Science

"Five ways to fix statistics"

Interesting, especially because it introduced me to a fine idea by David Colquhoun and then to his blog.  Two posts from his blog:

"We know little about the effect of diet on health. That’s why so much is written about it". (And he links to a short paper by John Ioannidis, "Implausible results in human nutrition research".)

"Placebo effects are weak: regression to the mean is the main reason ineffective treatments appear to work". 


"The Most Important Question In Science"

Yes, indeed. A whole lot of faulty thinking and argument could be substantially fixed if people would just follow the advice given here.

(Milton Friedman, of course, knew this very well.)

His goodness is of course in part intellectual, even Jewish. He is a man of the book. In the Money Seminar, and anywhere else he's standing, he asks always, persistently "How do you know?" It's a terrifying question, since most of the time we can't say how we know, because we don't know. The question feels like an assault if you're not ready for it. But of course Milton is seriously curious, looking for enlightenment, ready to take this or that side in the schule (the school of Hillel might claim that inflation is a wage-price spiral; the school of Friedman claims on the contrary that inflation is everywhere and always a monetary phenomenon). I am told [How do you know, Deirdre?] that when people these days say to him how good he looks, considering he is 90, he sometimes replies sharply, "How do you know? Are you an expert in gerontology?"