Research Areas

Several projects we are involved in combine a substantive interest in political science research questions with the development and refinement of quantitative methodologies for social science research. The approaches we develop bridge the gap between substantially relevant social science research and novel quantitative approaches adopted from statistics, as well as the natural, engineering and computer sciences.

Recent failures to predict landmark elections such as the U.S. presidential race and the Brexit vote heavily discredited polls as forecasting tools. We have been exploring alternative strategies that allow for accurate predictions of election results and realistic assessments of their uncertainty at reasonable costs. Currently, we are evaluating an instant runoff forecasting method for two-round elections in connection with the 2017 French presidential elections.

Party competition is a constitutive feature of democracies and a major channel through which electoral institutions affect political behavior. Yet how to measure it in a universal manner remains unclear. We have developed general measures of inter- and intra-party competition and have applied them to answer questions about candidate behavior and electoral turnout.

One of our overarching research interests is in rational and psychological accounts of political behavior. To what extent do candidates and voters behave strategically in order to not waste their votes on hopeless competitors? What consequences does this have for the composition of the elected bodies? How quickly do voters and elites adapt to new electoral institutions? Does campaign information help voters to make up their minds? These are some of the questions we have scrutinized in our research. Another major area of interest is unequal electoral participation and its policy implications.

Online communication is ever more important and pervasive for both our online and offline life. Our research specifically examines the role and effects of digital technology in political communication and, from a methodological point of view, the use of digital trace data in the social sciences. Two ongoing research projects place a particular emphaiss on studying the relevance and evolution of frames as well as attitude formation and change.

Most research on political behavior and public opinion relies on survey data. Yet survey data is often subject to errors that limit the validity and reliability of survey-based inferences. In our research we are trying to gauge and correct survey errors due to nonresponse, misreporting, cheating etc. We are also developing statistical models to make the most out of sparse survey data, for instance, to yield small area estimates from national survey samples.