working papers
conditionally accepted
-
Coal Rush: The Built Legacy of the Industrial Revolution and the Rise of the Radical RightNils Blossey, Lukas Haffert, and Lukas F Stoetzer2025Populism & Radical RightHistorical industrial centers have shifted to the right but have done so at different speeds and intensities. We argue that this variation can be explained by differences in the historical industrialization process. Communities that industrialized later and more intensively realign more toward the radical right today. This is because the built environment shaped by the original industrialization drives demographic persistence and neighborhood disadvantage. To examine our argument, we study the effects of nineteenth-century coal mining in Germany’s Ruhr area. We match the geolocation of over 1,000 mining shafts, historical plant-level employment data, and the spread of company housing with contemporary electoral results at the neighborhood level. For identification, we exploit the depth of coal deposits that governed the adoption of deep-shaft mining. The findings demonstrate how the path of economic development influences voting in the long run.
@techreport{blossey_haffert_stoetzer_2025_coal_rush, bibtex_show = {true}, status = {condaccept}, journal = {American Journal of Political Science}, title = {Coal Rush: The Built Legacy of the Industrial Revolution and the Rise of the Radical Right}, keywords = {Populism \& Radical Right}, author = {Blossey, Nils and Haffert, Lukas and Stoetzer, Lukas F}, year = {2025}, impact = {1}, preview = {wp_placeholder.png}, institution = {OSF Preprints, Center for Open Science}, type = {Preprint}, number = {tswm7\_v1}, doi = {10.31219/osf.io/tswm7\_v1}, url = {https://osf.io/preprints/socarxiv/tswm7\_v1} }
revise and resubmit
-
Emphasizing owned issues or riding the wave? An experimental investigation of effective issue emphasize strategiesLukas F Stoetzer, Heike Klüver, and Petra SchleiterAmerican European Political Science Association Conference 2022. 2022Voting DecisionsParty CompetitionThe selective emphasis of political issues is a key strategy for winning electoral support during campaigns. We present experimental evidence that an issue ownership strategy, which emphasizes parties’ perceived competence advantages, is more effective than a “riding the wave” strategy, which focuses on issues currently salient to voters. Using panel survey data from the 2021 German federal election, we show that the Christian Democrats could have gained support by highlighting their perceived competence on the economy. In contrast, ’riding the wave’ and focusing their campaign on the environment would likely not have increased their vote share.
@article{stoetzer2022, bibtex_show = {true}, status = {rar}, year = {2022}, title = {Emphasizing owned issues or riding the wave? An experimental investigation of effective issue emphasize strategies}, keywords = {Voting Decisions, Party Competition}, author = {Stoetzer, Lukas F and Klüver, Heike and Schleiter, Petra}, journal = {American European Political Science Association Conference 2022.}, impact = {2}, preview = {wp_placeholder.png} } -
Breaking Pre-electoral Coalition Commitments: Electoral Accountability and Satisfaction with DemocracyLukas F. Stoetzer, and Oke Bahnsen2026Coalition PoliticsRepresentationPre-electoral coalition signals are widely seen as essential for democratic accountability in proportional representation systems, as they help voters anticipate post-electoral governments. Among these, pre-electoral coalition commitments to enter or rule out specific coalitions provide particularly strong guidance to voters. However, we know little about how citizens react when parties violate such commitments. We argue that while electoral punishment for breaking coalition commitments is context-dependent, such violations reduce satisfaction with democracy more broadly. We test this claim using a quasi-experimental analysis of the 1996 New Zealand election and a survey experiment conducted during the 2020 Irish government formation process. We find that violations of coalition commitments consistently reduce democratic satisfaction, whereas electoral punishment of the violating party appears only in the survey experiment and depends on voters’ preferences over the coalition outcome. These results show how violations of informal institutional commitments can undermine democratic support even in the absence of strong electoral sanctions.
under review
-
How to Break Populist Parties’ Appeal? An Experimental Evaluation of Communication Counter-StrategiesHeike Klüver, Ferdinand Geissler, Felix Hartmann, Johannes Giesecke, Lukas F. Stoetzer, and Petra Schleiter2024Populism & Radical RightParty CompetitionRight-wing populist parties have been on the rise across Europe, but there is limited evidence on how mainstream parties can effectively counter this appeal. This study experimentally evaluates four communication counter-strategies—highlighting mainstream parties’ representation of citizen interests, emphasizing their performance, exposing populists’ self-serving motivations, and promoting alternative identities. We fielded a large-scale survey experiment among over 24,000 respondents in Germany using real-world messages drawn from social media, party manifestos, and press releases. The results indicate that exposing populist parties as self-serving is an effective counter-strategy, especially among voters initially supportive of right-wing populists, while other strategies show more limited effects. These findings contribute to understanding how political communication can mitigate the appeal of populist parties.
