27 August 2021
Analyzing the Unobservable: Structural Equation Modeling in Comparative Social Research
In case that an on-campus Summer School is not possible due to the Covid-19 pandemic, this module will take place online.
Sociological research is often interested in phenomena that cannot be directly measured, like xenophobia, meritocratic values or social trust. Such latent variables can be measured with the help of confirmatory factor analysis. Their relationships can be analyzed with structural equation models.
In the context of comparative research, we must make sure that our measurement instruments for latent variables are comparable between countries and cultures: Is our measurement instrument for xenophobia comparable across countries? Such questions can be answered with multi-group confirmatory factor analysis, which will be the first focus of the course. Once comparability of measures is established, we can start to ask substantial questions: Is the relationship between xenophobia and social trust the same in every culture? Such questions can be analyzed with multi-group structural equation models, which will be a second focus. Throughout the course, we will work with the statistical software Stata and data from the European Social Survey.
This module is part of part of the course Cross-national Comparative Research in the Social Sciences. It consists of two modules on statistical techniques that are useful in cross-national comparative research (and any other comparative research): Multilevel analysis and multi-group structural equation modeling. The two modules make use of the same software and data, which allows students to easily integrate the knowledge from both courses into their own substantive research projects. However, both courses also stand on their own and is it not necessary to visit both courses.
The course comprises 28 contact hours (8*3.5 hours). Upon successful completion, 4 ECTS (European Credit Transfer System) points will be awarded for the module. A single ECTS point is defined as the equivalent of 25 to 30 hours of student workload. This includes class hours, additional preparations for class activities, readings, assignments as well as final assessments.
Attendance: Participants have to attend at least 80 % of the classes.
Course leader
Prof. Dr. Alexander Schmidt-Catran
professor of sociology with a focus on methods for quantitative empirical research
Target group
Bachelor’s and Master’s students in the Social Sciences (e.g. Sociology, Anthropology) with basic statistical knowledge and high interest in quantitative research. Open for other disciplines.
Course aim
Learn how to use multi-group confirmatory factor analysis of latent variables such as xenophobia or social trust for establishing comparability of measures
Work with multi-group structural equation models for analysing research questions
Fee info
EUR 500: Includes all study materials and transcript of records