Multivariate Methods for Social Researchers
Lecturer: Eugène Horber
Modality: In presence
Week 1: 10-14 August 2026
Workshop Contents and Objectives
The goal of this workshop is to introduce tools dealing with many variables (multivariate analysis): building models with several variables and reducing complexity (data reduction). Statistical tools are only useful and meaningful when they serve a research project, based on a well-defined theoretical framework (research question, research design, hypotheses) and good quality data. The focus of this workshop is methodological, conceptual and practical, oriented towards the application of these tools to typical analysis problems in social research.
The objectives of this workshop are:
- Lie sound foundations of knowledge and skills with multivariate statistical tools for participants who had an introductory course in statistics and need to go beyond basics.
- Acquire practical skills with data and statistical software, as well as awareness of both the potential and shortcomings and limitations (assumptions, pitfalls) of commonly used statistical tools.
- Stress the importance of embedding the use of statistical tools in a full research process, from the initial research question, to data collection and analytics, as well as reporting the results.
At the end of the workshop, active participants should be able to
- Apply the appropriate statistical tools to their own research projects, within a well-designed and defined, theoretically grounded, as well as realistic (i.e. applicable) framework.
- Understand and critically assess publications reporting results based on statistical techniques.
Workshop design
Lectures and exercises with a focus on group projects. At least 50% of teaching time will be allotted to practical work.
Detailed lecture plan (dayily schedule)
Day 1.
Introduction to multivariate analysis; quick review of univariate and bivariate statistics, statistical inference and concepts of empirical research in the social sciences focused on a multivariate view of basic tools.
Day 2.
Multiple linear regression; regression assumptions and diagnosis; analysis of residuals; building models.
Day 3.
Regression with categorical variables, logistic regression and related techniques.
Day 4.
Unidimensional and multidimensional scaling: Likert, Guttman scales, data reduction (Principal component analytics, factor analysis).
Day 5.
Outlook: tree models and classification; AI and statistics. Presentation of participants’ projects.
Class materials
A website will provide all materials, including slides, exercises and project related materials.
Prerequisites
Background in basic statistics and statistical software. You should able to perform basic statistical analyses and data transformations (recoding, variable combination) using statistical software (SPSS).
If this is not the case or you are unsure, consider the 3-day preliminary workshop Statistics with SPSS for Social Scientists mandatory for you.
Recommended readings or preliminary material
Before attending the workshop, you should read some introductory texts to social research and basic statistical concepts. Some References and Advice here
What our participants appreciated most
"It was a great bundle of theoretical fundations, practical examples provided by the professor, and team work on a project. | am extremely happy that | had the chance to learn from such an inspiring team and that | was able to enjoy the USI campus. Thank you!"
"Loved that we had so muchtime to work on our projects"
Eugène Horber
University of Geneva and FORS, Switzerland
Eugène Horber is professor emeritus of methodology at the Department of Political Science and International relations, University of Geneva, as well as an affiliated researcher at FORS. He holds a PhD degree in Political Science and has taught social science methodology (both quantitative and qualitative), applied computer science, and statistics at the University of Geneva.
He was the director of the Swiss Summer School (Social Science Methodology) for over 25 years; main teaching activities in the past include the Essex Summer School, the Carcassonne Summer School, the PRESTA programme (EU programme for South America), Eurostat/TES, ENSAE (Paris) and ENSAI (Rennes).
His research interests and publications are in the area of statistical methodology (data exploration, visual data analysis), survey research and aggregate data analysis, as well as applied computer science (didactic software, hypertext, statistical software) and computer-assisted qualitative data analysis. He is the author of a software package for exploratory data analysis.