# What book would you recommend for non-statistician scientists?

What book would you recommend for scientists who are not statisticians?

Clear delivery is most appreciated. As well as the explanation of the appropriate techniques and methods for typical tasks: time series analysis, presentation and aggregation of large data sets.

-
 Could you be a little more precise? What type of analysis, in what context, etc. – dominic999 Jul 13 '11 at 5:38 Well, I'm talking about the basics, an overview of as much as is possible. – Fergus Barker Jul 13 '11 at 7:08 Statistics explained covers the basics using examples from the life sciences. The answers to this question may also contain recommendations that you'll find useful. – MånsT Jun 15 '12 at 13:59

Peter Dalgaard's Introductory Statistics with R is a great book for some introductory statistics with a focus on the R software for data analysis.

-

The answer would most definitely depend on their discipline, the methods/techniques that they would like to learn and their existing mathematical/statistical abilities.

For example, economists/social scientists who want to learn about cutting edge empirical econometrics could read Angrist and Pischke's Mostly Harmless Econometrics. This is a non-technical book covering the "natural experimental revolution" in economics. The book only presupposes that they know what regression is.

But I think the best book on applied regression is Gelman and Hill's Data Analysis Using Regression and Multilevel/Hierarchical Models. This covers basic regression, multilevel regression, and Bayesian methods in a clear and intuitive way. It would be good for any scientist with a basic background in statistics.

-

I'm going to assume some basic statistics knowledge and recommend:

-
The Statistical sleuth is used a the textbook on that great introductory course (there are 64 lectures in total) video.google.com/… – gd047 Jul 27 '10 at 18:36
The first recommendation seems appropriate but Flury's book in not concise and deals with a special topic and is not a general introductory book on statistics. – Michael Chernick Jun 15 '12 at 15:54
Oddly, the Flury book was linked to Jerry Dallal's excellent & free online test, The Little Handbook of Statistical Practice. I have corrected that, but leave the link to LHSP here--it's certainly a book I can highly recommend & it may be that @ars had intended to include it as well. – gung Jul 28 '12 at 3:04

Statistics

David Freedman, Robert Pisani, Roger Purves

Fourth edition: 2007, First edition: 1978

As an undergraduate studying philosophy, I was asked to analyze some data for a small study that I was working on with a physician. Needless to say, I found myself somewhat overwhelmed, but was able to get by by mimicking some old Stata code that a biostatistician friend had given me. The analysis turned out to be good enough to help get the study published, and I had suddenly become interested in this curious field of study called statistics.

The first book on statistics that I read was Statistics, by David Freedman and colleagues. What I liked most about it was its focus on explaining the fundamental concepts of statistical analysis (what do p-values actually mean, why is it important to visualize data, what does it mean for a test to be significant, etc) with concise and accurate language, but without too much mathematics. With that conceptual background, I found it much easier to go on to read more advanced literature with more advanced mathematics.

This book covers all topics covered in a first year statistics course, but does not cover time series or aggregation of large data sets. I feel it does a very good job at teaching a non-statistician how to think like a statistician. From there, adding new methods, like time series, should be relatively easy, and the non-statistician should be well on his way to becoming a life-long student of statistics.

-
+1 As someone who first learned stats from this book (right after the first edition came out :-), who has taught from it, and who has given many copies away to colleagues and clients, I can warmly recommend it. The exposition is marvelously clear. There's nothing outdated about the first edition, by the way: the third and fourth editions use more recent datasets but are otherwise about the same. This means the first (and now the third) editions are real bargains. (The second edition's ok, but it started to veer off course a little, a mistake that was corrected in subsequent editions.) – whuber Jun 15 '12 at 20:50

-

Statistics in Plain English is pretty good.

4.5 on Amazon, 11 reviews.

Explains ANOVA pretty well too.

-

A lot of Social Science / Psychology students with minimal mathematical background like Andy Field's book: Discovering Statistics Using SPSS. He also has a website that shares a lot of material.

-

Not intending to plug my book but it does seem to possibly apply. Last year I published a book with Wiley titled "The Essentials of Biostatistics for Physicians, Nurses and Clinicians". It is paperback and fairly concise 214 pages in total. It has the advantage for you that it emphasizes topics that are important in biological applications but may not be quite as concise as you would like to have for a 10 day self-learning course. "Introductory Statistics for Biology Students" 2nd Edtion by Trudy Watt and published by Chapman and Hall/CRC 1997 is another paperback that might be right for you. It is a little simpler than my book but does not include survival analysis which I consider to be a very important topic in biological studies (particularly clinical trials). Her book is 236 pages. I would also like to mention "The Cartoon Guide to Statistics" by Gonick. A humorous book but it also covers basic concepts very well and is exceptionally easy to read.

-
@Procrastinator. Thanks a lot. I really do apprciate your edits and those of MAcro, Huber and others. I will try to learn more and relieve some of the burden. – Michael Chernick Jul 24 '12 at 19:07

The Flaw of Averages by Sam Savage.

-
 A nice book and philosophical but not what the OOP is looking for. – Michael Chernick Jun 15 '12 at 15:30

The Drunkard's Walk: How Randomness Rules Our Lives by Leonard Mlodinow is an excellent book for laypeople. Enjoyable and educational.

It might not be a textbook, but it makes you think about the world in the right way.

-

So many wonderful recommendations! It's not quite what you asked for, but How to Lie with Statistics is short and quite wonderful. It doesn't directly teach the things you want, but it does help point out violation of assumptions and other flaws.

-
 This could be helpful but not an introductory text on statistics. – Michael Chernick Jun 15 '12 at 15:30
-
Good recommendations here but the first two are not concise books. – Michael Chernick Jun 15 '12 at 15:27

Probably the best basic, get the big picture / ideas book is going to be:

Robert Abelson's Statistics as Principled Argument

-

"How to Tell the Liars from the Statisticians" by Hooke. I am fond of its way of explaining the concepts of statistics to laypersons.

