Customer Reviews

Mathematical Statistics (Chapman & Hall/CRC Texts in Statistical Science)

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

 

 

With so many good graduate texts available to choose from I think it is more helpful for me to tell you what this offers that some of the others do not and then you can decide for yourself if it is for you. It is not the best written text and is highly mathematical. It does introduce some modern concepts such as empirical likelihood,the jackknife, and the bootstrap and devotes a chapter to the important topic of generalized linear models. However, the bootstrap is only mentioned briefly as a resampling tool that can be used in the context of generalized linear models. Its more general use as a competitor to the jackknife for estimation of variance or standard errors and in the context of confidence intervals is not mentioned.

An avantage of the book for some is that it includes chapters on probability theory. It only includes the topics needed to understand distribution theory for inference purposes and to be able to understand asymptotic distributions. To achieve this Knight covers the basic rules of probability and its measrue theoretic basis in chapter one. Inportant convergence results needed to develop the asymptotic theory of statistics is cover in Chapter 3 where the concepts of convergence in probability and convergence in distribution are given along with key results such as the weak law of large numbers and the central limot theorem. This is a nice feature as it makes the book self-contained and the student does not need a book or separate course in advanced probability since the essence of that material is covered in Chapters 1 and 3.

I do applaud the author for covering important topics such as generalized linear models along with the standard parametric theory of hypothesis testing and estimation. Another topic not commonly covered in the affect of model mispecification on results.

So if you are interested in learning about the jackknife, empirical likelihood and the Bayesian approach, this is one of very few advanced books that covers all these topics and still covers probability and standard statidtical theory (e.g. Cramer-Rao inequality, Lehmann-Scheffe theorem, Neyman Pearson lemma, uniformly most powerful tests, Basu's lemma, and sufficiency and efficiency concepts.

I used this book in a class that I took for intermediate statistics, a mid-level graduate course in the statistics department at a top university. This is an advanced book on mathematical statistics. In general, I found that if I had significant exposure to a topic previously, I could easily understand the text. Sections of the book containing topics new to me, however, were difficult to understand. The biggest failure of this book to me is that it serves as a poor reference tool. Definitions, theorems, and important concepts are often hidden in paragraphs nestled between examples. There is no appendix or list of distributions, moment generating functions, etc. The index is incomplete. A second negative aspect of the book is that there are many typos. The extent of the typos often increases the difficulty of understanding the text.

Overall, this book should never be used for an undergraduate course. I also do not recommend this book for use in a graduate course, unless some later edition with heavy revisions is released. I strongly recommend not buying this book for personal use as a reference tool.

This book is very challenging for a beginner student in statistics. I am not sure how it would be if one had some preparation in the subject, but I still think that it should rather be used as a source for excercises and some concepts. However, using it can be a little frustrating sometimes. Let me give you an example: If you want to do some exercise which requires you to know a density function of a certain random variable then you have to look somewhere else because this book fails to provide you with a table of most common distributions. Also there are a lot of typos. Some exercises overcomplicate things, e.g. they are not about understanding a concept but giving you hard calculations to do. OK, as far as positive things about it: Some examples are good, definitions are precise, and most of exercises will make you think. Overall, I would recommend it only if you have mastered the basics. It would probably be useful with an excellent teacher who would explain the concepts to make you look at the material from a different point of view.

I checked several statistics books(Casella, Degroot and Wasserman), but this is the best one. Easy to read, and comprehensive. Graduate students who are not so good at math can read comfortably.

While statistics is a very interesting subject and can be taught in many interesting ways, it is by no means taught in any student friendly way in this book. To begin with, this book is about as interesting of a read as its cover design. It starts off well with some review of general concepts, but then as you go on, things that could and should be explained are skipped using phrases such as "it could be clearly shown that..." or "it is easy to see that...". Unfortunately, to the average student of statistics, not everything is trivial. Aside from that, there are many typing errors and by that i mean A LOT. This book can be greatly improved first with the fixing of all those typos, and second with better explanations and less assumptions of "clearity" to the reader. Also, the discussion should be made less dry by perhaps relating what is taught to some familiar concepts. Overall, unless you've been studying statistics for a while, dont expect to know what you're reading at all times.

This book does not, as the editorial review claims to be "a balanced, practical text that brings life and relevance to a subject so often perceived as irrelevant and dry". Rather, it makes what is already irrelevant and dry as even moreso. Not only that, but there are many typos that only confuzzle the reader even more. It reads as though it was written by a robot. There is no human touch to this text at all. This book is "clearly" and "obviously" inconclusive and I for one would love to never see it again.

Maybe it would be easier to read and comprehend if Keef used the cartoon format for his statistics textbook.