Customer Reviews

Structural Equations with Latent Variables

5 star:		(5)
4 star:		(3)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

The software Lisrel was developed to model and analyze data using structural equation models which involve the introduction of latent variables. Although this topic has historically been most commonly used in the social sciences including psychology and sociology, it is finding a wide range of applications as statisticians encounter more and more problems where it is appropriate to use latent variables.

Bollen provides a thorough treatment of the topic that has advanced some since the publication of the book . This is still the best source for a detailed account of the methods. Bengt Meuthen at UCLA was one of the pioneers of the methodology and his books and papers provide good additional sources for the reader who wants to understand the theory and the software tools.

Bollen is comprehensive and thorough, sometimes even to a fault. This work is definitely for the advanced academic who works with structural equations on a regular basis - certainly not for the faint of heart.

Although the field has made advances in the thirteen or so years since this book was first published, I found that Bollen has weathered well, is still suitable for instruction and practice, and should remain so for many years to come.

This is the definitive textbook on structural equation models with latent varibles. It is a bit dated now, and the treatment is rigorous, but any serious user of LISREL, EQS, etc. should purchase and read this book. For what it's worth, Bollen is not only a good writer, but a great teacher.

This book sticks out like a beautiful flower in a bed of weeds.

There is a dearth of really good books available to sociologists interested in deepening their knowledge of statistics in general, and structural equation modelling in particular. Most of the books I encounter are either too simple or too difficult. There is either too little, or too much detail. I suppose that in general writers cater to two types of statistics users: the statisticians, and those applied users who prefer the "black box" approach in running their regressions. A good example of a bad statistics books for sociologists: the often-cited econometrics textbooks that read like dictionaries by providing lists of formulas rather than integrating knowledge by carefully explaining general concepts.

And then there is Bollen's 1989 SEM book. Almost a class by itself. He achieves a balance between comprehensiveness and comprehensibility ideal for the sociologist. Bollen excels at the task of explaining statistics. If someone is thinking of taking a first class in SEM, I would recommend getting a copy of the syllabus. If the syllabus doesn't include this book, I would recommend that the person consider dropping the class and reading this book if at all possible.

Someone needs to convince Bollen to write books about every statistics topic he has mastered, and then to learn other topics and write books on those. Bollen is a resource that should not be missed.

I've taken a class in Structural Equations Modelling from Ken Bollen at UNC Chapel Hill. Naturally the book was the main reference for the course, so I went over most of it during the semester. The coverage in classes differed from the book material due to the recent developments, and my impression was that Prof. Bollen planned on writing a second edition of the book (although he is not quite happy about the pricing strategy of Wiley). For many things marked as "currently open problem" in the book, the solution, at least a partial one, have been found in the last 15 years of research in this dynamically developing area. The book is still a compendium of important basic results, and will be one for at least another 5 to 10 years, until another modern and rigorous treatment appears in press; so far this is the best reference I've seen.

Bollen's book about structural equations is exceptionally good. Bollen does not only write and explain very well and thorough but he also is able to communicate a difficult subject. I would guess that even people form rather non-mathematical fields will find this book great because he really tries to explain even small technical details frequently neglected.

With great book I mean not only, that it provides you with insights about structural equations but also with enough practical knowledge to use this information in your own problems. This is in contrast to, e.g., Pearl's book about causality.

Yes, this book is already same years old, but I would nevertheless say it is the best book available about structural equations. The progress made in the last decade is more computational rather than a conceptual.

I strongly recommend this book to anyone who has at least some basic math skills or decides now to be more interested in. This should be enough to start reading. Two thumbs up!!

Great book that mixes a bit of philosophy of science in with the rigorous mathematics. This book has over 10,000 citations according to Google Scholar. For most of us, that is all we need to know!

The editor of this book, clearly was an amatoeur. He/she designed
the book such as only chapters are numbered (sic!), but sections or
subsections not (but equations are numbered fortunatelly)

In result the (often used) references point to the whole chapters, while
sections/subsections would be much better. So Bollen says some stuff
was described in more details in chapter so-and-so, and the reader
have to waste the time searching through 50 pages or so...

Shame on Wiley...