Analysis of Longitudinal Data

Rating:
Summary: the long awaited second edition
Review: The second edition is much like the first and is at least a year behind the original schedule. See my review of the first edition to understand why this is a classical. The promised advances in missing data are included and a new author Haegerty has been added. The missing data chapter is three times longer than in the first edition. They cover what they promised. They also mention some of the econometrics literature including the work of Nobel Laureate James Heckman but admit in the preface that they do not know that literature very well and hence do not cover it in depth.

In the past two years Verbeke and Molenberghs have produced a highly competitive book that deals in detail with pattern mixture models and other missing data methodology but curiously Diggle et al. do not reference it even though they do cite some of Molenberghs work.

Rating:
Summary: the long awaited second edition
Review: The second edition is much like the first and is at least a year behind the original schedule. See my review of the first edition to understand why this is a classical. The promised advances in missing data are included and a new author Haegerty has been added. The missing data chapter is three times longer than in the first edition. They cover what they promised. They also mention some of the econometrics literature including the work of Nobel Laureate James Heckman but admit in the preface that they do not know that literature very well and hence do not cover it in depth.

In the past two years Verbeke and Molenberghs have produced a highly competitive book that deals in detail with pattern mixture models and other missing data methodology but curiously Diggle et al. do not reference it even though they do cite some of Molenberghs work.

Rating:
Summary: Excellent, highly recommended!
Review: This book was written by three very prestigious authors, two of which work at The Johns Hopkins University(Dr. Liang and Dr. Zeger), and Dr. Diggle, who is working in England. These three are very well known and respected characters in their field of work, and this book is an excellent reflection upon the research and work they have done over the years. Watch out! the key word is: GEE

Rating:
Summary: already the classic book on longitudinal data analysis
Review: When this book came out in 1994 there was a great need to look differently at clinical data on subjects. Typically such data would have repeated measurements over time for many subjects but for only a few time points (say three to five). Standard analysis of variance methods do not properly account for within patient correlation between measurements. Time series analysis generally is good for treating long series (but usually only one or a few). In the clinical setting we often are considering hundreds of patients over short time intervals. This book is clearly written for intermediate level statistics students.

The field is important and rapidly developing. Though slightly dated the book is still an excellent introduction to the subject and a very good reference. However, a second edition is in the works and should be out in about one year. I recently took a short course from the authors and I know that the second edition will have some nice features including the latest advances for dealing with missing data and ways to combined the information from time to event data with the repeated measures data. It may be that if longitudinal data analysis is important to you, read the first edition at your favorite university library and save your money for the second edition.

The book includes some nice treatment of the important but often neglected topic of sample size determination.