At 08:52 AM 9/25/2003 -0400, Gary King wrote:
You should never pay much attention to the alpha in the
extended model,
unless you believe the model holds exactly. The best use of the extended
model, as the basic model, is to go all the way to the precinct-level
estimates...
I understand the argument for burdening the estimation stage with lots of
hypothesized covariates (although I don't like it!). I also understand the
problem with using the alphas as statistical summaries. They just move
around the bivariate normal from which the precinct estimates are drawn;
they do not reflect the actual relationship between covariates and
estimated behavior that emerges after the bounds constrain precinct
estimation. But I'm not clear on what you are prescribing when you say
people should go to the precinct estimates if they want to report the
relationship between various explanatory variables and the (estimated)
group behavior. I thought that's what I was doing with the second-stage
regression. Can you provide any guidance?
(Was it Huey Long who accused a campaign opponent of
being a "known
matriculator"?)
Apparently it was George Smathers, a senator from Florida. He is quoted as
having said the following about Claude Pepper (see
http://www.lincolninstitute.org/archives/kennedy/0303.html)
- was a shameless extrovert
- practiced celibacy before his marriage
- had a sister who was a thespian in wicked New York
- matriculated before entering college
- was a well-known heterosexual in college
steve voss
--------------------------------------------------------------
D. Stephen Voss, Asst. Prof URL-
http://www.uky.edu/~dsvoss
1603 Patterson Office Tower Phone- (859)257-4313
Dept. of Political Science "It was my duty to bring
University of Kentucky the facts to light, and there
Lexington, KY 40506-0027 I must leave it." Sherlock Holmes
--------------------------------------------------------------
-
EI mailing list served by Harvard-MIT Data Center
Subscribe/Unsubscribe:
http://lists.hmdc.harvard.edu/?info=ei