ABSTRACT
Following the explanation of the measurement model in Chapter 4, this chapter discusses the testing of relationships across constructs. The concept of path analysis is introduced and a detailed example given. Next, a discussion of full structural modeling is presented where the measurement model is included in the testing of relationships across constructs. Further, the specification search function is discussed in how to trim a structural model. This chapter also explains the importance of control variables and how to specifically test a control variable in a structural model. Concluding the first part of the chapter is a discussion on alternative models and why this can add to the validity of your findings.
The second half of the chapter addresses the topic of two group analysis and testing relationships across groups. The remainder of the chapter looks at troubleshooting potential problems in structural models. How to find and address multivariate outliers, how to determine if your data is not normally distributed, and what to do if you get error messages like your model is not positive definite or an iteration limit has been reached in the analysis are just some of the questions that are addressed.
