Customer Reviews

Mixed Effects Models in S and S-Plus

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

 

 

Mixed effects linear models are very useful particularly in medical research (e.g. device or drug trials). Pinheiro and Bates provide comprehensive coverage of both linear and nonlinear mixed effects models with many applications. Implementation is illustrated using the S programming language and the software package SPlus.

Bates is an expert on nonlinear regression and hence the emphasis on the nonlinear models as well as the linear ones. These models are very useful for handling repeated measures data with missing observations. Such data often arise in clinical trials and these models have been used to do the intnt to treat analysis that is often required in regulatory submissions to the FDA, Also some variables are quite naturally modelled as a random effects component in the model.The specific clinical site for investigators in a multi-site trial is one common example.

Even though the title of this book is mixed effects models in S and S+ but this is a wonderful book for a person to learn mixed effect models in R. If you read this book carefully and also use the R to practice examples. Then you will get a lot from the learning process. Of course you should has a basic background in linear model before you read this book.

I strong recommend this book to whom needs nonlinear mixed models of longitudinal data in R.

Every statistician should has this book.

At first sight, there are a lot of SPlus/R commands in the book which one may expect to learn a lot about using nlme. However, I found there is a lack in explanation of the command, if not missing. For e.g., in Chapter 1, the book talks about nested classficification models and gave the command in Splus/R, with the model equation right in front of me, I still can't figure out why in the command ...... random=list(Dog=~day,Side=~1) .... can't figure out the logic of this command in relation to the equation. I know this is not an introductory book for R, but a lot of time, when we want to use R or Splus the first time, it's not b'cos we want to do simple statistics, so a bit more explanation of the commands will be helpful, rather than following the commands blindly. Furthermore, I'm not even talking about R programming. Having said that, I still want to emphasize it is a good book written for the topic and package.

I have deducted one star from my review as the Kindle version has not been properly proofread and so (for example) lme is variously rendered as 1me and Ime in the text, although the actual code reproductions appear correct. I have been running the code in R and have not noticed any errors.

This is a useful book for using the nlme and lme4 packages in R, as it covers the theory of mixed effects models and provides practical examples of their analysis in S. The code can be used in R, as I have been doing, although the output may differ somewhat from that provided in the book.

If you wish to learn about mixed effects models, and in particular if you are using R, I recommend this book.

It is difficult to understand who this book is written for. The authors are clearly clever guys but strangely the book does little to explain the command structure for modelling in the authors' own package! It would be useful to explain the models in the context of random effects/nested modelling in conventional ANOVA for newbies. Please authors if you do a second edition show how really simple models are programmed.

A good example of its beginner unfriendliness is the first example(!) to have a vector within a dataframe of exactly the same name. Not a good idea in a text book. You would have thought the editor would have had something to say about that.

The book has excelent presentation (theory and practical), overall a lot applications with R (my favorite)...If you want to be update in applied statistics, in my opinion, you should have it...