Bryan Caplan wrote:
> Chris, would you mind telling us what the variances of u1, u2, and u3
> were? I think you gave us all the necessary info, but I'm too lazy. :-)
The standard deviations were about 1, 10, and 2 in the log-wage,
AFQT, and years of education equations, respectively.
> A while back I said that the Card/Krueger minimum wage study was at the
> 90th percentile of quality for published empirical work. I would put
> Murray and Herrnstein at the 99th percentile of quality. Applied work
> is never perfect, but I have to think you're being way too harsh.
And I turn think you're being way, way to quick to turn a blind
eye to all the serious problems here. I doubt most of the Bell Curve
would ever have got through peer review at any refereed economics
journal, for very good reasons. I'd put the Bell Curve in the bottom
5% of the distribution of published empirical work in economics. And
I'm probably being far too generous.
> What empirical problems *do* you think are "pretty easy"?
Most questions that do not implicitly or explicitly require
causation to be determined. Causal questions where there exists
large truly exogenous variation, such as in solid controlled
experiments or good natural experiments.
Chris Auld (403)220-4098
Economics, University of Calgary <mailto:[EMAIL PROTECTED]>
Calgary, Alberta, Canada <URL:http://jerry.ss.ucalgary.ca/>
