Posted by Jay_Bravest_Face on March 1, 2008, at 23:20:55
In reply to Re: AD's no better than placebo research » Giorgio, posted by Larry Hoover on March 1, 2008, at 16:53:49
Lar, you probably already know of this great little book (yes, it is littlewhich makes easy reading for those not keen on stats, like me!). Its called "How to Lie With Statistics"
I got it as a reader for my second year social research course. I read it in about a day! Dont let that fool you though, because you (anybody) will be coming back again and again to it. It is not THE ANSWER, and dont think you can use it to go around telling everybody any stats they present are wrong. Its just *good to know* info. It pokes into the methodology and logic in compiling and crunching stats, especially ones that seem to suggest they are failproof. I am not going to rehash the whole book here, but here is the title of its 10 chapters, which give some hints to its directions. Anybody can read this book, and for those with even a bit of statistical education, this will be a real treat.1. The Sample with the Built-in Bias
2. The Well-Chosen Average
3. The Little Figures That Are Not There
4. Much Ado about Practically Nothing
5. The Gee-Whiz Graph
6. The One-Dimensional Picture
7. The Semi-attached Figure
8. Post Hoc Rides Again
9. How to Statisticulate
10. How to Talk Back to a Statistic
The book is cheapgrab yourself a copy!
poster:Jay_Bravest_Face
thread:815587
URL: http://www.dr-bob.org/babble/20080221/msgs/815651.html