Monday, 14 March 2016

Read Andrew Gelman, He Is Our Master In Everything Statistical

I’m doing this to prove I’m still alive. I’ve been working hard on my Riemann-Roch essay when I haven’t been, you know, doing the day job and hefting iron, and this weekend, giving the ornamental grasses their winter savaging and keeping the lawn from getting out of control.

If you’re remotely interested in the philosophy of science, statistics, or read science articles in the press, or even research in medical, sociological and psychology journals, or you're into business analysis and insight, or are likely to believe “studies” that prove people with more facial hair are likely to be more sexist, then you need to be reading Andrew Gelman’s blog, and especially this paper on the Garden of Forking Paths. A few weeks of doing so, and catching up with the previous posts, will soon set you straight. I would go so far to say that you should not read one more newspaper “science” article until you understand the idea behind the Garden of Forking paths. Also read his contribution to the ASA report on p-values.

Professor Gelman is the only other person I’ve read who references Imre Lakatos, who was a philosopher of science at the LSE in the late 1960’s and early 1970’s. Recently, as with this post Professor Gelman has been hitting his stride on the issue of low-quality and junk research. That particular post is as polemical, prescriptive and on-the-money as anything Lakatos ever wrote.

He has serious technical chops, having co-created a Bayesian inference language called STAN, and co-authored the first textbook on Bayesian statistics I’ve found convincing. He also has serious common sense and if you don’t believe a statistician can demonstrate sprezzatura then read this.

(Now I’m going back to trying to be sure that, just because I can embed a Riemann surface into a complex projective curve, doesn’t mean that it is a projective curve. I think it means that Riemann surfaces are like cross-sections of projective curves.)

PS: The title is a light-hearted reference to what Gauss said about Euler: "Lisez Euler, lisez Euler, c'est notre maître à tous."

No comments:

Post a Comment