Visiting Scholar in Political Science, Yale University
MSc Student in Social and Economic Data Science, University of Konstanz
I study political methodology as well as American and comparative political behavior. I am currently a Visiting Scholar in Political Science at Yale University for the 2025–26 academic year, hosted by Shiro Kuriwaki, and an MSc student in Social and Economic Data Science at the University of Konstanz. My research focuses on causal and descriptive inference under selection bias and measurement error, and on the generalization of findings across populations under distributional and conditional heterogeneity, in the context of political representation and electoral behavior. I received a BA in Politics and Public Administration from the University of Konstanz, graduating with distinction as the top student in my cohort, and completed exchange studies at the University of Leeds, the University of Zurich, and ETH Zurich. Since 2022, I have worked as a research assistant to Peter Selb and previously as a teaching assistant to Susumu Shikano.
jens.wiederspohn@yale.edu
Yale University
New Haven, CT 06520
Case studies of polling failures focus on within-election differences in poll accuracy. The crucial question of why polls fail in one election but not in others often remains a matter of speculation. To develop a contextual understanding, we review and unify theories of election features suspected of encouraging polling errors, including mobilization, candidacies, polarization, and electoral conduct. We extend a Bayesian hierarchical modeling approach that separates poll bias and variance at the election level and links error components to electoral predictors. Investigating 6,375 preelection polls across 318 US Senate elections from 1990 to 2022, we find an overall trend toward smaller but more uniform errors. Poll variance exhibits a weak negative association with mobilization and polarization. Until 2004, front-runners and incumbents were overestimated, but there is little evidence that polls are biased for female or minority candidates. Finally, Republicans in states with lower levels of state democracy are slightly underestimated in recent years.
Surveys are vital for studying political participation but often face two key challenges: individuals who engage in politics are overrepresented in surveys, and participation is overreported by respondents. Both phenomena inflate marginal survey estimates. But how do they affect inferences about social differences in political participation? Extant research does not provide a clear answer. We employ directed acyclic graphs to encode assumptions about survey participation and responding, and determine potential biases. We then use validated votes and social characteristics of both respondents and nonrespondents from the Swedish National Election Studies 1998–2022 to empirically assess biases in estimated turnout disparities across age groups, gender, education, income, nativity status, and place of residence. We find that both selection and measurement biases have been rising and often reinforcing each other to attenuate social differences in turnout. We discuss implications for survey research, advocating for routine sensitivity analysis in studies of political participation and beyond.
Ensuring that elected representatives reflect majority preferences is a central challenge in first-past-the-post electoral systems, motivating reforms such as mixed-member and ranked-choice voting. Yet such institutional designs can generate new distortions at the district level. I examine this possibility in Germany’s mixed-member proportional system using the Condorcet criterion, which identifies as most representative the candidate who would defeat all others in pairwise contests. Leveraging pre-election survey data on candidate and party evaluations, I simulate hypothetical head-to-head contests among district candidates and aggregate predicted choices to the population of actual voters in each constituency. Across the 2021 and 2025 Bundestag elections, a pronounced divergence between plurality outcomes and Condorcet-consistent outcomes emerges. In 2021, all sixteen AfD district winners are highly likely Condorcet losers, while many unelected Social Democratic candidates are likely Condorcet winners. The pattern intensifies in 2025, with all forty-six AfD district winners appearing to be Condorcet losers and many CDU/CSU candidates estimated to be Condorcet winners. These findings indicate that mixed-member systems can generate district outcomes in which plurality winners are not the candidates who would receive majority support in pairwise comparisons.
Public debate in Germany has increasingly focused on proposals to simplify access to means-tested basic income support. One such proposal involves introducing a grace period during which applicants’ assets and housing costs are not immediately subject to strict eligibility checks. We examine which groups would benefit from such provisions. The findings suggest that individuals with completed vocational training and longer employment histories would be the primary beneficiaries. These reforms would therefore align with the political objective of more strongly recognizing prior contributions to the labor market and could help alleviate concerns about downward mobility among the middle classes.
About this website: This website is based on Shiro Kuriwaki's website, uses code from Minimal Mistakes, Github Pages, uses some CSS from Matt Blackwell's website at the time, and is inspired by Sirus Bouchat's website, Andrew Hall's website, and Hanno Hilbig's website.