Tobias Widmann
  • About
  • Publications
  • Work in Progress
  • Grants & Awards
  • Tools
  • Teaching
  • Media
  • Contact

About

Tobias Widmann

I am an Associate Professor of Political Science at the Department of Political Science at Aarhus University, Denmark. I study political communication, emotions and morality, and political behavior, with a methodological focus on computational text analysis and natural language processing.

Previously, I was a PhD researcher at the European University Institute in Florence and held visiting positions at the Hertie School Data Science Lab and the University of Wisconsin-Madison. My research has been supported by the Fulbright Schuman Scholarship, and my work has received awards including the Rudolf Wildenmann Prize. I publish in leading journals, including The Journal of Politics, Political Analysis, Comparative Political Studies, and the British Journal of Political Science.

A slice of the space my methods work in: 64 German terms from the ed8 emotion dictionary, placed by the word2vec vectors trained for Widmann & Wich (2022) and projected to two dimensions (t-SNE). Positive and negative vocabularies separate on their own — hover a word or an emotion.

  • Anger
  • Fear
  • Disgust
  • Sadness
  • Joy
  • Enthusiasm
  • Pride
  • Hope

Research interests

See publications →

Data I work with

  • 0parliamentary speeches — eight national parliaments plus all 16 German state parliaments, up to six decadesiScience 2024 · Political Behavior 2021 · BJPS
  • 0tweets from citizens, plus the Twitter/X accounts of MPs, parties and news mediaPolitical Psychology 2022 · PSRM 2024 · JOP 2025
  • 0party press releases and tweets from three European countriesPolitical Psychology 2021
  • 0Economist articles spanning 177 years, 1843–2020Comparative Political Studies 2024
  • 0crowd-coded sentences in six languages, the training data behind my classifiersPolitical Analysis · Tools
  • 0Western democracies compared — from parliamentary debates to party Facebook posts and Manifesto Project positionsBJPS · Political Analysis
  • 0kinds of context linked to text: wind-turbine construction records, daily weather, COVID-19 case counts, election timing, party positionsJOP 2025 · iScience 2024
  • 0open-source tools for measuring emotion and morality in text, free for academic researchTools

How I work with it

  • Fine-tuned transformers
  • Custom dictionaries
  • Word embeddings
  • Crowd-coding & validation
  • Staggered difference-in-differences
  • Vector autoregression
  • Panel fixed effects
  • Cross-national comparison

Measuring what politicians say · explaining why and with what effect