Top critical review
4 people found this helpful
Not particularly good. Superficial and not always very clear
on 5 August 2016
Statistics is a very difficut subject, and requires a lot of mathematical knowledge and maturity in order to deeply understand its logic and its many subtleties. Most introductory books for undergraduates tend to greatly simplify the theoretical material needed to master statistics, in order to fit the usual background of students. These books often result in oversimplification of the topics, and the final product is a stream of horrible "introductory" books which are barely intelligible when, for example, they pretend to explain difficult concepts without the necessary theoretical concepts needed and without the necessary rigour.
Having said this, let me make my point about this book. This is an introduction to statistics which is not at the completely rigorous level, and tends to the applications rather than the theory. As such, it manages to convey some statistical content, however, contrary to what many reviews tend to believe, is not a rigorous nor a profoundly written book on statistics. The background required to the reader is rather elementary - just a little of calculus, and no use of the more advanced topics like measure theory necessary to develop statistics at the rigorous level - is made throughout the text. All the proofs are quite elementary and given only for particular cases (e.g., discrete variables, continuous variables), and almost never in the general case. The explanations, however, are often insufficient and unclear. Having to accomodate for an audience of readers without background in measure theory, the authors almost invariaby opt for "shortcuts" when they introduce advanced concepts - and almost invariably fail to provide clear and convincing explanations. For instance, in Chapter 9, which is the technical core of the book. where the basic properties of estimators are introduced, the discussion of sufficient statistics, and of the Rao-Blackwell theorem is awful. In fact, without the introduction of the notion of sigma-algebra, it is virtually impossible to explain what a sufficient statistic really is, and this text fails completely to explain this idea in convincing and clear terms. The discussion following the Rao-Blackwell theorem is unintelligible to say the least: no explanations are provided for the method of finding MVUEs, and the examples in that section only serve to apply th recipe. Also, I found the rest of this chapter very poorly written.
As I wrote before, this book is definitely practice-oriented, and there are many exercises. However, most of the problems, after the insufficient level of the theoretical part, seem just some mechanical applications of a set of rules rather than a real test for understanding of the material. Therefore, most of the exercises are just variations over a single type of problem, which is okay if you want a "hand" in doing things, but is absolutely useless if you wish to understand in depth your subject of study. Just do a couple of problems for each section, and that will be enough.
In conclusion, this text falls in the category of "statistics made easy for undergraduates who must know how to apply things rather than understand how and why things work", and it is fairly okay if your aim is that. But is useless if you expect to understand and master the concepts of Mathematical Statistics the proper way. Not recommended.