I'm pretty dumb, but the Clarke and Primo article seemed really dumb.
Downs created a model for why political parties acted some way, but only the empirical observation actually seems to be useful (or perhaps the idea that politicians act much like competing businesses). At best, Clarke and Primo managed to suck all the air out of their own argument by not citing an extension of the Downs model beyond the observation that parties converge on the electoral centre.
On Arrow's impossibility theorem, they take stupid to 11, by discussing a proven mathematical theorem in the context of not testing it with data, AS A DEFENCE OF NOT TESTING THEORIES IN THE SOCIAL SCIENCES!
That is not even wrong. It misunderstands what a theorem is in the mathematical sense, what is required to trust a theorem, and then abuses this ignorance in defence of the idea of writing up untested models.
The worst thing is I half-agree with the authors: there is room in academia for proposing a model before it is tested. But given that proposing a model is really easy, and testing a model is the actual hard work, journals have a good reason to not review untested models.
But this was published on April 1; maybe it's a joke?
2 comments:
Note that the physics envy article about political scientists gives two examples of useful theories in the study of politics.
Of course neither was developed by a political scientist!!
I'm pretty dumb, but the Clarke and Primo article seemed really dumb.
Downs created a model for why political parties acted some way, but only the empirical observation actually seems to be useful (or perhaps the idea that politicians act much like competing businesses). At best, Clarke and Primo managed to suck all the air out of their own argument by not citing an extension of the Downs model beyond the observation that parties converge on the electoral centre.
On Arrow's impossibility theorem, they take stupid to 11, by discussing a proven mathematical theorem in the context of not testing it with data, AS A DEFENCE OF NOT TESTING THEORIES IN THE SOCIAL SCIENCES!
That is not even wrong. It misunderstands what a theorem is in the mathematical sense, what is required to trust a theorem, and then abuses this ignorance in defence of the idea of writing up untested models.
The worst thing is I half-agree with the authors: there is room in academia for proposing a model before it is tested. But given that proposing a model is really easy, and testing a model is the actual hard work, journals have a good reason to not review untested models.
But this was published on April 1; maybe it's a joke?
Post a Comment