2 of 2 people found the following review helpful
This review is from: Multiple Imputation and Its Application (Statistics in Practice) (Hardcover)
Customer review from the Amazon Vine Programme (What's this?)
This is certainly a very comprehensive book on missing data imputation, but one that is equally hard to follow! Like many before me have mentioned, it is not for the wide audience, although one could gather that from the title alone, but even within the circles of multiple imputation (MI), certain people will struggle to come to terms with all the terminology!
The first chapter of the book is a clear, and at the same time highly technical, presentation of the problem and its consequences. This is a must for anyone who wants to at least find out a bit more about the issue and how serious it can be.
The second chapter justifies the process of multiple imputation (again in a technical manner), but also gives some handy tips at the end of the chapter for people who want to implement it in their research.
Chapters 3-5 discuss in detail the theory and practice of MI in a variety of settings (cross-sectional data, binary and ordinal and unordered categorical data), whilst chapters 6 and 7 complete the discussion by bringing in non-linear relationships and the issue of interactions, which is quite common in real life!
The advanced section of the book is for those that are really into MI. I know that I will be making more use of Chapter 8, as I will be training in survival analysis and hopefully will be utilising some large datasets, which by definition are going to include some missing data!
All in all, missing data is something that researchers have to face on a daily basis and this book is a helpful resource in making sure that this can be dealt with effectively.
Oh, and before I forget...unlike other books that simply present the theory and long-winded mathematical expressions (oh, wait, this one does that as well!), this book also comes with example datasets that you can load onto your favourite software (be that SAS, Stata, etc.). What I found quite helpful is that the REALCOM software can now be downloaded and run directly from within Stata (hurrah!) if need be!