Well, the OLS is very different from the ordinal probit. The former is a
model for a continuous variable whereas the latter is the model for an
ordianl variable. It is true that the latter assume a normal distribution
for the underlying continuous variable, but the coefficients for the
ordinal probit do not have the same interpretation as the OLS
coefficients.
Kosuke
On Wed, 16 Apr 2003, Stanislav Markus wrote:
So I got rid of the intercept, and ran optim()
successfully, estimating
3 coefficients and 2 tau values. However, the values for coefficients
now differ from those yielded by OLS (e.g. 0.6 vs 0.3) when intercept
was included.. Is this theoretically justifiable?
Thanks,
Stan
****************************
Stanislav Markus
Ph.D. Candidate
Harvard University
Department of Government
e: smarkus(a)fas.harvard.edu
t: 617.513.5407
-----Original Message-----
From: gov2001-l-admin(a)fas.harvard.edu
[mailto:gov2001-l-admin@fas.harvard.edu] On Behalf Of Kosuke Imai
Sent: Monday, April 14, 2003 10:12 AM
To: gov2001-l(a)fas.harvard.edu
Subject: [gov2001-l] ternary()
You might want to replace the last line of ternary() with
points((X[,2] - X[,3])*s3, X[,1], pch=pch, ...)
This way, your input for "pch" won't get ignored!
Kosuke
_______________________________________________
gov2001-l mailing list
gov2001-l(a)fas.harvard.edu
http://www.fas.harvard.edu/mailman/listinfo/gov2001-l
_______________________________________________
gov2001-l mailing list
gov2001-l(a)fas.harvard.edu
http://www.fas.harvard.edu/mailman/listinfo/gov2001-l