-
The Effect of Political Arguments on Voting DecisionsLukas F Stoetzer, and Denise TraberSwiss Political Science Association (SPSA) Annual Conference in Zurich 2019 and Midwest Political Science Association conference in Chicago 2019. 2019Voting DecisionsThe question of what role political arguments play in democracy has taken centre stage in times of populism, post-truth discourse, and identity politics. It is relevant to analyse how political arguments affect policy preferences, but also how these effects carry over to the election of representatives. In this article, we test the persuasive effect of policy arguments on voting decisions. We combine a persuasion experiment using policy arguments with a conjoint experiment, in which voters choose between two candidates who differ in various policy positions. We conduct the experiment with two representative samples of the German voting-age population in June 2021 and August 2022 before the Election. In addition to testing the downstream effects of arguments on candidate choice, we distinguish between a persuasive effect of policy arguments (i.e. they affect vote decisions by changing the direction of policy preferences) and a priming effect of arguments (i.e. they affect vote decisions by changing the salience of policy preferences). The findings show that persuasion plays a more central role in how arguments shape voters’ decision making. The results have a multitude of implications for the role of political arguments in Democracies, political campaigns, and party competition.
@article{stoetzer2019vote, bibtex_show = {true}, title = {The Effect of Political Arguments on Voting Decisions}, keywords = {Voting Decisions}, author = {Stoetzer, Lukas F and Traber, Denise}, year = {2019}, impact = {2}, status = {review}, journal = {Swiss Political Science Association (SPSA) Annual Conference in Zurich 2019 and Midwest Political Science Association conference in Chicago 2019.}, preview = {wp_placeholder.png} } -
Beyond Persuasion: Protest’s Direct Behavioral Impact on BystandersVioleta Haas, Tim Wappenhans, Ferdinand Geißler, Felix Hartmann, Daniel Bischof, Johannes Giesecke, Macartan Humphreys, Heike Klüver, and Lukas F. Stoetzer2025Despite decades of scholarship on protest effects, we know little about how bystanders – citizens who merely observe protests without participating – are affected by them. Understanding the impact of protest on bystanders is crucial because their latent support, normative beliefs, and concrete actions can make or break a movement’s broader societal impact. This study addresses this gap with a novel field experiment in Berlin, Germany, in which pedestrians were randomly routed either past (treatment) or away from (control) three large-scale Fridays for Future climate strikes. A one-month follow-up survey assessed the persistence of effects. We find no detectable impact on climate attitudes, vote intentions, or norm perceptions but a substantial increase in immediate donations to climate causes. These results suggest that protest is more likely to influence bystanders through immediate behavioral cues than through changes in attitudes or norms. The findings challenge the prevailing assumption that protests reshape public opinion directly and call for renewed theorizing that centers on observers’ immediate behavioral activation rather than just opinion change mechanisms.
-
Visual presentation raises trust in election pollsElias Koch, Simon Munzert, Linus Hagemann, Hannah Rajski, Thomas Gschwend, Cornelius Erfort, and Lukas F. Stoetzer2026Election ForecastingOpinion polls are central to voters’ understanding of electoral competition, shaping expectations, strategic choices, and turnout. But how the news media present polls varies widely, and whether that presentation shapes public perception is largely unknown. We conducted a visual conjoint experiment with 23,444 participants during the 2025 German federal election campaign, randomly varying the visual format and technical content of the same poll report across 160 versions. Visual formats, including bar charts, dot plots, and line graphs, were consistently perceived as more appealing and trustworthy than text-only reports. The trust gains were largest among participants with lower data and visualization literacy and less accurate prior beliefs about the poll standings, whereas the appeal of visualization was greatest among the most literate. Technical features mattered less and not always favorably: disclosing statistical uncertainty reduced the perceived attractiveness of a report, while reference values shifted trust depending on the benchmark used. These findings suggest that visual reporting holds consistent advantages over the text-only format that still dominates poll reporting in the news media, particularly for audiences who struggle most to interpret quantitative information.
-
Stacking Election ForecastsLukas F. Stoetzer, and Cornelius Erfort2026Election ForecastingMethodsAccurately predicting election outcomes is challenging due to model uncertainty, limited data availability, and the sequential nature of electoral prediction. While election forecasting increasingly relies on diverse models that combine different theories, predictors, and data sources, forecasts are still commonly treated as a competition among individual models rather than combined into ensemble forecasts. We propose the use of stacking weights, computed from leave-future-out predictive performance, to systematically combine forecasts from different election forecasting models. Our framework accounts for model uncertainty, the complete predictive distribution and weights forecasts according to their past predictive performance. We demonstrate the approach using applications to U.S. presidential elections and German federal elections. The results show that stacked forecasts improve on individual and alternative ensemble models in terms of predictive density of the realized outcomes.
