- Tue 21 Apr 2015 09:00 - 15:00
- Mon 20 Apr 2015 09:00 - 15:00
Seminar Room 2, Badia Fiesolana
This 2-day workshop, financed by the eduLIFE project via the ERC funds, consists of two parts. Part 1 deals with introductory and intermediate aspects of the increasingly popular methodology of structural equation modeling (SEM; latent variable modeling, LVM) in the political, social, and behavioral sciences. In this part, a coherent introduction to the basics of the methodology is provided, including model identification issues, implied covariance and mean structure, parameter estimation, (robust) maximum likelihood estimation and (asymptotically) distribution-free estimation, as well as model fit evaluation.
Part 2 of the workshop is focused on longitudinal data analysis and modeling using the SEM/LVM framework. This part is concerned specifically with fitting unconditional and conditional (covariate-based) models to data from repeated measure studies, which are increasingly popular in the social sciences. Frequently arising empirical issues in longitudinal designs, such as incomplete data, nesting of subjects (of special relevance and frequency in industrial/organizational settings) as well as non-normality of longitudinally followed measures, are also adderssed in this part of the workshop.
Throughout the workshop, multiple empirical examples are provided, as well as input and output files of the highly popular SEM package Mplus are discussed in detail (a self-study booklet on using Stata for SEM is also provided to the participants). The workshop is primarily application oriented and provides instrumental skills and assistance to students and faculty engaged in empirical research in the political and social sciences.
Location: Seminar Room 2, Badia Fiesolana
Affiliation: Department of Political and Social Sciences
Organiser: Hans Peter Blossfeld - EUI - Department of Political and Social Sciences
Attachment: Workshop Outline
Organized and financed by: the European Research Council within the framework of the eduLIFE project.