I’m Hauke, Assistant Professor of Computational Political Science at the Department of Political Science and the Digital Science Center of the University of Innsbruck, Austria.

I develop and apply computational methods for the comparative study of political communication strategies. For example, in a paper with Ronja Sczepanski (Sciences Po Paris) in the British Journal of Political Science, we present a novel, deep learning-based method for extracting mentions of social groups from political texts that has received the PolMeth Europe Best Poster Award for Innovative Data Science in 2022. And in a recent paper forthcoming in the Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP), my co-authors and I develop and evaluate LLMs in the measurement of scalar socail science constructs, like the negativity of a campaign ad, the level of arousal in a political speech, the intensity of fear and anger expressed in open-ended survy responses.

I’m particularly interested in the role of rhetorical strategies in democratic representation, electoral competition, and legislative politics. For example, in a Journal of Politics paper, Tarik Abou-Chadi, Pablo Barberá, Whitney Hua, and I apply machine learning methods to Twitter data to investigate how challenger and mainstream parties adapt their anti-elite rhetoric to coalition formation incentives and their standing in the polls. From 2026 to 2029, I will pursue this research interest in the third-party funded project Group Appeals in Parliamentary and Electoral Debates that I will co-lead with Lena M. Huber (University of Mannheim).

Another part of my research focuses on multilingual text analysis and aims to equip scholars and data analysts with innovative tools for comparing political processes and decisions across diverse contexts and languages. For example, in my paper “Cross-Lingual Classification of Political Texts Using Multilingual Sentence Embeddings” published in Political Analysis, I show how translation-free, state-of-the-art transfer learning strategies can enable reliable cross-lingual political text analysis. In a paper forthcoming in Political Science Research and Methods, my co-authors and I validate open-source machine translation models for cross-lingual measurement with computational text analysis methods like topic modeling or transformer classifier fine-tuning. And in a Computational Communication Research paper co-authored with Fabienne Lind (University of Vienna), we provide a guide to multilingual text analysis for political and communication scientists.

Read more about my research and professional activities in my CV. And feel free to email me or contact me on LinkedIn!