New Book

My new book, Understanding Statistics: Basic Theory and Practice, is now available. You can get it here.

This is an introductory text, assumes no prior knowledge of statistics, and doesn’t require calculus. Those of you who already practice statistics will find it rather elementary. Those of you who have always wanted to learn something about the topic but never really got around to it — enjoy!

UPDATE:

Commenters requested a table of contents, which is given below. I’ve also added the table of contents to the “preview” (which includes the TOC and the first 10 pages of chapter 1).

Someone mentioned getting it to brush up on time series analysis. That is not covered — this is really “Statistics 101.” I do have an elementary introduction to time series available here. Don’t let the title put you off, its focus is astronomical data but the methods are quite general and you don’t need to know anything about astronomy. It’s not your usual time series text, it’s geared to physical science data and doesn’t really cover topics like autocorrelation or ARIMA models, but I think it’s an excellent intro to studying time series data, especially for those who’ve never done it before. I’m also working on a graduate-level text about time series (which will include the usual stuff covered in such texts, and more to boot).

And, I have more books in the works, including an Introduction to Fourier Analysis, and a Brief Introduction to Bayesian Statistics.

There are requests for an e-reader version. I haven’t quite figured out how to do that yet, but I’m looking into it.

Contents:

Introduction
Preface: the Value of Data

I. Basic Theory

1. What Is Statistics?
2. Average
3. Histograms
4. Densities and Distributions
5. Probability
6. Dispersion
7. Box and Whiskers
8. Expected Value
9. Expected Values
10. A Miracle Happens
11. Normal
12. Uncertainty
13. Hypothesis Testing
14. Descriptive Statistics

II. Basic Practice

15. Binomial Distribution
16. Uniform Distribution
17. Chi-Square Distributions
18. Multinomial Distribution
19. Student’s t Distribution
20. F Distribution
21. ANOVA
22. Covariance and Correlation
23. Regression
24. Smoothing
25. Testing the Distribution
26. Outliers
27. Non-Parametric Statistics
28. Case Studies
29. Linear Models
30. Bayesian Statistics

StatCoverSmall

About these ads

32 responses to “New Book

  1. Glad to see LaTeX is still used in textbooks nowadays. I will probably buy the book to refresh my 20-something year old experience of time series analysis. Thanks for all your work, Tamino.

    [Response: This does not cover time series analysis. It's really "Statistics 101." I do have an elementary introduction to time series available here. Don't let the title put you off, its focus is astronomical data but the methods are quite general and you don't need to know anything about astronomy. It’s not your usual time series text, it’s geared to physical science data and doesn’t really cover topics like autocorrelation or ARIMA models, but I think it’s an excellent intro to studying time series data, especially for those who’ve never done it before. I’m also working on a graduate-level text about time series (which will include the usual stuff covered in such texts, and more to boot).]

    • Gavin's Pussycat

      I for one am also happy to see LaTeX used… it’s the best for homebrew typesetting, bar none. But it requires a little bit of TLC to make it look great… and looking at the first twelve pages I see a problem already: standard LaTeX has a poor algorithm for placing floats on the page, and there are several pages with just one figure and lots of whitespace around.

      What I use is the commands

      \renewcommand\floatpagefraction{.9}
      \renewcommand\topfraction{.9}
      \renewcommand\bottomfraction{.9}
      \renewcommand\textfraction{.2}

      which should mostly prevent this problem. You’ll also get fewer pages, meaning a smaller price for the dead-trees version.

      Hope this can still be useful.

  2. I tried doing stats at Uni. I must have missed the ‘enjoy’ part. ;).

    [Response: Maybe you just didn't have the right book!]

    • Horatio Algeranon

      Edgar:

      Horatio took a “Probability and Statistics” course (30 years ago) and also somehow missed the “enjoy” part (along with the “Probability” and “Statistics” parts).
      The material was a bit , shall we say, inaccessible (“impeniterrible” might be a better term) — as was the professor, who spent most of his time in his office dealing himself blackjack hands…and in Vegas testing his latest counting method.

