Using LLMs and Text-as-Data in Political Science Research

Schedule

Time	Title	Presenter
09:00-09:45	Introduction	Xavier Fernández-i-Marín
09:45-10:30	LLMs for Text Annotation: Promises, Patterns and Pitfalls of a new Methodological Approach	Arnault Chatelain
10:30-11:15	Machine Bias. How much, and how, are LLMs biased? An Experiment	Etienne Ollion
11:15-11:45	Break
11:45-12:30	Measuring Populism with Large Language Models: Prompt engineering as qualitative social science	Petter Törnberg
12:30-13:15	Scaling Political Texts with ChatGPT	Gaël Le Mens
13:15-14:30	Lunch
Parallel afternoon session 1
14:30-15:15	Central Banking Under Pressure: A Textual Measure of Dominance and Coordination	Lauren Leek
15:15-16:00	Deep Active Learning for Data Mining from Conflict Text Corpora	Mihai Croicu
16:00-16:45	Coding Political Regimes with LLMs	Nils B. Weidmann
16:45-17:30	More Discretion, More Regulation? Examining Regulatory Production in the United Kingdom, 1991-2022	Nir Kosti
Parallel afternoon session 2
14:30-15:15	Party discipline and dyadic representation	Oda Nedregard
15:15-16:00	Holy Words, Contentions Politics: Discourse on Abortion Rights in Religious Sermons	Nela Mrchkovska
16:00-16:45	DOPEH – Dynamics of online political elite hostility: The audio-visual packaging of negative and uncivil campaign ads on meta	Philipp Mendoza
16:45-17:15	Political Sycophancy of Large Language Models: Do they align with German voters?	Jan Batzner
17:30-18:30	Keynote	Fabrizio Gilardi

Venue

Room “Maria Soteras”, 4th floor. New building (white one). Facultat de Dret, Universitat de Barcelona Avda Diagonal 684, 08034 Barcelona

View Larger Map

Enter the white building (shown here, named “Facultat de Dret”) and get one of the elevators to the fourth floor. Room “Maria Soteras” is just in front of the elevator when getting out in the fourth floor.

Organizers

Aina Gallego (UB), Gaël Le Mens (UPF) & Xavier Fernández-i-Marín (UB).