Books to make Statistics Interesting
December 20, 2013 2:35 PM   Subscribe

If I'm going to succeed in a field I am considering (bioinformatics), I'm going to need to learn a good deal of statistics. I have limited experience in the subject and have never found the details of it especially compelling. Can you recommend some books on statiistics (not necessarily light on the details) that are well written and interesting? Something designed to get the layman interested in the details could be good, but textbook recommendations would also be good.
posted by garuda to Education (20 answers total) 38 users marked this as a favorite
You can't go wrong with Nate Silver's "The Signal and the Noise". Light (very light) on the technical details, but if you want to get excited about statistics, that's the place to start.
posted by downing street memo at 2:48 PM on December 20, 2013 [2 favorites]

May I ask a question? Why are you considering a field that is very heavy in stats if you don't find stats compelling or interesting?
posted by sid at 2:53 PM on December 20, 2013 [8 favorites]

The Signal and the Noise is an awesome book, agreed. For something slightly more technical, consider Motulsky's Intuitive Biostatistics: -- it introduces you to an huge variety of statistical concepts and does so gently, although keep in mind that it has a biostats-oriented slant (ie, it's written for clinical researchers), which could end up being a positive depending on where you go in bioinformatics.
posted by un petit cadeau at 2:57 PM on December 20, 2013

The Manga Guide to Statistics

The Cartoon Guide to Statistics by Larry Gonick!!!
posted by Chocolate Pickle at 3:22 PM on December 20, 2013

Head First Statistics
posted by epersonae at 3:57 PM on December 20, 2013

I bookmarked this a few months back. I have yet to read it so can't offer personal opinions, but it was recommended on BoingBoing.
posted by PaulaSchultz at 3:58 PM on December 20, 2013

Intuitive Biostatistics is one of the texts that my university uses for the undergraduate biostatistics course - I haven't TAed it but I have heard good things about that text.
posted by pemberkins at 4:00 PM on December 20, 2013

Something designed to get the layman interested in the details could be good

I got a lot out of

Lady Luck by Warren Weaver,


Probability and Statistics for Everyman by Irving Adler-- which goes surprisingly deep, and which I absolutely loved.
posted by jamjam at 4:36 PM on December 20, 2013

Seconding the Cartoon Guide to Statistics. I actually used it to study for my first stats final in grad school.

The book Freakonomics seems to be falling out of favor but it is the thing that got me really excited about multivariate regression before I went to grad school, as a person who literally changed majors in undergrad so I wouldn't have to take stats. Multivariate regression analysis is now one of my favorite things, no joke. They do get a bit loosey-gosey with their analysis, but as an entree to the field, it's very engaging.
posted by lunasol at 5:09 PM on December 20, 2013

The older editions of David S Moore are very readable textbooks, with extremely intuitive explanations. Be sure to get older editions; the newer ones still have his name on them, but he says he has not been active in the revisions since he retired.
posted by wittgenstein at 5:48 PM on December 20, 2013

Learning statistics because it is a requirement is deathly dull. Learning statistics because you need it to solve a problem you are really interested in is a totally different story. Read one of the breezier books recommended above and then once you start digging in and you know why just averaging those numbers in excel isn't going to cut it any more and it will become a lot easier to learn.

I recommend Kennedy's A Guide to Econometrics (the edition does not matter) as a supplemental text. He treats each topic three times, first at a high level, almost all words and pictures little math. Then in a second section he busts out the maths and the literature review and in a third section he covers practical notes and pitfalls.
posted by shothotbot at 5:49 PM on December 20, 2013 [1 favorite]

How to Lie with Statistics is an interesting read, especially with how people skew statistics to their advantage, such as media stories.
posted by deinemutti at 6:36 PM on December 20, 2013

2008 question "Best way to re-learn statistics?" The resources listed are mostly ways to learn statistics the first time, never fear.
posted by amtho at 7:50 PM on December 20, 2013

May I ask a question? Why are you considering a field that is very heavy in stats if you don't find stats compelling or interesting?

I find the questions that can be answered with bioinformatics very interesting. I've long been floundering around doing research in the biological sciences, at times finding myself staring at large datasets I've helped create and wanting to engage them in a productive way. I'm likely to give this field a go, for personal and professional reasons, but I want to delve into the background before I get in too deep. My hypothesis is that statistics is, in fact, quite interesting, and that I just have not approached it in the right way.
posted by garuda at 5:47 AM on December 21, 2013

Thank you, oh great patrons of the internet!

I definitely want to check out Intuitive Biostatistics at some point, but I think I'll start with one of these shorter ones first.
posted by garuda at 6:27 AM on December 21, 2013

i'd suggest finding a free biostats data set on line, getting R, and playing around with the data. any introductory biostats book will do to give you a guide. playing around with the actual data will be the part that can help you decide if it's actually interesting, or if you're just interested in the idea of it. i know for a fact there are biostats packages in R. you could try to follow their vignettes.
posted by cupcake1337 at 7:29 PM on December 21, 2013

The Handbook of Statistical Practice
Statistics Done Wrong
Choosing a Statistical Test (just for reference)
Khan Academy (helpful in some instances)

Also, the book Statistics by Freedman, et al served me well enough.
posted by quadog at 1:03 AM on December 22, 2013

These might be more poppy than what you're looking for, but they lit a fire for the topic in me personally:

The Drunkard's Walk by Leonard Mlodinow
Fooled by Randomness by Nassim Nicholas Taleb
posted by rollick at 12:27 PM on December 22, 2013

More generally, I think you'll want different books for the goals of (1) make statistics interesting, versus (2) learn the nuts and bolts of statistics. I have yet to find one book that successfully accomplishes both. If your first step is to make yourself find statistics more interesting (without necessarily learning any statistical nuts and bolts), many of the pop-sci stats books recommended above (The Drunkard's Walk, The Signal and The Noise...) books will serve you well.

Also, since you mentioned large datasets, I'd like to give a hat tip to Practical Computing for Biologists, which I found to be a useful intro to a smorgasbord of computational methods for big data. (Skip it if you're already up to speed on the computational side of things.)
posted by pemberkins at 1:39 PM on December 22, 2013

Rather than a book I will commend you to Tableau Software and their data sets so you can visualize this stuff. Very nice way of dealing with large datasets.
posted by ptm at 1:44 AM on December 23, 2013

« Older Regarding diet/exercise, what ...   |  I am involved with a tiny nonp... Newer »

You are not logged in, either login or create an account to post comments