Textual Measures of Populism (TEMPOP) for the Analysis of Party Competition and Political Behaviour

Research question/goal: 

The project "Textual Measures of Populism (TEMPOP) for the Analysis of Party Competition and Political Behaviour" seeks to contribute scientifically to research on populism, party competition, and political behaviour in conceptual, methodological, and analytical terms. A quantification of populism eases the scientific study as well as the societal discussion of populism and its causes or consequences. In the course of the project, the information obtained on the degree of populism of politicians and political parties is used to answer research questions on patterns of political competition between populist and mainstream parties as well as on the impact on individual political behaviour. The project applies statistical models measuring populism from political text (party manifestos, political speeches and (social) media), thereby crossing contextual and language barriers, and contributes to the analysis of causes and consequences of populism.

Current stage: 

The project has found internal partners at the MZES and thus gained momentum. The planned research can be divided into three work packages: 1) the acquisition of comprehensive relevant text data (party manifestos, parliamentary speeches, news articles, and social media data) for a selection of countries over time, 2) the annotation of the data and the development and training of a multilingual machine learning model for the detection of populism across text sources, and 3) the analysis of interactions between political and media actors regarding their populist rhetoric. The submission of a jointly developed DFG grant proposal is scheduled for 2023.

Fact sheet

Funding: 
MZES
Duration: 
2018 to 2024
Status: 
in preparation
Data Sources: 
party manifestos, political speeches, Tweets
Geographic Space: 
16 established European democracies

Publications

Journal Articles

Storz, Anna, and Julian Bernauer (2018): Supply and Demand of Populism: A Quantitative Text Analysis of Cantonal SVP Manifestos. Swiss Political Science Review, 24, issue 4, pp. 525-544. more

Conference Presentations

Rehbein, Ines, Josef Ruppenhofer and Julian Bernauer (2021): Who is we? Disambiguating the referents of first person plural pronouns in parliamentary debates. [Proceedings of the 17th Conference on Natural Language Processing (KONVENS 2021), Düsseldorf, Germany, September 06th to September 09th, 2021] more
Bernauer, Julian, and Federico Nanni (2019): Cross-Lingual Topical Scaling of Political Text using Word Embeddings. [General Online Research (GOR), Köln, March 06th to March 08th, 2019] more