这是第一版，听说已有第二版，我把论坛币取消了，这个附件供大家做个参考好了。

Multilevel and Longitudinal Modeling Using Stata [ILLUSTRATED]  (Paperback)
        Review
A strength of thebook are the exercises at the end of each of the chapters. …The authorsare to be commended for helping foster the appropriate use of theseflexible regression models.
-The American Statistician, Vol. 60, No. 3, August 2006

Astrength of the book are the exercises at the end of each of thechapters. …The authors are to be commended for helping foster theappropriate use of these flexible regression models.
—The American Statistician, Vol. 60, No. 3, August 2006

Astrength of the book are the exercises at the end of each of thechapters. …The authors are to be commended for helping foster theappropriate use of these flexible regression models.
-The American Statistician, Vol. 60, No. 3, August 2006

Thedescription of all models is clear. …There are ample plots and outputtables to support each analysis, and very little mathematical contentto interfere with the flow of exposition. There are exercises at theend of each chapter, and they, like the examples, span a wide range ofapplication areas.
-Biometrics, September 2006

Thedescription of all models is clear. …There are ample plots and outputtables to support each analysis, and very little mathematical contentto interfere with the flow of exposition. There are exercises at theend of each chapter, and they, like the examples, span a wide range ofapplication areas.
—Biometrics, September 2006

Thedescription of all models is clear. …There are ample plots and outputtables to support each analysis, and very little mathematical contentto interfere with the flow of exposition. There are exercises at theend of each chapter, and they, like the examples, span a wide range ofapplication areas.
-Biometrics, September 2006

This is auseful reference source for researchers involved with multilevelmodeling. It gives a fairly comprehensive treatment of methods foranalysis of multilevel data, with a particular focus on random effectsmodels. Rabe-Hesketh and Skrondal’s work would also be quite helpful asan adjunct text for courses on multilevel modeling. It could serve as astand-alone text for courses that focus on applications andimplementation of the methods… . One of the appealing features of thebook is the use of interesting data sets throughout to illustrate theapplication of the methods. In addition to the data sets used in thetext, many more data sets form the bases of interesting exercisesprovided after each chapter. All of the data sets can be freelydownloaded from a website provided by the authors. Another usefulfeature is the detailed Stata commands for all the results presented,which will allow the reader to easily conduct the analyses on their owndata sets. A strength of the book is the clear and detailedexplanations of how to interpret all the models presented; thegraphical depictions of the models are particularly helpful in thisregard. …
—Brian Leroux (University of Washington), Statistics in Medicine, 2008  

      Product Description
Presenting a thoroughand accessible treatment of generalized linear mixed models, also knownas multilevel or hierarchical models, Multilevel and Longitudinal Modeling Using Stata explains the models and their assumptions, applies methods to real data using Stata, and shows how to interpret the results.    Beginningwith the comparatively simple random-intercept linear model withoutcovariates, the text develops the mixed model from first principles,familiarizing the reader with terminology, summarizing and relating thewidely used estimating strategies, and providing historicalperspective. Once this mixed-model foundation has been established, thetext smoothly transitions to random-intercept models with covariatesand then to random-coefficient models. The middle chapters apply theconcepts defined earlier for Gaussian models to models for binaryresponses (e.g., logit and probit), ordinal responses (e.g., orderedlogit and ordered probit), and count responses (e.g., Poisson). Modelswith multiple levels of random variation are then considered, as wellas models with crossed (nonnested) random effects.
The most complete and up-to-date depiction of Stata’s capacity for fitting generalized linear mixed models, Multilevel and Longitudinal Modeling Using Stata serves as an ideal introduction for Stata users wishing to learn about this powerful data-analysis tool.