City: Barcelona
Language: English
Credits: 0 EC
About
Structural equation modeling (SEM) is a very general statistical technique, as it has regression analysis (and therefore t tests, ANOVA, and correlation analyses), path analysis, and factor analysis as special cases. It is also possible to combine the advantages of these techniques, which makes SEM one of the most general and most flexible techniques available to researchers. As a result, SEM presently is also one the most widely used techniques in the social and behavioral sciences.
This course will introduce you to the fundamentals of SEM by first translating univariate regression models into mean and covariance structure (MACS) analyses. The first day will also show you how path analysis is more general than the general(ized) linear model and better able to facilitate testing hypotheses about mediation. The second day will introduce measurement models for latent variables (factor analysis) and show how a “full SEM” consists of both measurement and structural components (i.e., factor and path models). Day 3 will cover tactics for evaluating data–model correspondence.
Course leader
Beth Grandfield - Utrecht University