Under Review
How Snap Election Calls Shape Vote Intention Uncertainty (with Alejandro Fernández-Roldán)
Abstract:
While prior research has examined attitudinal and behavioral responses to snap elections, in this paper we study whether snap election calls influence vote intention uncertainty, measured through “don’t know” responses in public opinion surveys. We find that snap election calls immediately and significantly reduce vote intention uncertainty. We demonstrate this through two cases: the 2025 German Federal Election and the 2024 UK General Election. Leveraging a quasi-experimental unexpected event design in Germany and individual-level panel data in the UK, we show that snap election calls accelerate the timing of vote preference formation. Yet the distribution of these effects differ across contexts: in Germany, reductions are concentrated among non-partisans, whereas in the UK they are more broadly distributed across the electorate. Our core contribution is to show that snap election calls operate as a distinct political mechanism that triggers more rapid engagement and evaluation, thereby shifting the baseline level of vote uncertainty prior to the campaign. In doing so, we advance understanding of how disruptions to the expected electoral calendar shape the formation and expression of voters‘ preferences.
Manuscripts in Preparation
Electoral Polling as a Public Service? An Analysis of Spain’s CIS Performance in Pre-Election Estimates Over Time (with Alejandro Fernández-Roldán)
Do Events Change Minds or Samples? A Framework for Identifying Compositional Bias in Event-Focused Causal Inference
Abstract:
Major political events can shape public attitudes and behaviors but may also affect who responds to surveys. This paper brings together insights from political psychology and survey methodology to argue that such events can bias causal estimates if shifts in participation, especially through unobserved factors, are related to outcomes of interest. Using the increasingly popular design of unexpected events during survey fieldwork (UESD), I propose a framework to disentangle the true causal effect of an event from compositional bias, i.e. bias from changes in sample compositions after the event.
I outline strategies to adjust for observable imbalances and extend sensitivity analyses to assess how strong unobserved confounders would need to be to threaten substantive (causal) conclusions. I demonstrate this approach using the rally-around-the-flag effect after the 2015 Charlie Hebdo attacks in France, then apply it to replicate 14 published UESD studies on terrorist events and rally-style outcomes like government approval and trust in political institutions. By addressing both observable and unobservable forms of bias, this framework improves the robustness of causal inference in event-driven research and strengthens the assessment of public opinion measurements‘ credibility in dynamic political settings.
Reassessing Predictors of Vote Choice and Electoral Turnout: The Role of Attitude Strength (with Harald Schoen)
Political Drivers of Survey Nonresponse: The Role of Attitudes and Local Political Context (with Joel Ardiaca Arnau)
The German Snap Election and the Challenge of Compositional Bias in Unexpected Event Designs
Klara Müller 