[PEN-L:6592] RE: on econometrics

["Fellows, Jeffrey" <jmf9@xxxxxxx>]

Without econometrics, I doubt if I could argue that homicide was related to
poverty and income inequality instead of race and "southern culture."
Getting past the latter two myths is difficult enough as it is. Thankfully,
there is decent theory behind the empirical relationships, including how one
gets from income inequality to diffused interpersonal violence.

I am not a skilled econometrician, and I would love to get any suggestions
others have for empirical methods that will be taken seriously by applied
scientists. Please send references.

Jeff

-----Original Message-----
From: Michael Perelman [mailto:michael@xxxxxxxxxxxxxxxxx]
Sent: Monday, May 10, 1999 11:42 AM
To: pen-l@xxxxxxxxxxxxxxxxxxx
Cc: jdinardo@xxxxxxxxxxxxxxx; lbo-talk@xxxxxxxxxxxxxxx
>Subject: [PEN-L:6580] on econometrics


Can an econometric exercise ever produce anything more than a hint of an
interesting idea?  I suspect that a skillful econometrician can find a
relationship
between any two arbitrary data sets.

As econometric techniques become more sophisticated, become more
brittle.

How is a sophisticated econometric model more convincing than a simple
scatter diagram?

--

Michael Perelman
Economics Department
California State University
michael@xxxxxxxxxxxxxxxxx
Chico, CA 95929
530-898-5321
fax 530-898-5901