This book unifies and extends latent variable models, including multilevel or generalized linear mixed models, longitudinal or panel models, item response or factor models, latent class or finite mixture models, and structural equation models. Following a gentle introduction to latent variable modeling, the authors clearly explain and contrast a wide range of estimation and prediction methods from biostatistics, psychometrics, econometrics, and statistics. They present exciting and realistic applications that demonstrate how researchers can use latent variable modeling to solve concrete problems in areas as diverse as medicine, economics, and psychology. The examples considered include many nonstandard response types, such as ordinal, nominal, count, and survival data. Joint modeling of mixed responses, such as survival and longitudinal data, is also illustrated. Numerous displays, figures, and graphs make the text vivid and easy to read.
About the authors:
Anders Skrondal is Professor and Chair in Social Statistics, Department of Statistics, London School of Economics, UK
Sophia Rabe-Hesketh is a Professor of Educational Statistics at the Graduate School of Education and Graduate Group in Biostatistics, University of California, Berkeley, USA.
Written by well-known experts in biostatistics and educational statistics, it presents a uniform approach to enriching both theoretical and applied latent variables modeling that also can be used in any branch of natural science or technical and engineering application. ... Numerous interesting examples ... are considered. ... Written in a very friendly and mathematically clear language, rigorous but not overloaded with redundant pure statistical derivations, the book could be exceptionally useful for practitioners. ... This book is a really enjoyable and useful reading for graduate students and researchers along with [those] from any field who wish to use modern statistical techniques to solve practical problems.
- Technometrics, May 2005, Vol. 47, No. 2
[This] is a book written for people who like to construct and to read about very general theories and modeling strategies. ... It is also a very useful book for statisticians who have specialized in one area ... and would like to learn more about another area ... . The book ... is very well written. The presentation is concise; many issues are well illustrated graphically. Altogether, the authors have written an excellent, imaginative, and authoritative text ... .
- Biometrics
This is perhaps the only book that uses the latent modeling framework to address a range of data analytical situations. ... it provided a great introduction to this field.
Dr. S.V. Subramanian, Harvard University
Overall I found the book an exceedingly valuable reference that would be ideal for graduate-level courses devoted to generalized latent variable modeling. It is very straight forward to build around it a comprehensive course where the statistical section is complemented with a multidisciplinary set of illustrative examples which can be easily replicated as both the datasets and the software are available on-line. In addition, the impressive book's breadth and depth make it an essential reference for any researcher interested in understanding the state-of-the-art methods and potential applications in latent multilevel, longitudinal and structural equation modeling.
-- Journal of the American Statistical Association
This is a very impressive book...an excellent book. I have no hesitation in recommending readers to buy this book.
-The Stata Journal (2005)
Who will profit from reading this book? On the one hand, it is a book written for people who like to construct and to read about very general theories and modeling strategies. It is also a very useful book for statisticians who have specialized in one area...and would like to learn more about another area.
The book itself is very well-written. The presentation is concise; many issues are well illustrated graphically. Altogether the authors have written an excellent, imaginative and authoritative text on the difficult topic of modeling the problems of multivariate outcomes with different scaling levels, different units of analysis, and different study designs simultaneously.
- Biometrics, March 2005