in preparation
-
Measuring Attribute Relevance in Conjoint AnalysisLukas F. Stoetzer, and Asya Magazinnik2026MethodsResearchers often rely on conjoint analysis to examine which attributes are most relevant for selecting a profile. However, standard estimands are not well suited to assess the relevance of attributes. In this paper, we introduce a novel approach to measuring attribute relevance in conjoint analyses by drawing on causal attribution methods. We propose the Average Probability of Necessary and Sufficient conditions (APNS) as a new estimand that captures the probability that a change in an attribute level is both necessary and sufficient for a decision. We show that under a set of assumptions related to monotonicity and separability, the APNS can be identified nonparametrically as a function of the widely used Average Marginal Component Effect (AMCE). Using simulations and applications to two conjoint experiments, we demonstrate how the APNS provides a nuanced measure of attribute relevance in decision-making. The APNS is highly correlated with respondents’ subjective issue importance measures, supporting its validity as a tool for examining the role of attribute relevance in choice behavior.
-
Precision-Weighted Aggregation of Citizen Election ForecastsLukas F. Stoetzer, and Hannah Rajski2026Election ForecastingMethodsCitizen forecasts aggregate individuals’ electoral expectations to predict election outcomes, typically weighting all respondents equally and implicitly assuming homogeneous forecast quality. Yet forecasting accuracy varies across individuals, suggesting that weighting more capable forecasters more heavily could improve performance. Existing weighting approaches, however, rely on information about forecasters, such as political knowledge or past accuracy, and therefore require historical data. We propose an elicitation-based approach that weights forecasts by respondents’ subjective precision. We derive oracle weights that minimize the mean squared error of weighted vote-share forecasts and show that a set of feasible precision-based weights can approximate them. To estimate precision directly, we introduce Manski-style elicitation questions that collect both expected vote shares and subjective uncertainty assessments. Using original pre-election survey data from approximately 14,600 eligible voters in the 2025 German federal election, we forecast seven parties’ vote shares nationally and across 299 electoral districts. Nationally, precision weighting reduces RMSE from 1.39 to 0.95 percentage points. In district-level forecasts, additional corrections for overconfidence yield the largest improvements, although gains are more modest.
-
Coordination Spirals in Proportional ElectionsLukas F. Stoetzer, Elias Koch, and Thomas Gschwend2026Coalition PoliticsRepresentationFragmentation in proportional representation (PR) systems is often seen as a source of difficulties for government formation and electoral accountability. What can hold these party systems together? We argue that PR also contains an inherent coordination mechanism that concentrates competition on the two leading parties. When polls signal a close race at the top, and both parties hold balanced potentials in the electorate, voters shift support away from smaller parties toward the frontrunners, thereby reducing party system fragmentation. We show formally that when voters derive utility from which party wins, equilibrium behavior exhibits a coordination spiral in close contests: strategic voting feeds back into greater concentration on the top two. We test this argument in two studies. First, linking pre-election polls to outcomes across 320 PR elections, we find that the combined top-two vote share rises by up to 10 percentage points in close races, with effects emerging roughly two to four weeks before election day. Second, panel data from the 2021 German campaign — a close and balanced race — show that the same dynamic operates at the individual level. Together, these findings reveal a previously underappreciated stabilizing force that arises endogenously from electoral competition in proportional systems.
-
Pre-electoral Coalition Strategies in Multiparty SystemsThomas Gschwend, Lukas F Stoetzer, and Indridi IndridasonEuropean Political Science AssociationConference 2018 and Swiss Political Science Association (SPSA) Annual Conference in Zurich 2019 and 2021. 2021Coalition PoliticsRepresentationGovernment coalitions are a foreseeable and central feature of governance in multiparty systems. This fact often compels parties to make coalition politics part of their pre-election campaign effort, signaling their preferred coalition to the electorate. In this article, we derive theoretical expectations that take voters reaction to pre-electoral coalitions into account under what conditions pre-electoral coalitions form and which parties form such coalitions. The model reveals that the ideological configuration of the coalition matters: Parties favor joining pre-electoral coalitions with partners that are found on the same side of the ideological spectrum. Bringing together data about pre-electoral coalitions in 398 legislative elections from 22 advanced industrialized democratic countries from 1946 to 2014 permits us to support this hypothesis. The finding that parties are more likely to enter bloc pre-electoral coalitions has important implications for accountability and representation in proportional systems.
@article{gschwend2019, bibtex_show = {true}, status = {conf}, project = {pec}, year = {2021}, impact = {3}, title = {Pre-electoral Coalition Strategies in Multiparty Systems}, keywords = {Coalition Politics, Representation}, author = {Gschwend, Thomas and Stoetzer, Lukas F and Indridason, Indridi}, journal = {European Political Science AssociationConference 2018 and Swiss Political Science Association (SPSA) Annual Conference in Zurich 2019 and 2021.}, preview = {wp_placeholder.png} }