STA2201: Methods of Applied Statistics II, Winter/Spring 2006

Structural Equation Models

http://www.fisher.utstat.toronto.edu/~brunner/2201s06

Lecture: Tuesday 10:10 - 11:00 a.m. and Thursday 10:10 a.m. - 12:00 noon, WI 523.

• Instructor: Jerry Brunner
• Office: 6025B
• Phone: 416-978-0673
• email: brunner@utstat.toronto.edu
• Office Hours: Tuesday 11:10 a.m. - 1:00 p.m.

Text: Structural equations with latent variables, by K. Bollen. It is available at the bookstore.

What's it about? Structural equation models represent a generalization of the usual linear model. They consist of systems of linear equations, which may include latent (unobservable) random variables that are not error terms. A random variable may appear as an independent variable in one equation and a dependent variable in another. Important special cases are regression with measurement error in the independent variables, path analysis, factor analysis and the so-called "hierarchical linear models." We will work through a good part of Bollen's text, and also learn how to do it with SAS. There will be a little bit of R early in the course.

Evaluation: There will be weekly quizzes given at the beginning of class on most Thursdays, starting January 19th. Quizzes will be based closely on weekly homework assignments. Computer portions of the homework may be handed in as part of the quiz. Non-computer portions are to be done as preparation for the quizzes, and are not to be handed in. No makeup quizzes will be given.