Areas of Specialization

Working Papers

Panic at the Courtroom: Can Legislative Action against Discrimination in Court Reduce Violence?
Draft available upon request

Abstract: Can legislative action against discrimination in court reduce violence? In recent years, we have seen more and more legislation trying to address emerging regulatory gaps in the protection of gender identity and sexual orientation rights. While previous research documents great benefits of these policies when targeted at the LGBTQ+ population, much less is known about perpetrator-centered approaches. I examine whether the prohibition of the LGBTQ+ “panic” defense, a legal strategy that seeks to blame the victim’s sexual orientation or gender identity for the defendant’s violent behavior, affects the prevalence of hate crimes. To identify the effect of legislative interventions banning this practice, I leverage the gradual roll-out of LGBTQ+ “panic” defense bans across the US and implement a difference-in-differences design. The results show that introducing LGBTQ+ “panic” defense bans lead to significantly lower levels of anti-LGBTQ+ hate crime rates. Besides contributing to the literature on minority politics, policy feedback and hate crime prevention, insights gained from this work are of assistance to policy makers and minority rights advocates.

Changing the Rules of the Game: Populists in power and constitutional retrogression
Draft available upon request

Abstract: As populists rise from periphery to power, the question of whether and how these actors threaten democratic institutions becomes ever more pressing. While previous research has focused primarily on the determinants of populist electoral success and its consequences for voters, established parties, and overall democratic quality, this article examines the impact of populist parties’ government participation on the most fundamental of democratic institutions, the constitution. Drawing on a comprehensive dataset, covering information on constitutional changes and populist parties across 30 European democracies (1989-2019), I investigate how the entrance of these actors into power impacts the content of constitutional redrafting. Using a series of matching and difference-in-differences models, I find no support for the notion that populists in power are going after constitutionally enshrined horizontal accountability. Neither de jure executive power nor de jure judicial independence significantly shift after populists enter the government. The same results hold when accounting for the level of constitutional rigidity. The findings make an important contribution to the literature on populist parties, constitutional change and democratic backsliding.

Does protest affect bystanders? experimental evidence
– with Daniel Bischof, Ferdinand Geissler, Johannes Giesecke, Felix Hartmann, Macartan Humphreys, Heike Klüver, Lukas Stoetzer, and Tim Wappenhans –

Abstract: A rich body of observational work documents effects of protest on various political outcomes such as election results, policy change, public opinion and political behavior. Yet, the mechanisms behind these audience effects are largely unknown. We argue that observing political protest functions as a strong and highly visible cue that conveys summary information about the distribution of attitudes and behaviors in society, creating the impression of a swaying public opinion. To test this argument, we conduct a field experiment in the context of a large-scale climate strike organized by Fridays for Future (FFF). We recruited citizens in the city of Berlin and randomly allocated them to a treatment (exposed to the protest) and a control condition (not exposed to the protest). By manipulating participants’ exposure to the FFF protest, we test whether bystanders – individuals unintentionally drawn into the role of observers – update their norm perceptions, environmental attitudes and behaviors in response to witnessing the climate strike. A few weeks after the protests, we conduct a follow-up survey among participants to measure long-term effects. We find that while bystanders adapt their social norm perceptions and behaviors, attitudes remain initially unchanged. These findings fill an important gap in the literature by unveiling the mechanisms through which protests translate into shifts in political outcomes.

The Police as Gatekeepers of Information
– with Ashrakat Elshehawy, Arun Frey, Tobias Roemer, Sascha Riaz –

Abstract:
We use large language models to quantitatively examine the role of the police in shaping public opinion about immigration and crime. While a large and growing literature examines the political biases of media reporting and its effects on political attitudes, this line of research generally overlooks the critical role of the police as primary information gatekeepers in the context of crime. The police, as the initial contact in crime incidents and reporting, have the ability to mold media narratives by choosing which crimes to report and whether to release information about perpetrator characteristics such as nationality or ethnicity. The police are thus causally prior to the news media in the sense that the news media can generally only report information that the police have chosen to disclose. We systematically examine the role of the German police and its reporting behavior using big textual data on more than one million police press statements released between 2013 and 2019. We use large language models that excel at their contextual and semantic understanding to uncover bias in the nature of crime reporting on perpetrators’ characteristics. We focus on extracting explicit and implicit indicators of perpetrator ethnicity. We compare our results across traditional NLP techniques and different LLM outputs, particularly Llama, Gemini Pro, ChatGPT 3.5 and ChatGPT 4. We further evaluate our findings against human coders. In a second step, we use a regression discontinuity approach to (i) test whether the police act strategically in what information it chooses to release to the news media prior to elections and (ii) assess the effectiveness of such strategies in influencing electoral outcomes. This research will have important implications for the study of immigration attitudes by highlighting the interdependence between the news media and the police in shaping public discourse.

The German Hate Crime Dictionary (GHCD): Measuring Spatial and Temporal Trends in Hate Crimes using
Quantitative Text Analysis
Draft available upon request

Abstract: This paper proposes a new technique for measuring spatial and temporal trends in hate crime incidents based on quantitative text analysis. Applied to a novel and unique data set of over 11,000 police breaking news published between 2015 and 2019, the GHCD proves to be a fast, reliable and valid instrument for detecting descriptions of hate crimes in textual data. The GHCD thus provides policy makers and researchers with an effective tool to assist in the development and evaluation of geographically targeted interventions and the investigation of causes and effects of prejudice motivated crimes.

When and why do Parties Respond to Protest? Parties’ Responsiveness to Environmental Protests
– with Jóhanna Bjarndóttir, and António Valentim –

Abstract: When do parties respond to protests? A growing body of literature studies the extent to which elites respond to protests, finding mixed results. This paper explores the conditions under which political parties are more likely to respond to protests. We do so by exploring three key expectations in the literature: i. the role of being in power or opposition, ii. that of party ideology and alignment with the protest and iii. that of electoral threat. We test for these using difference-in-differences and regression-discontinuity in time designs on geocoded data on environmental protests, which we combine with party press from 2018-2021. We explore variation in state-level cabinet status in Germany to explore the role of parties’ cabinet status and use state-of-the-art quantitative text analysis to explore the mechanisms of when and how parties respond to protests. Overall, this study’s findings contribute to our understanding of not just whether but when and how parties respond to protests.

Selected Work in Progress

Coalition governments and party competition: The electoral implications of ministerial portfolios
– with Heike Klüver, and Jae-Jae Spoon –

The Electoral Effects of State-Sponsored Homophobia
– with Konstantin Bogatyrev, Tarik Abou-Chadi, Lukas Stoetzer, and Heike Klüver –