      Not sure how he fared, except to note that he remained at the University until he retired (so he probably didn’t do that well, right?)

      Tamino: Congrats on the book. Sure it will be a classic some day (if it isn’t already)

      [Response: Thanks.

      Too many math teachers spend too much time trying to show you how smart they are. They should be showing you how smart *you* are.]

      • “Don’t know much about algebra”–or statistics, but I do know a bit about teaching, and “They should be showing you how smart *you* are” is insightful.

  3. Will there be an e-reader version ?

  4. Congrats, Tamino, on completing and publishing the book. Can you post a table of contents?

    [Response: Done.]

  5. Tamino,

    Congratulations. Book duly ordered. Keep up the good work!

  6. Congrats!

  7. John Cartmill

    I looked at the preview, but did not see a table of contents.
    Could you post one here or on the Lulu site.
    Thanks.

    [Response: Done.]

  8. Looks great, but 60 dollars is out of my price range. Is there a chance of an e-book version some day?

    [Response: I'm looking into it. I know $55.55 isn't cheap, but it's 484 pages with lots of equations and that's inexpensive compared to the usual price for such books. In fact, I looked at the text used for Statistics 101 at a local university, and the list price was $389 -- which I consider to be a bit "over the top."]

    • I just edited a book with fewer pages…and thrice the price!

      Not that I’ll be seeing much of that ;-(

    • IMO, textbook publishing has become a bit of a racket. “Over the top” prices have become the norm, and seem to continue escalating.

      [Response: You're not the only one who thinks so.]

    • Put me down for the e-book when it arrives. I can’t carry 484 pages around with me anymore. Lulu should be able to help you with ebooks – I am doing a similar thing myself (although in a quite unrelated areas). You might also consider putting it on the Amazon Kindle Store or iBooks.

  9. Tamino,

    This is a timely release for me. I am teaching a course in elementary statistics (introductory, non calculus) at the local community college. I purchased a copy which I will place on the reference shelf in our library for the students use. I am sure there will be ways I can use it in the class as well.

    How do you handle the data for the exercises? Is there a disk with the text or websites students are referred to?

    I am looking forward to receiving it as I am sure it will be very helpful to me and to my students.

    [Response: I'm preparing a separate volume with exercises, and with instructions on how to use R for the calculations, specifically so folks can use it as a "Stats 101" textbook. I intend to put data sets on the web, except for those that already are.]

  10. Can you send a free copy to Anthony Watts? On second thoughts maybe two or three!

    [Response: Hmmm...]

    PS I am going to buy a copy. I have forgotten 80% of the stats I learned in university and need to relearn and refresh! This looks like it might be just the book. I’m hopeful that the explanations in the book are as good as what you provide on the site!

  11. For someone with little to no interest in the subject I have to say that ‘The Joy of Statistics’, which might still be available on the BBC iPlayer, was one of the better documentaries I’ve seen recently. Hans Rosling does an excellent job. Believe it or not it was highly entertaining, especially his animated charts such as life expectancy -v- wealth. It was like watching the Grand National.
    http://www.bbc.co.uk/programmes/p00cgkfk

    [Response: I too am very impressed with Rosling.]

    • Warming: gratitious advertisement. Hans Rosling’s gapminder.org allows you to make the same type of graphs he presented in the programme. Now you only need to become as entertaining. Alternatively, you can follow some courses at Karolinska where he has his official day job.

  12. Purchased, look forward to receiving.

  13. This looks great – I hope you’ll followup if/when it makes its way into a kindle-friendly format.

  14. Another vote for an e-book version … !

  15. Copy ordered, I’m not teaching any courses where it would be a suitable text at the moment, but maybe some of my colleagues do. A good introductory text that focusses on the “why” rather than the “how” of statistics is badly needed, there are far too many “cookbooks”, which encourage people to use statistics without understanding what they are doing, which is a recipe for disaster!

