DATIC Workshop: SEM

Among the topics covered to be covered are the basic terms, path analysis, confirmatory factor analysis, and latent variable modeling. Topics NOT to be covered are categorical and ordinal outcomes, models with feedback, and models with means. Also, we do not plan to discuss explicitly longitudinal modeling, but some of the methods can be used for longitudinal models. Participants will learn how to set up a structural equation model (SEM) and be able to interpret the results from their analyses. Participants are expected to have a working knowledge of multiple regression. Instruction consists of lecture, computer workshops, and individualized consultations. The emphasis will be practical with minimal emphasis on statistical theory. The instructors and a teaching assistant will also be available for individual meetings.

The computer workshop will use the program AMOS graphics. No prior experience with AMOS is required. Also in meetings with individual participants, alternative programs (e.g., LISREL, EQS, Mplus) can be discussed.

Because the number of spaces is limited, please apply soon to reserve a space. To register, send an email to DATIC staff, and complete the registration form (click here to get the form).

Participants are encouraged to bring their own data, though it is not necessary to do so. They should contact the staff beforehand to ensure that data are appropriately formatted for analysis.

Curriculum

Day 1: Basic Terms and Path Analysis

Review of Multiple Regression
Basic Vocabulary of SEM
Steps in SEM

Specification
Identification
Estimation
Testing

Path Analysis
Model Testing and Respecification
Estimation of Path Models
Lab

Day 2: Confirmatory Factor Analysis

Latent Variables
Brief Description of Exploratory Factor Analysis
The Basics of Confirmatory Factor Analysis
Identification
Problems

Discriminant validity
Heywood cases
Empirical underidentification
Single indicator variables with known reliability

Respecification

How to respecify
Tools for Respecification

Lab

Day 3: Hybrid Models

Combining Path Analysis and Confirmatory Factor Analysis

Measurement Model: CFA
Structural Model: Path Analysis
df of model = df of CFA + df of Path Analysis

Model Comparisons
Measuring Fit
Respecification of the Measurement Model (CFA)
Lab

Day 4: Mediation and Moderation in SEM

Mediation in Latent Variable Models

Basic mediation equation
The Four Baron & Kenny Steps
Using SEM to estimate and test mediation
Specification error and alternative models
Bootstrapping to test indirect effects

Moderation

Definition
Linear and nonlinear moderation
Multi-group options in AMOS 13
Tests of moderation

Dealing with Problem Models
Lab

Day 5: Remaining Issues

Missing Data
Non-normal Distributions
Alternative Programs
Standardization
Specialized Latent Variable Models
Examples of Longitudinal Models in SEM
Limits and Potential of SEM

Noon LUNCHEON