Linking in the stats
This fall, I squirreled away 400 e-mails to myself with links to interesting papers or blogs or commentary. Lots of them were things I wanted to stick on my blogs. Now I will try to slowly weed myself down to none again. It will likely results in multiple linking posts, so I declare this to be the first in the series, and it will be all about stats.
First up, Telliamed revisited’s post on the 10 commandments of statistics. Post it prominently on the classroom walls.
This one I have linked to before, but, hey, let’s repeat the good stuff. The p-curve page. Includes the Paper, the app, the user’s guide and supplementary materials. Use it on your favorite area of research.
Speaking of p-curve, Here is a paper (pdf) from Gelman and Loken on how multiple comparisons can be a problem, even when all practices are non-questionable. (Now, I hope that link will work).
A path to learning is to get exposed to What Not To Do! And, the least painful way to do that is to observe other failures, or at least read about them kind of in the abstract. Statistics Done Wrong is an excellent opportunity to do this. It is, um, amazing to realize how many of those misconceptions one has held…
NeoAcademic is a blog from an I/O perspective (Industrial/organizational psychology that is), and Richard Landers posted a series of commentary on a paper comparing Null hypothesis with effect size. I link in the last one (because that is the one that I sent myself), but you can easily get to the other installments from his post.
I also think I’ve linked in Felix Schönbrodt’s post before, but also worth repeating. At what sample size does correlations stablilize?
The collected works of Tukey. In Google Books.
Well, I’m down to November. There is more to come, but I have to sort it through. Probably a second post of stats links.