Hacker News Reader

Thanks for the Mayo recommendation. I am currently going through her Statistical Inference as Severe Testing, and find it very tiring. So far very little is about actual statistics and the meat of disputes, but rather about meta questions and the he-said, she-said of those disputes. Maybe the prior book Error and the Growth of Experimental Knowledge is a better exposition of the ideas?

View on HN

I got mich out of it, but only on the meta level also. Afair she never goes into technical detail of e.g. Neyman Pearson, but covers things like how to overcome the Quine-Duhem-problem with statistics, or how to stengthen Popper’s theory. But on that level I learned a lot, and it trickles down to what is important in actual experimental design and why.

Reply on news.ycombinator.com