Introduction to Large Language Models for Social and Political Data Analysis

Key topics covered

Large language models are increasingly used to classify documents, extract information, summarise material, and support other tasks in social and political research. For researchers approaching these methods for the first time, understanding what a model does is the starting point for deciding how to use it and what to make of its results.

This course provides a structured introduction. It begins with the concepts needed to understand language models: how text is represented, how models learn from training data, and how they generate responses. These ideas are introduced through intuitive explanations and examples, without assuming prior training in machine learning.

The course then develops the basic steps of using LLMs for research: specifying a task, preparing inputs, writing instructions, and interpreting outputs. Text analysis provides the main setting for learning these steps, drawing on examples from social and political research. Evaluation is introduced alongside the applications and then examined more systematically through questions of accuracy, measurement validity, robustness, and reproducibility. The final session places these methods within a broader range of research uses.

Across the five three-hour online sessions, participants will work with supplied examples and Python notebooks in Google Colab. Short exercises will involve modifying instructions, examining outputs, comparing results, and discussing methodological choices. By the end, participants should have the conceptual grounding to read LLM-based studies critically and begin using simple workflows in their own research.

Day 1: How do language models work?

We begin with the foundations: how computers represent text, what language models learn from data, and how they produce responses. The session introduces key concepts such as tokens, embeddings, and context, and explains how LLMs relate to earlier text-as-data methods. Lectures, visual explanations, and simple demonstrations will help participants build a shared vocabulary and interpret examples of model behaviour.

Day 2: Working with an LLM

The second session introduces the basic process of using a model: defining a task, supplying instructions and examples, and obtaining usable outputs. We consider how model choice, prompt wording, and the information provided affect responses. Through guided Python demonstrations and short exercises, participants will modify a supplied example and produce a simple comparison of outputs under different instructions.

Day 3: Using LLMs to analyse social and political data

This session connects basic model use to empirical research through selected tasks such as text classification, information extraction, and summarisation. We examine how a research concept becomes a coding instruction and how individual outputs become an analytical dataset. Worked examples and guided notebook exercises will help participants follow a small analysis, inspect the resulting labels or summaries, and explain what they represent.

Day 4: Assessing the quality of an LLM analysis

How can we tell whether an analysis is useful and credible? This session introduces evaluation through concrete examples: comparing outputs with human judgments, interpreting errors, checking sensitivity to analytical choices, and distinguishing predictive accuracy from measurement validity. We also address bias and reproducibility. Participants will examine supplied results and prepare a brief assessment of their strengths, limitations, and possible improvements.

Day 5: Further applications and responsible research use

The final session broadens the perspective through selected examples, which may include working with document collections, supporting qualitative or survey research, and generating synthetic responses. We consider how these uses differ from analysing observed data and what additional assumptions they require. Discussion and a short synthesis exercise will bring together method selection, evaluation, privacy, and transparent reporting. Participants will outline the main considerations for a chosen application.

Post-course assignment

For the extra ECTS credit, participants will complete an independent assignment. They may extend a supplied course example or critically examine a published application of LLMs in social or political research.

The submission will explain the research task and method, discuss the main analytical choices, interpret the results or evidence presented, and assess validation, limitations, and reproducibility. Participants conducting a small analysis will also submit their annotated notebook. The indicative report length is 2,500 words. Assessment will focus on understanding and applying the methodological principles introduced in the course.

Prerequisite knowledge

Basic programming skills in R or Python are required. All demonstrations and course materials will use Python. No prior knowledge of large language models is assumed.

Participants will need a free Google account to access the course notebooks in Google Colab.

Before the course

Approximately 6-8 hours before the course to read 4-5 assigned papers introducing large language models and their applications in social and political research. Readings will be provided in advance. During the course, additional reading will be limited to one paper per day.

Learning commitment

The course includes three hours of live teaching each day for five days, alongside readings and other learning activities.

Once you have registered and paid, you will receive access to our Learning Management System (LMS) approximately two weeks before the course start date. Course materials, including pre-course readings, will be available here. Allow approximately 20 hours per week before the course to review the materials and complete readings and any pre-course tasks.

During the course week, you should allow approximately two to three hours per day for preparation and assignments, in addition to the live teaching sessions.

The course offers the opportunity to earn three ECTS credits. To earn a fourth credit, you will need to complete a post-course assignment involving approximately 25 hours of work.

Disclaimer

This course description may be subject to subsequent adaptations (e.g. taking into account new developments in the field, participant demands, group size, etc.). Registered participants will be informed at the time of change.

By registering for this course, you confirm that you possess the knowledge required to follow it. The instructor will not teach prerequisite items. If in doubt, please contact us before registering.