Customer Reviews

Statistics, Third Edition

5 star:		(26)
4 star:		(6)
3 star:		(3)
2 star:		(1)
1 star:		(4)

 

 

If you are looking for a good (non-technical i.e. not involving higher math) introduction to statistics, this is the book for you. As a TA at Cal (Berkeley..Go Bears!), I worked for Roger Purves (one of the authors of the book) and I taught out of this book. Needless to say, I got to know the book rather well. I highly recommend it.

However, if you are looking for a slightly more advanced introduction to statistics/probability, I would suggest something along the lines of Probability by Jim Pitman or Mathematical Statistics and Data Analysis by John Rice. If you are really serious about probability theory, you might want to try Statistical Inference - by G. Casella and R.L. Berger.

Many introductory statistics texts suffer from one of two ailments. Either they incorporate too much mathematics for non-statisticians or they provide oversimplified and sometimes incorrect explanations. This text is excellent and is favored by many statisticians who teach the introductory service course for non-statistics majors. The book provides excellent and insightful explanations. It is written by well-known Berkeley statisticians with great theoretical and applied experience, so it is not oversimplified or inaccurate. On the other hand Friedman and his co-authors took pains to minimize the necessary mathematics. It covers all the topics one would want to include in a first course. Real examples are used throughout to illustrate the value of the methods. These include clinical trials and observational studies, telephone surveys and opinion polls and some models in genetics.
Discussion of the data snooping issue is important, particularly as we move into an age where data mining is now feasible with current computing power.

I had used this book for my Introduction to Statistics class, and I think it was the best $60 I had ever spent. The authors explain all the major statistical concepts clearly and succinctly, drawing on a variety of samples and adding a touch of humor. The math in this book is more than doable; anyone with a basic grasp of algebra and a willingness tackle numbers is the only prerequisite to understanding this book. However, the authors also try to convey to the reader that there is more to statistics than just numbers. How an experiment is constructed, how polls are taken, what biases exist, and how assumptions are made are all integral parts of statistics. This book is applicable to almost every subject--political science, sociology, sciences, engineering, etc. There are enough exercises in the chapters to assist the reader in reviewing the concepts. I can't stress this enough: Buy this book and you won't regret it!

If you buy just one book on statistics, this should be it. If mathematics scares you, get this book. If you are interested in how and why statistics works, get this book. If you want to improve the way you teach statistics, get this book.

For people first studying statistics, note that this book is written in ENGLISH! All formulas are written in English, not arcane mathematical symbols. For example, the formula for the arithmetic mean or average is:

The average of a list of numbers equals their sum, divided by how many there are.

That's it, no summation symbols and no variables with subscripts. The average is also described as:

Average of a list = sum of entries divided by the number of entries

The standard deviation (SD) is described:

SD = square root of (average of (deviations from the average)^2 )

(A deviation from the average is just the number minus the mean for the entire set of numbers. I've used "^2" to represent "squared" or "raising to the second power".)

The book is both easy and enjoyable to read. It is interesting reading and not just for statisticians. You get to read about important applications of statisitics in the real world (often including relevant historical details). There are also very well thought-out excercises that are realistic and yet can be easily computed by hand.

When I first found this book, I had finished by Ph. D. and had taught statistics for a number of years. Even though this is an introductory text book, I still learned a lot! It actually explains many important concepts that are often buried in the mathematics of other books. (For example, how many students understand the concept of "regression toward the mean"?) It completely changed the way I taught statistics. Especially when you are first starting to study statistics, you don't want the mathematics to obscure the statistical concepts. I've seen far too many students being able to do much of the mathematics but not having a clue about the statistical concept behind the method. They could do the computations but wouldn't know why they were doing them or when the method was appropriate to use.

The book consists of 29 chapters and covers design of experiments (comparative experiments), descriptive statistics (histograms, mean, standard deviation, normal distribution), correlation and regression, probability, chance variability (expected value and standard error), sampling (surveys, chance error), chance models (measurement error, genetics), and tests of significance (large sample tests for the mean and proportions, t-tests, and Chi-square tests).

This book is a rare gem. You can find piles of books with Statistics symbols/equations and hard-core problems, but how many of them really teach you the meaning of what you're doing?

Statistics is a kind of data-compression - you start off with with a bag of data and you extract certain "features" such as averages, standard deviations etc... ...this allows you to say general things about the entire dataset (avg/SD, etc.) or claim associations between multiple datasets with varying degrees of confidence (correlations) or even predict the value of one variable if you know the other (regressions).

The dangerous thing is, if you are not careful about how you "compress" this data or about what you do or don't do with the dataset (like dealing with outliers), your conclusions may be ENTIRELY INVALID! By using specific examples, this book teaches you to look at what are you doing before you do the analysis and then how to look at your results after you do your compression (running statistical studies).

