Best way to re-learn statistics?

October 23, 2008 7:26 PM

Best way to re-learn statistics? Enough that I wouldn't be lost in an advanced, fairly data-intensive psychology class. I took the course at Ga. Tech some years ago, but my old textbook is sooo boring and awful.

I'd love a shiny skinny book with actual math that was relatively easy to follow. I'm hoping to review over Christmas break. Book is better than web site.

Naturally I've forgotten most of calculus already, too, but will re-learn if needed. Can anyone offer any guidance?

I seem to remember that the stats class I took was mostly memorizing what different terms / distributions meant. I want to know more than "a coefficient of correlation ranges from -1 to +1".

No, I haven't searched Amazon yet, but I'm hoping there's a better selection process for me than reading endless reviews there. Thank you!

posted by amtho to education (17 comments total)

Get the Freedman, Pisani and Purves textbook. I TA'd for two of these professors, using the book to teach the classes. Best text hands down.
posted by mikeand1 at 7:41 PM on October 23, 2008

For me Excel was a good way to play with the statistical methods/formulae/etc. Just take your text book, load up excel, and play around.
posted by ian1977 at 7:48 PM on October 23, 2008

Find the specific methods you need to use and buy the books from Sage's "Quantitative Applications in the Social Sciences" about them.
posted by PueExMachina at 7:56 PM on October 23, 2008 [2 favorites]

I have Essentials of Statistics for the Behavioral Sceinces by Gravetter and Wallnau, and while not being the most thrilling read, it is extremely accessible and does a fantastic job of making the concepts clear through skillful wording, reptition (but not to the point of exasperation), and excellent examples.

I think it's one of those things in which you'll just have to find the textbook you like the most, read the material, and do some practice problems. Perhaps you could find some instructional videos online as well to supplement.


Or you could always just get a really nice calculator.
posted by Defenestrator at 8:22 PM on October 23, 2008 [1 favorite]

I just took a statistics class this summer. I have the book still, in new condition. If you wanna take if off my hands, send me a msg.
posted by disaster77 at 8:34 PM on October 23, 2008

These are good (except the calculator suggestion, sorry Defenestrator - but the text could be interesting).

More please, though! I'll check out all of these. mikeand1, is the text you recommended still in print? I couldn't even find any for sale, new or used, on Amazon. The later printing at the UNC library is missing too -- which is a kind of recommendation, I guess.
posted by amtho at 9:22 PM on October 23, 2008

I am sure I would second Freedman, Pisani, and Purves (for sale here, starting at $5 including shipping) if I had read it. I have a singly-authored stats textbook by Freedman that is absolutely outstanding.
posted by Maxwell_Smart at 9:35 PM on October 23, 2008 [1 favorite]

It's an old book, but it's the cornerstone of any course in statistics..

How to Lie With Statistics
posted by matty at 9:39 PM on October 23, 2008 [1 favorite]

Basic Statistics - Tales of Distributions is an excellent textbook. It's written in conversational English and should catch you up in no time. I second the suggestion to fire up excel and go along with the textbook - you aren't likely to be doing paper and pencil stats in a graduate class. You really don't need the most current version of the textbook - the basic material is the same mostly the problems are changed up for the publishers to make new textbooks.
posted by bigmusic at 9:47 PM on October 23, 2008

It may sound lame, but the Dummies/Idiots books are really pretty good for stuff like this.
posted by davidmsc at 10:22 PM on October 23, 2008

don't fire up excel, use a real stats package instead (for all sorts of reasons). JMP is good and has a trial version, or for free forever you can get R. More on R here.
posted by singingfish at 12:50 AM on October 24, 2008 [1 favorite]

Advanced Placement Statistics Exam "prep books" are thorough, succinct and readable. I recommend Hinders' 5 Steps to a 5. They cover descriptive statistics, probability, design of studies, and inferential statistics up to Anova. As prep books, they emphasize distinctions such as why you residual plots besides r to judge a regression. There are ten times as many practice questions as you need--but unlike textbooks, all questions are solved, often in detail. These are aimed at busy students, and old editions cost $5.
posted by gregoreo at 2:59 AM on October 24, 2008

I am thirding Freedman, Pisani, and Purves. I have kept my copy of this book from the 1980's because it is still better than others. The newer edition should be even better. I also agree that the SAGE series of quantitative apps for social science is a good choice.
posted by eaglehound at 7:10 AM on October 24, 2008

I would recommend Intuitive Biostatistics. It takes a great deal of time explaining why and when you should use stats, and how to interpret the results, using lots of examples and clear discussions. It's a great reference, and the author (Motulsky) is also the guy who owns GraphPad Software, makers of Prism, which I can't say enough good things about as a basic stats + graphing package. Graphpad offers fun free full-function trial versions. And it's cross-platform to boot.

Plus it's nice when you file a bug report and realize the guy writing back to you is the company owner. That's service for you.
posted by caution live frogs at 7:11 AM on October 24, 2008

Here's the Freedman Pisani Purves book on Amazon. Still in print, but you probably want to by a used copy.
posted by mikeand1 at 8:49 AM on October 24, 2008

Older AskMe about Statistics podcasts, and a few other resources.
posted by dog food sugar at 7:26 PM on October 24, 2008

If you want a light hearted not too deep treatment, check out the Cartoon Guide to Statistics
by Larry Gonick & Woollcott Smith.
posted by fings at 4:52 PM on October 26, 2008