  16. Out of pedagogical interest, Tamino – does your book require the reader to learn statistical tables (chi-squared, normal, etc.) or have you skipped them in favour of excel functions?

    [Response: The book does not include tables (like books did when I was a youth), it assumes the reader will get values from some software program. But it does describe those statistical functions -- that's kind of the point, so the reader understands what they are and what they mean. I want the book still to be relevant long after software has changed beyond recognition.

    I'm writing a follow-up which will contain problem sets and more examples, as well as "how-to" use R for solution. I know ExCel has an impressive array of built-in statistical functions, but I find it clumsy as hell. I recommend you learn R -- the time you spend doing so (and it's really not that hard) will be repaid many times over by the time you save using it.]

  17. Stats can be a bit dull, unless you have data that you care about. Much of my research is experimental work that is analyzed with ANOVA. Once I have some data, I really enjoy statistics.

  18. Tamino, thanks for the reply. It’s good to see you’ve written a textbook that moves beyond old-fashioned tables. When I teach stats I have stopped using tables in favour of excel. I use excel because it is ubiquitous, not because I somehow think it superior – I can use R too, but my students are studying for epidemiology so I am teaching them in Stata. Also, I think it’s really important for scientists to be able to use excel with a high degree of competence, even though (I agree) it’s clumsy as hell. In the modern world no company uses any other spreadsheet package, so advanced excel skills are a really good thing to teach students.

    (Incidentally, I’ve had a lot of problems with R, it has serious problems. I’m not a software fascist, I’ll use whatever I’m pointed at – heavens, today I was using SPSS – but I don’t trust R).

    [Response: I've never had any problem with R at all, I trust it completely. Oh well, one man's poison ...

    You're right about ExCel. It's ubiquitous, every company uses it, that it's worthwhile for people to learn. But -- it's not what the additional book of exercises will be about.]

  19. Noting all the noise on contrarian sites about ‘priors’, I’m wondering whether you go into that area at all (in the Bayesian chapter?). I haven’t noticed any blog posts from serious sites (like yours) that discuss the issue and would like to understand what the fuss is about.

    BTW, I enjoyed your book, “Noise”, although it was a little smaller than I anticipated :-)

    [Response: The final chapter (about Bayesian statistics) is just a teaser. This book covers "Statistics 101" for non-calculus, and any attempt at a serious introduction to Bayesian statistics would be out of scope. But I wanted readers to know that it *exists$.]

  20. Okay copy ordered. Next question, any recommendations for a good text on non-parametric statistics for environmental investigations? I’d like to have a text to brush up on my knowledge since university and pass around the office.

    [Response: Good question. There's a text called "Nonparametric Statistical Methods" by Hollander and Wolfe, which in my opinion is good in some ways and not in others. It's reasonably clear and thorough, but hardly what I'd call "engaging." But by the time you get to nonparametric statistics, you're probably beyond the point of being clueless about why this stuff works. But Hollander and Wolfe is "dry" enough that it probably won't get too much attention from being "passed around the office" unless your coworkers are pretty motivated. BTW, there's a reasonable chapter on nonparametric statistics in my book.

    If anybody knows of an outstanding book for nonparametric stats, let us know.]

  21. I’m not currently in need of an introductory stats book, but when I know of those who do, I’ll send them to yours. I’ve learned a great deal here, and I know that you’ve got tremendous skill in explaining complex concepts.

  22. Purchased with glee. You are the best person I’ve ever read who is able to makes statistics comprehensible and rewarding. Statistics = The Art of Extracting Meaning from Data. This is an immense challenge at many levels; the world is awash with data but struggles to agree on it’s meaning. While I always read your posts with interest I’ve long wanted a more formal text to ‘fill in the gaps’ as it were.