Longitudinal Structural Equation Modeling

Course Description

Exposure to a wide variety of statistical models as special cases of the General Linear Mixed Model with latent variables. HD FS 530 Longitudinal Structural Equation Modeling (3) This course provides a broad overview of structural equations modeling as a method for studying developmental processes in Human Development and Family Studies. In this course, students gain a thorough hands-on understanding of a wide variety of statistical model types as special cases of the General Linear Mixed Model (GLMM) with latent variables. Specific statistical model types covered include: exploratory and confirmatory factor analysis; linear, nonlinear and multivariate latent growth curve modeling; quasi-simplex modeling; longitudinal factor modeling; multi-group factor analysis, including a concise introduction to behavior genetic modeling; mediation analysis; testing for measurement equivalence; MANCOVA with nonstandard within-subject covariance structures; outlines of statistical selection theory and principal component analysis. The presentation of these diverse model types as special instances of the same GLMM is helpful to understanding their relationships and differences and considerably streamlines applied statistical modeling. Each of these statistical model types are commonly used to analyze data from studies in the field of Human Development and Family Studies and illustrative examples are provided. Each model type is explained at 4 levels: 1) in terms of a set of simultaneous model equations; 2) as a set of matrix equations; 3) as a graphical model; and 4) as a Lisrel input code. All model assumptions are made explicit and the interrelationships between the 4 levels of model representation are emphasized. Then the model is applied to simulated and real data. The obtained model fits are assessed in terms of various statistical criteria and conclusions are explicitly drawn based on standard statistical decision theory. Selected models from studies of Human Development and other social sciences are interpreted in terms of content and possible pitfalls in their interpretation are discussed. For each modeling technique appropriate background publications, lecture notes and advanced reading material on nonstandard topics are provided.

Code

HDFS 530

Credits

3

Subject

HDFS