32 of 33 people found the following review helpful:
5.0 out of 5 stars
Easy to read and understand; needs a second edition, October 3, 2004
No one expects statistics to be easy reading, especially when it concerns complicated models such as structural equation modeling (SEM). Nevertheless, this book manages to do just that.
Schumacker and Lomax have successfully put together a guide that explains to beginners (like myself) in simple terms how the whole thing works. As with most books that treat complex models, some basic knowledge of statistics is preferable before you begin to read it. But if your statistics is rusty and you have only vague impressions of probability sketches in your memory, fear not! Schumacker and Lomax are kind to us poor souls, and begin by introducing some basics in chapter 1 to prod your memory: terminology, variable scales, how to treat missing data, outliers and normality. And in chapter 2, they discuss correlation and covariance.
Before talking about structural equation models, Schumacker and Lomax dedicate chapter 3 to a number of statistical methods on which SEM is built. This chapter gives a basic overview of regression, path analysis and factor analysis. The review of these methods helps you to understand SEM better later on. They also provide an excellent understanding of the methods, in case you have not used them before or it's been a while ...
The rest of the guide covers SEM: how to develop and measure a model (chapters 4 and 5), how the model parameters are estimated and how you can check for reliability and validity (chapter 6), and checking for goodness of fit of your model (chapter 7).
In chapter 8, you are shown some examples of computer outputs by two software packages that can conduct SEM, EQS5 and LISREL8-SIMPLIS.
Chapter 9 goes into more detail on models and diagrams (regression, analysis of covariance, path, measurement and structural models). For those that feel by this point that they've gained enough experience, advanced topics such as cross validation, simulation, bootstrap and jacknife methods as well as multiple same and interaction models are covered in chapter 10. And for the super-keen, the technical bits are covered in chapter 11 (health warning: you better be up to speed on matrix algebra).
The great thing about this book is that you most likely will be able to run models and interpret results by chapter 7, and you don't need to go into the nitty-gritty if you don't want to. On the other hand, the details are there if you need them. In essence, the authors start at the beginning, building up slowly until you are able to handle a basic model, before going into more complex issues.
One drawback, I have found, is that this book was published in 1996. That's nearly a decade ago, and (fortunately) computer power and statistical modeling has come a long way since then. The authors, for example, are convinced that WordPerfect is the software of choice for word processing and that at some point in the future it would be possible to copy and paste diagrams into a word processing program. We've come a long way since then. Today, many SEM packages exist that are much more user friendly than some of the older stodgy packages that require you to enter data in a very specific way and interpret results by going through reams of data output. As such, the authors (or publishers) probably ought to update it with a second edition.
Nevertheless, the strength of the book lies in its guidance and explanatory power. And even if you use a different package, you can skim through the data outputs they use, and focus on your model, how to construct it, and of what pitfalls to beware.
I highly recommend this for anyone starting on SEM - your modeling days will be much happier with this guide.
Help other customers find the most helpful reviews
Was this review helpful to you? Yes
No
