一、課程簡介:

This course basically covers Structural Equation Models (also known as `Causal Analysis` or `Simultaneous Equation Systems`). Specifically, the course introduces the LISREL (Linear Structural Relationships) modeling techniques, which are very useful in dealing with substantive problems in behavioral sciences and marketing. Applications of LISREL can be found almost everywhere. LISREL is powerful in areas like measurement and hypothesis testing. With respect to measurement, it is capable of examining reliability, dimensions of constructs, and validity. As for hypothesis testing, it produces parameter estimates and goodness-of-fit indices which can be used for testing of assumptions, theory verification, or even comparison of competing models. People who use survey or experimental design will find the LISREL methodology applicable to a wide range of models such as exploratory and confirmatory factor analysis, path analysis, and econometric models for time series data.

The format for this course will be primarily lecture. Examples drawn from marketing, behavioral sciences, psychology, organizational science, corporate strategy, international management and other related fields will be used to illustrate the ideas. This course does not involve a high level of mathematical sophistication, but a basic understanding of statistics and multivariate techniques is required. Theoretically, the course couldn`t be too much a burden for students who don`t even like quantitative methods. A sincere desire to learn is considered more important.

Since this is a course on methodology, a number of homework assignments and a term project will be required. Details on the term project will be announced in due time. Participants in this class should at least learn how to run the PC version of LISREL.

二、先修課程:

三、參考書目:

Joreskog, K. G. and D. Sorbom (1993), LISREL 8 & PRELIS 2 for Windows Manual, Chicago:
SSI, Inc. (required)
Joreskog, K. G. and D. Sorbom (1993), LISREL 8: Structural Equation Modelin