I was reluctant to buy this book at the bookstore at first, but after having read the entire text, I believe this investment was money well spent. If you don't believe me, check it out in a bookstore before you buy the book.

Good luck!

D. Messer asks how his/her review can be so different from Alice's. That's easy. They're both right. They are looking at the same book through different lenses.

If the goal is to explain what's really going on "under the hood" in an introductory class, there's no better book than FPP. However, it works its magic by discarding things that don't matter--like mathematical notation. I would argue that students who take a course using FPP would be able to successfully tackle problems that would cause others to shrug their shoulders from lack of comprehension.

However, it comes at a price. Students do not see the mathematical notation that is necessary to pursue the subject at advanced levels. Imagine, if you will, being an instructor with a class of students who had completed an introductory course with excellent grades and wouldn't recognize the formula for Student's t statistic in standard notation.

I've used FPP in a university setting. It works fine, BUT there has to be careful coordination with teachers of follow-up classes so that they can adapt their methods accordingly, that is, for a class who has a deeper understanding of the principles of statistics without having been exposed to the usual formulas in the traditional way.

I've just finished studying this book. It's just absolutely delightful. Aftger having taken brain numbing statistics courses in graduate school, this book is like having an expert friend to talk to about the real basis for things. The authors are very thorough in developing baisc statisical theory through examples and practical problems, not to mention interesting and relevant historical background. It's basically a book on learning how to think statistacally, correctly! Common pitfalls togehter with discussions of famous and not so famous goofs and misapplicatinos of statistical methods are throughout the book used (not to poke fun, although it is fun) to develop a second nature in basic concepts. While the book is thick, the reading is easygoing and friendly. It won't take very long for most people to get through it. Concepts are developed progressively on firmly developed and well explained basic ideas. It's as much, if not more, a book on critical thinking as it is on the techniques of elementary statistics. -- Jack Penkethman

I am not a statistics major, nor do I tend to excel in mathematics, but I am capable of achieving if I put enough energy into a subject...that wasn't the case with my stat class last semester, which used this textbook.

This book takes the role of a friendly teacher who dumbs down the material so we "not-so-mathy" students can understand what's going on. The problem, however, is that this book speaks in riddles, teaches in examples, stories, and fake conversations between mathematicians of the past, and doesn't spell out in any clear way what the method is for solving certain types of questions. Also, after using this text for Stat I, I moved on to take Stat II and was pretty lost. In Stat II, they use "scary" language such as p, q, n, instead of "big number" "small number" and "box" (which is used in this text). I found it was much easier for my mind to grasp the consistency and methodology of statistics when using a different, more "advanced" texbook.

Stat can be a very difficult thing to understand when you're treated like a baby. Even my TAs hated this textbook, which kind of says a lot (mainly that a lot of students are confused, and can't get any help from outside tutors who don't speak the pseudo-stat language of this book)

I would not recommend this text to anyone. If you're thinking about taking a Stat class where this text is used, you'd be better off waiting a semester until you can enroll in a class where the teacher values actually learning statistical language.

I expected this to be dry and mechanical like lots of other math texts - too-technical proofs, homework questions irrelevant to the material, insufficient explanations for why things are the way they are. This book really surprised me because it wasn't "mathy" at all. It doesn't just throw proofs at you expecting you to wade thru page upon page of math notation until you understand - it gives you the intuitive side of important concepts, which means you only need common sense, not an intensive mathematical background to get everything. The examples they picked simplify rather than confuse the concepts. Each easily and naturally leads to the next. If there's anything not thoroughly elaborated, they were sure to cover it in the homework questions, by gently prodding the reader along towards the answer step by step instead of smacking them in the face with impossible problems. Homework questions supplement the material perfectly and basically leave you with a full and well-rounded impression of what the concepts mean as well as when and why to use them, not just how to plug numbers into some formula.
If anything, I'd say this book errs on the side of caution in that in some sections it could pick up the pace a little. But then again, you could always just skip the easier parts.

I taught an introductory statistics course with this book two years ago. I have to say that *I* learned a great deal preparing for class as I read it--there is a lot of insight and intuition here that you won't generally find anywhere else. Teaching out of it is tough, though, because you don't have the math to hide behind. For those of us used to math, formulas can be a comforting thing. For most students, they're usually just intimidating and the object of many blank stares. IMO, for students who will take only one class in statistics, learning out of this book would be very helpful in a way many other books would not be. For students who will take more than one, gaining a strong conceptual foundation will be helpful as well.

My guess is that the students complaining about this book don't know how good they've got it. You could be stuck with a book that focuses on how to do statistics with Excel or the like, in which case you'll basically learn nothing of subsequent value. :)