Different types of empirical research are convincing to different people. I, and
I gather Bryan, tend to like the type where a problem is approached from several
different angles using perhaps relatively simple econometrics and is integrated with
other types of evidence including historical evidence (non-quantitative empirics)
and a modicum of theory told with a plausible story. The Bell Curve does well on
this measure.
The other sort uses of empirical research estimates a single statistical model
using the very best of the modern econometricians arsenal. Typically, only one data
set is used. The Bell Curve fails this test miserably.
The two sorts of research are not incompatible but almost no one does both
types well and many who do the latter type of research are disdainful of the former
and vice-versa.
By the way, as a paradigmatic example of the first sort I would mention
Friedman and Schwart'z A Monetary History of the United States. Chris, I wonder
what you would say about the quality of this work? (And I mean today not circa
1965). I still think it is great even thought it fails test number 2 very badly.
Alex
