<< 1 >>
Rating: 
Summary: Book Description
Review: A classic, time-honored introduction to the theory and practice of statistics modeling and inference-revised to reflect the changing focus of Statistics and the mathematics background of today's students. Coverage begins with the more general nonparametric point of view and then looks at parametric models as submodels of the nonparametric ones which can be described smoothly by Euclidean parameters. Although some computational issues are discussed, this is very much a book on theory. A second volume treating more advanced topics is in preparation
Rating:
Summary: Great Graduate Guide
Review: As a graduate level book, it renders main results in statistics at an advanced level readily accessible. Topics range from estimation theory to asymptotic analysis. Ideal for mathematical statician.
Rating: 
Summary: excellend introduction
Review: Bickel and Doksum is of course a standard text in mathematical statistics. The development of the theory is thorough but not as complete as books by Lehmann or Cassella and Berger. I felt that Bickel and Doksum occasionally left out important theorems and details. Also, the examples given were sometimes unhelpful or irrelavent. I personally disliked their penchant for the hypergeometric distribution and Hardy-Weinberg equilibriums. Nevertheless, the writing is clear and the book gives a solid background of the fundamentals. Personally, I prefer Casella and Berger for a more complete overview.
Rating: 
Summary: Not as good as the 1977 edition
Review: I'm surprised of the many positive reviews for this book. Matter of fact is, THIS IS NOT A GOOD BOOK, whether it be used as a graduate text or as a reference book. I tried to use the book as a reference to my statistical theory course.
It came as a shock to me that during the discussion of UMVUE's the is no reference (I didn't find it anyway and the is certainly no mention of it in the index) to complete sufficient statistics. How can that be? Also, I'm not sure if there is discussion of equivariance in the book for example. I mean I saw it mentioned, no definition though, and no mention of it in the index. These points are my arguments for it to be the wrong book to use as a textbook.
The style of the book makes it hard to read too, and I say this NOT BECAUSE IT IS AN ABSTRACT BOOK, but because it is poorly referenced to ITSELF on past arguments. The index is almost useless. Oh, and the typos...........LOTS OF TYPOS. These facts affect the use of the book as a reference.
Conclusion, the 1977 edition is a much better textbook. This one has its good points (such as choosing a combination of a sufficient statistic and an ancillary statistic without loosing any information from your sample. A handy trick for some estimation problems, like say mixtures) but they are to few to supress the weaknesses. For contemporary courses the Casella and Berger book is widely used and although not a complete textbook either, (you would need references for it too) it is much more useful than Bickel and Doksum.
Rating: 
Summary: A beta version of half a book
Review: If you're looking for a single reference on math stat, this isn't it, because it's vol 1 of a 2-volume set. Vol 2, which will include for example most of the material on non-parametric statistics, is not planned to appear until 2003. Possibly by then a new edition of vol 1, correcting all the typos, will be available, and then the set may be worth considering. The proofreading of the current edition of vol 1 was so sloppy that it cannot be recommended. The authors didn't even bother to run the text through a spell checker ('terrabytes'). The index is a bit skimpy, too.Until the book is completed, most of the errata corrected, and (dare one hope?) the index improved, better choices would be 'Statistical Inference' by Casella and Berger, or even Bickel and Doksum's earlier single-volume text!
Rating:
Summary: Great Book for your whole life!
Review: This is definitely not an easy book, and there're typos in this book, but all of this can not stop this book to be an excellent book.
My professor has been working with this book for years, and he has published some solution for this book, but he still does not think that he has solved every prolem in this book completely or perfectly, and he can still find some interesting solutions from students. This book is just so illuminating.
This book would serve as an excellent reference and also the textbook.
Rating:
Summary: This book couldn't even serve as toilet paper.
Review: This is one of a number of good first year graduate texts on statistical theory. It was used at Berkeley for their students in the late 1970s and early 1980s. These authors put together many interesting and challenging exercises at the end of the chapters. However they did not provide solutions to any of the problems. When Marc Sobel was a graduate student, his father Milton, a statistics professor convinced Marc to work out every problem in the book! Marc did this and eventually he and Milton put together a solution manual which was published. In the process a number of mistakes were caught and corrected.
If you get the book try to get the solution manual as well. It will greatly deepen your understanding of the material and help you through the difficult problems. My review pertains to the original book by Bickel and Doksum that was published in 1977 by Holden-Day. I was under the misimpression that Prentice-Hall was publishing a reprint of the old book due to its popularity and the non-existence of the original publisher. I was apparently mistaken as the title indicates Volume 1 of the Second Edition. This was pointed out to me by a reader. In general if you have doubts check with the reviewer about the edition for books with multiple editions. Often reviewers write reviews for a particular edition and when the new edition comes out it accidentally gets moved. This can happen when amazon removes the page for the old edition because it no longer carries it. Reviews of old editions can still be helpful since the heart of the book usually remains the same and the quality of writing of the authors does not often change much. Of course a review of the new edition would be better if it has information on changes and additions and any particularly attractive new features of the book.
Rating:
Summary: An important reference
Review: This new version will become a classic. Like its predecessor, the 1977 edition, it is an important reference that is at the forefront of contemporary statistical knowledge and will "stay young" for many years to come. The book gives a rigourous and carefully detailed presentation of estimation and testing, in the univariate and multivariate setting. Moreover, it presents both the frequentist and the bayesian viewpoint, covers asymptotics and deals with algorithmic issues (e.g. the EM algorithm). On top of theory, practical issues are raised all over the text; the many examples and problems are very relevant for applications. The curious reader can learn the "why" of many popular statistical procedures of much use nowadays, like logistic regression, e.g.. Any statistician with quantitative background should have it. This a top statistical text, and I have used it very often, both for my classes and for my own research.
<< 1 >>
|
