact on their system evaluation.

For one, my contribution provides a valuable empirical test of well-established theories in democracy research within a novel issue dimension. Moreover, climate change salience and concern are likely to rise even further in the future. Thus, understanding the threat for core components of democratic systems posed by this concern also carries important implications beyond the scientific community.

A growing number of citizens says that they are no longer free to speak their mind. This raises an important question in the study of democracy. What sanctions of free expression exist in a liberal democracy, how probable are they and what costs do they entail? And how accurate are citizens’ perceptions?

We model the degree of free speech as the product of the probability of being sanctioned for speech and the associated costs of those sanctions. Using multiple original surveys from Germany, we relate this heuristic model to empirical data and present three studies on citizens’ perceptions of free speech.

Study 1 shows that citizens distinguish regular discursive risks from manifest negative consequences. While the former are perceived as much more probable, both carry comparable costs for free speech. Study 2 evaluates which topics are sanctioned and by whom. Exploiting randomized issue primes we find few differences in reported sanction probabilities. However, citizens do not feel restricted by close social ties but by more distant democratic actors. Last, study 3 demonstrates that citizens’ expectations of being sanctioned for speaking freely are to a substantial degree rooted in actual experiences. Taken together, these findings have significant implications for our understanding of liberal democracy and its vulnerabilities.

Does political polarization harm citizens’ subjective sense of free speech? Seminal accounts in political science such as the ‘spiral of silence’, conformity pressures in social networks, or online echo-chambers point to social in-group mechanisms that can restrict citizens’ free expression. We offer a novel alternative argument that focuses on a different aspect of political polarization. We decompose polarization into two parts: in-group attachment and out-group aversion.

Specifically, we argue that out-group aversion reduces normative constraints on citizens’ expression and therefore has a liberating effect on subjective free speech. We test our argument using data from two original surveys fielded in Germany and including a new measure of out-group aversion. The results broadly support our argument and suggest that out-group aversion not only significantly increases subjective freedom of speech, but is of far greater relevance than contrasting in-group mechanisms. This finding has important implications for our understanding of political polarization, public opinion and democratic culture.