Workshop - Structural Equation Modeling (SEM) - June 23 - June 26th, 2011

Three-and-a-half day workshop introducing participants to the mathematical underpinnings of SEM models, emphasising practical applications in current statistical approaches using MPlus software.

Structural equation models (also referred to as “SEM models”) have become popular in the applied health sciences as well as the Social Sciences. These models are quite general, and subsume many of the multivariate techniques typically employed by researchers including regression models, factor analysis, analysis of variance/covariance, principal components analysis and path modeling.

A strong feature of structural equation models is that they allow for the estimation of parameters in models with multiple indicators for each construct. More recently, SEM models have provided an approach to the estimation of parameters in growth curve models for longitudinal data and an approach to the problem of the unbiased estimation of parameters in the presence of missing data.

Times: Thursday June 23 to Saturday June 25 (9:00-5:30 each day), Sunday June 26 (10:00-2:30)

Location: Hickman Building Room 105, University of Victoria, Victoria, BC

Fee: $750. Fee includes comprehensive course notes and refreshments.

Instructor

Doug Baer is a Sociologist and current Academic Director of the UVic Branch Statistics Canada Research Data Centre. Dr. Baer has taught SEM courses for over 20 years, most notably at the Inter-University Consortium for Political and Social Research Summer Program in Quantitative Methods in Ann Arbor Michigan.

Course description

The workshop will introduce participants to the mathematical underpinnings of SEM models, but will emphasize practical applications in current statistical approaches using MPlus software.

Topics will include:

The relationship between path models and SEM models
Conceptualizing latent variable-manifest variable relationships
Assessing the fit of models and using diagnostics to improve models
Simultaneous analysis of models in multiple groups
Models for non-normally distributed data
Models for means and intercepts.
Introduction to growth curve models and other models for panel (longitudinal) data

Prerequisite

Participants should have a reasonably strong background in multiple regression (including dummy variable regression). Some experience with or knowledge of factor analysis would be beneficial but is not essential.

Register now!

Sponsored by Population Data BC, in partnership with the Division of Continuing Studies, University of Victoria