Piece on the arbitrariness of “statistical significance” and the effect it has in medical research.
Not terribly good argument that brain science doesn’t disprove God (or Darwin, I hope, but the headline is clearer than the essay).
Even some conservatives recognise that Darwin didn’t lead to Hitler, and this one makes an interesting distinction or two.
I really wish people wouldn’t put Bayesian method up as the ‘solution’ to p-values. They’re not (except perhaps sociologically). p-value abuse can be done in the Bayesian framework too (I even have a paper on some methods for doing this!), and the “Bayesian” solution can be done just as well in a maximum likelihood framework.
The problem is that to do this stuff properly, you have to be able to understand what you’re doing, and to think. Unfortunately, statistical methods are treated as oracles that mysteriously give you The Answer once you’ve worked out which bit of the computer’s entrails to sacrifice.
Unfortunately, statistical methods are treated as oracles that mysteriously give you The Answer once you’ve worked out which bit of the computer’s entrails to sacrifice.
Can I steal that?
Of course.
Oscar Wilde (I think) was once involved in an exchange that went something like “I wish I had said that” – “Oh you will, you will”.
He did indeed but the riposte, from Whistler I think, was “you will Oscar, you will”.
Not terribly good argument that brain science doesn’t disprove God (or Darwin, I hope, but the headline is clearer than the essay).
..and the author used to be a professor!?