Science in action: The Veronique de Rugy story
So Veronique de Rugy put out a paper entitled Stimulus Facts. In it, she ran a regression of the money spent by Obama’s “stimulus” package recently, to see if it was actually being spent where unemployment is highest. It turns out it’s not, and most important factor she found in determining where money was spent, was whether the area was represented by a Republican or a Democrat. Democrats got more, you see.
Like any good scientist, she published her data, her formulas, all the results she got. This was all put out for the world to see, to criticize, and even to reproduce if desired. She then got some criticism, and re-ran her tests taking into account the suggestions she got, and then published a new version of the report without making the old version go away, even.
And as it turns out, whether a Republican represents you became the third most important factor in whether you got stimulus money. State capitals got a lot more, states with higher income levels (wealthier and presumably higher costs of living) got more, and then areas represented by Republicans got less.
Agree or disagree with her methodology, she’s putting the data out there and deserves to be commended for it. Let’s see her critics on the left take what she’s done and fix it, if what she’s done is so wrong.
And an aside to Nate Silver: it’s funny that you tell Veronique de Rugy to publish five models when you yourself only publish one. Well, you publish the results of one. While I fully understand you not wanting to give away the store, I think you should reconsider what you do from the porch of your glass house. After all, to take your own advice you could try running versions of your process with different sets of assumptions, fudge factors, and other subjective tweaks, and explain what’s going on with the different results. But you don’t, so it’s a pretty cheap shot to criticize her for sharing your own failing.