9 August 2019
Introduction to Structural Equation Modeling: Confirmatory Factor Analysis with Mplus
The course focuses on measurement models and their application within the Structural Equation Modeling (SEM) framework. We will show how a theoretical model, represented by measurement models, can be applied to empirical data and how to assess its fit to the data through the measurements' covariance matrix. Confirmatory Factor Analysis (CFA) is an important and basic aspect of the SEM-framework and its understanding and application to data is the core learning aspect of this course. Also, CFA is a necessary conceptual precondition to understand and apply the structural aspect of SEM, path modeling. Therefore, the course deals with concepts and applications of CFA such as assessing construct validity and reliability of a measurement model as well as the interpretation of calculated results. The topics addressed in the course include different modeling techniques of CFA such as single measurement models, simultaneous CFA (SCFA), the Multiple Group Comparison of the CFA (MGCFA), and the higher-order CFA. If time permits on the last day, we can peak into topics as CFA with categorical data, path-modeling, how to handle missing data, or longitudinal analysis. Throughout the course we will work on examples provided by the lecturers using the popular SEM software package Mplus. For data preparation we accommodate needs of SPSS- or Stata-users.
Prof. Dr. Jost Reinecke is professor of quantitative methods of empirical social research at the University of Bielefeld, Germany.
Georg Kessler is a researcher at the Institute for Methods of the Empirical Social Sciences at the University of Bielefeld
Participants will find the course useful if they
(on the level of their research questions)
- work with models that involve a complex structure of variables involving latent concepts and their relationships to each other;
- have a strong deductive framework and want to verify theoretical assumptions derived from substantive theories;
- need information on measurement quality (validity and reliability testing)
- want to apply SEM to their future analysis.
(on a more basic level)
- want to get an introduction into Structural Equation Model (SEM)-framework;
- have had prior experience with SEM, but no formal training;
- they have had prior training, but still find the whole matter rather complicated;
- they want to further their understanding of Mplus.
While this course is introductory in nature, its theoretical input should be dense enough to help more advanced users to effectively brush up their knowledge.
- we strongly encourage participants to familiarize themselves with and have a conceptual/mathematical understanding of variance, covariance, correlation, standardization, hypothesis testing (t-test, chi-square), and regression analysis [for compact refreshing we recommend http://davidmlane.com/hyperstat/];
- basic knowledge of matrix notation [a short refresher can be found on https://www.youtube.com/watch?v=G16c2ZODcg8];
- knowledge of what a linear equation system looks like and how it can be solved;
- handling of system files (.sps; .dta; …) and transformation to portable or ASCII-data files (.dat; .csv; .txt; …) [for SPSS users: a good preparation is to import .txt-files into SPSS and use SPSS-syntax to get data; for Stata users: a good preparation is to use the stata2mplus ado in Stata to get Mplus input and data file simultaneously];
- as introductory reading we also recommend studying the chapters 1 to 3 of the Brown book (cited in the course literature);
- basic familiarity with Mplus (can be acquired in the short course “Using Mplus for Latent-Variable Modelling: An Introduction” in week 0) and familiarity with writing syntax (Mplus input - as taught in the class-is syntax only) [we recommend looking into chapter 5 of http://www.statmodel.com/ugexcerpts.shtml].
For the duration of this course, GESIS will provide participants with access to the required statistical software packages.
By the end of the course participants will:
- know how to define a latent construct through a measurement model;
- comprehend the mathematical and statistical foundation of SEM;
- be able to read, understand, and interpret an Mplus output;
- transfer the theoretical knowledge to applied research projects;
- in general be enabled to acquire the set of skills they need for their individual projects.
- Certificate of attendance issued upon completion.
- 4 ECTS points via the University of Mannheim for regular attendance and satisfactory work on daily assignments and for submitting a paper/report of about 5000 words to the lecturer(s) up to 4 weeks after the end of the summer school (EUR 50).
EUR 300: Student/PhD student rate.
EUR 450: Academic/non-profit rate.
The rates include the tuition fee, course materials, the academic program, access to library and IT facilities, coffee/tea, and a number of social activities.
10 DAAD scholarships are available via the Institute of Sociology and Social Psychology (ISS) of the University of Cologne.
5 scholarships for participation in one main course are available from the European Survey Research Association (ESRA).