Measuring policy positions of political agents
Date:
09/11/2026
Organised by:
London School of Economics and Political Science
Presenter:
Emeritus Professor Michael Laver
Level:
Intermediate (some prior knowledge)
Contact:
Venue: Online
Training for PhD and MSc students in the design of social research, quantitative and qualitative analysis.
Measuring policy positions of political agents by Michael J Laver (Emeritus)
Measuring the position and salience of agents' preferences on key issues is a core empirical project for social science. Methods used to do this include: surveys of experts, political elites and voters; scaling of legislative roll-calls; expert, crowd-sourced or automated analyses of text. Large language models (LLMs) can now do this much more quickly and cheaply.
Differences between methods highlight important conceptual and methodological questions about what we mean by, and how we measure, issue position and salience. These include the separability of estimated positions and saliences on different issue dimensions. We focus on a range of different approaches to measuring the position and salience of agents' preferences, but particularly on the systematic and replicable use of LLMs.
Cost:
Free
Website and registration:
Region:
Greater London
Keywords:
Data Collection, Data Quality and Data Management , AI and machine learning
Related publications and presentations from our eprints archive:
Data Collection
Data Quality and Data Management
AI and machine learning