Comments on: Influential Statisticians

by: Influential Statisticians « Permutations

Sun, 17 Jan 2010 22:39:32 +0000

[…] I recommend Seth Roberts’ comments here and Andrew Gelman’s comments here. […]

by: Sarosh Motivala

Wed, 13 Jan 2010 19:06:32 +0000

Great post Seth. To add to the mix, in my own work, I would add Peter Bentler for his contributions to structural equation modeling and Reuben Baron’s and David Kenny’s papers on mediation vs moderation very insightful.

As far as statisican-educators go, although thick and intimidating at first glance, Tabachnick and Fidell’s Using Multivariate Statistics is superbly written.

I wonder if there is a stats/methodology book on self-experimentation?

by: seth

Fri, 08 Jan 2010 21:15:37 +0000

David, thanks for the recommendation, the book sounds really interesting. Since a couple people have mentioned Tufte, I will say that I have learned nothing from his books. I find them annoying, starting with such titles as “Envisioning Information”. Tufte is, however, a genius entrepreneur.

by: Siah

Fri, 08 Jan 2010 20:29:59 +0000

A friend on twitter recommended your post. and enjoyed it a lot. I’m going to get John Tukey’s book. Thanks a lot and go bears

by: Mike Cooke

Fri, 08 Jan 2010 16:21:18 +0000

BC Canada had a high school Probability & Statistics grade 12 course in the 80s. I used Tukey’s EDA and Tufte’s books with considerable success. At its peak the course attracted as many students (especially girls) as the standard pre-university Algebra 12 course. 20 years later we hired a former student as my vice-principal who says that course was the best in his high school career. Then came the qualitative literacy movement (good) and the BC Ministry of Education mysteriously dropped the course (bad) to minimal protest (sad). Now there’s piddling disconnected tidbits of P & S in the K-12 curriculum. Bring back the good old days!

by: David

Fri, 08 Jan 2010 11:28:27 +0000

You might find this interesting if you haven’t already read it:

http://www.amazon.com/Probability-Theory-Logic-Science-Vol/dp/0521592712

by: Kevin Wright

Thu, 07 Jan 2010 15:23:09 +0000

I’ve used S-Plus and R for many years, but now work almost exclusively with R. I’ve filed many more bug reports for R than for S-Plus, so I find your suggestion that R has fewer bugs to be surprising. Also, from my own experience, almost every semi-annual release of R has major changes that break my existing code. S-Plus was much more stable.

YMMV.

Kevin

by: Alex Chernavsky

Thu, 07 Jan 2010 13:08:25 +0000

I have two of Edward Tufte’s books. They’re both excellent. He’s also quite a good speaker.

by: seth

Thu, 07 Jan 2010 01:41:21 +0000

Thanks, Aaron, glad to hear it. I was told by an editor that Exploratory Data Analysis was published only so that the same company could publish another book (with Mosteller) that came out at the same time.

by: Aaron Blaisdell

Thu, 07 Jan 2010 01:30:56 +0000

I agree with your list, Seth. Actually, because of you I bought a copy of Tukey’s EDA and am working through it one chapter at a time. I finished chapter 4 last night. On to Chapter 5 tonight. It’s great stuff!

I’ve always been a fan of plotting data to understand it through visuals before even doing other summary stats and definitely before applying any inferential stats. You’re right that psychology (maybe other science disciplines as well?) have really lost touch with this and overemphasize inferential stats.

I’d add a purely graphing book to your list: Edward Tufte’s The Visual Display of Quantitative Information.