This blog is an exercise in scalability. Instead of sending a long email to one person, I put the email in a blog where thousands can read it. Instead of devoting three hours to a referee report that will only be read by two people (the author and the journal editor), I do the equivalent here on the blog.
When the American Statistical Association asked me to participate in a workshop to give writing advice for a select group of young researchers, I agreed to participate in this program, as long as the authors were willing to have their articles and my comments posted on the blog. I think my advice on writing research articles had much more effect being posted on the web than it would’ve had, if I’d kept it in that meeting. (On the other hand, my advice benefited from having those two student papers to push against. If I’d just tried to give general advice without the context, I don’t think it would’ve been so useful to anyone.)
I’ve tried to be scalable for many years before I started blogging. Rather than put a huge effort into preparing a new class, I’ll center the class around notes which become a book. The readership of Bayesian Data Analysis is orders of magnitude more than all the students who’ve taken my courses on Bayesian statistics.
But now there are programs such as Coursera and institutions such as Wikipedia that can easily reach thousands. So there’s a lot more to scalability than I’ve done so far.
I don’t think I’m the only one who would love to see you do a Coursera course on Bayesian Data Analysis or multilevel models.
Yes!
Yes, that would be really wonderful !!
Ditto!
Maybe something like reviewing Coursera course material and pedagogy or something meta-Cousera in the sense of critically accessing this “about to be terribly abundant free material” that needs be audited and jointly processed to be best taken advantage of in its current and future full potential (i.e. how to nudge it to be continually getting less and less wrong).
Just to look at things in the other direction for a minute,it might be assumed (erroneously in my case) that bloggers see little if any distinction between what they write in a direct e-mail to person x and what they are prepared to have spread over all the world, or to have shared with whomever person x feels like telling.
Hey Prof. Gelman, I felt like you’d find this article as another opportunity for a “teachable moment” in the misinterpretation of statistics: http://www.project-syndicate.org/commentary/mitt-romney-rejects-his-natural-voters-by-jeffrey-frankel
(I’m not a Republican btw – I just think bad arguments don’t help my side)
Pingback: Could someone please lock this guy and Niall Ferguson in a room together? « Statistical Modeling, Causal Inference, and Social Science