Alex Tabarrok wrote:
As a thought experiment, imagine you were on
> an NSF committe and were presented with a proposal to study the minimum
> wage by using a natural experiment, surveying fast food restaurants etc.
> - i.e. everything which Card and Kruger did. I believe that just about
> everyone would have applauded the proposal as original, clever, and
> worthy of support. Consistency requires that we not abandon this
> perspective when the conclusion turns out to be something we don't agree
> with (isn't changing priors the point of doing research?)
>
Isn't one of the most important parts of doing "quality" empirical work
the proper collection of data? Hasn't it been estabished that
the C-K failed this test miserably?
Also, the idea that C-K studied a "natural experiment" is by no means
uncontroversial in the profession. I think there is a symposium on
Card-Krueger in Industrial Relations and Labor Review (forget the year),
where Dan Hammermesh, among others, attacks the Card-Krueger study - he
explains why it cannot be viewed as being either "natural" or an
"experiment."
Alex Robson
UC Irvine