As for explaining the motivations of statisticians, "The Lady Tasting Tea" is good reading.

-
 I think these are both very nice books that would make excellent reading for the OP. But neither is a concise introductory text on statistics that the OP is requesting. – Michael Chernick Jun 15 '12 at 15:43

It is a bit old, but I have found Chris Chatfield's book,

Statistics for Technology: A Course in Applied Technology

to be an excellent introduction.

It was how I first learned about statistics from a conceptual point of view.

-
 only 3 reviews on Amazon...sketchy – Neil McGuigan Dec 15 '10 at 20:32 That's all? I'm surprised, cos it is a really good book. The same author wrote an excellent introduction to Time Series Analysis as well. – kaybenleroll Jan 5 '11 at 19:41

As a first introduction to the topic i liked Data Analysis: A Bayesian Tutorial.

For a deep and philosophical discussion of the underlying ideas of quantitative scientific reasoning i recommend Probability Theory: The Logic of Science. This book does not serve as a good introduction, though. It's only recommended for persons who want to know, why bayesian statistics is the way it is and/or are interested in a historic review of bayesian statistics.

-
 I respectfully disagree regarding your second recommendation. Though certainly an interesting book, it is almost surely not the place to start for a non-statistician. Especially someone concerned about clear delivery. – cardinal May 2 '11 at 3:43 You're right, it's certainly not a good first read on the topic and not written in a concise way. Still, i thought for a scientist interested in how a consistent way of doing statistical analysis might be built, it makes an interesting read. But definitely not an introduction. I'll update my answer to reflect that. – Thies Heidecke May 2 '11 at 4:17

For the rudiments of statistics: http://www.bbc.co.uk/dna/h2g2/A1091350 and http://www.robertniles.com/stats/

For a good guide to data visualisation: http://www.perceptualedge.com/ - in particular, try the Graph Design IQ test at http://www.perceptualedge.com/files/GraphDesignIQ.html (requires Flash)

NB these are orthogonal - there are lots of statistics experts who are terrible at data visualisation, and vice versa.

-

I recently found Even You Can Learn Statistics to be pretty useful.

-

I have recently had this website pointed out to me. It covers a number of books useful for new statisticians, with some targetted discussion of their various strengths and weaknesses, and a summary right at the bottom.

-
 This discusses a lot of the books listed. It is a very nice link for the OP. It doesn't mention anything about books specifically directed to biologists though. – Michael Chernick Jun 15 '12 at 15:51

I strongly recommend "Statistics for Experimenters: Design, Innovation, and Discovery , 2nd Edition" by Box, Hunter and Hunter. Must-read book for any scientist doing statistical analysis of their experiments. There's a companion R package (BHH2) as well.

-
 This is a very good text but is not concise and emphasizes design rather than provide a general introduction to statistics. I think that many of the responses are getting away from the original question from the OP. – Michael Chernick Jun 15 '12 at 15:47

"Theoretical Statistics"
Keener, Robert W.
1st Edition., 2010, XVII, 538 p.
Hardcover, ISBN 978-0-387-93838-7

-
-1 How is this introductory and how does mentioning it answer the question? – Michael Chernick Jun 15 '12 at 15:48
@MichaelChernick Did you ever read the question? It says: "What BOOK would you recommend for non-statistician SCIENTISTS?" Downvoting something just to "vote" isn't really logic. Besides, if you would have read the book, you would know it takes a pretty scientific look at statistics. Knowing the book (among a truckload of others) and having read the question (in contrast to you) convinces me that my answer validates as correct and may be helpful in regard to the OP's request. Anything else I can explain to you? Tsss... – user7997 Aug 3 '12 at 1:30
I am sorry that I upset you by my downvote. I thought it was appropriate to explain the reason for the downvote. The OP wants an introductory book and not something on theoratical statistics. If you don't answer the question you deserve a downvote. Also if it were a good book for beginners you should explain why. I gave an answer on June 15 and recommended 3 books appropriate for biologists physicians and nurses with an explanation as to why they are useful. The cartoon guide to statistics is very good for general audiences. – Michael Chernick Aug 3 '12 at 4:55
But if you think that inspite of the title the book is valuable to scientists and add it to your answer I will retrack my downvote. My answer indicates that I did read and properly respond to the question. – Michael Chernick Aug 3 '12 at 4:56
When I took Keener's theoretical statistics sequence in grad school this book hadn't been produced so we still used a "working" version of the book - it was very challenging but it was great. I think it approaches the subject a bit differently from from some of the "standard" theoretical statistics texts (e.g. Casella and Berger, Lehmann) and I think I prefer it. To be sure, this book requires a strong math background, although it is largely self contained. I'm glad to see its out now - I'll be sure to get a copy. – Macro Aug 3 '12 at 11:31

If you're to use SPSS, I'd recommend this book: Data Analysis for the Behavioral Sciences Using SPSS by Weinberg & Abramowitz. It is very well written and accessible. Note that it doesn't cover time-series, though.

-

That'll depend very much on their background, but I found "Statistics in a Nutshell" to be pretty good.

-
I found Statistics in a Nutshell to be seriously flawed in terms of wrong/missing figures, mistakes in formulas, bad explanations and the book doesn't even have tables for critical values. This is especially bad at places where the authors write "and since the critical value for this is foo, this is significant", leaving the reader totally unclear about where this foo value comes from. The book does have a good intro section but should be edited eventually to make it good. Just look at the errata page for the book and be stunned at all the errors. – xmjx Oct 17 '10 at 9:07