yes, variables need to be identified if they're nominal, or else the
algorithm will treat them as if they're continuous. The latter isn't
really much of a problem in practice in many cases except where the
nominal categories aren't also ordered. So dichotomies and ordered
categories seem to work ok. The manual has more info on this.
Gary
On Tue, 12 Jul 2005, Sanjoy Bhattacharjee wrote:
Prof. King ...
do we need to mention if some of the "FULLY OBSERVED"
variables in our data set are Nominal(_AMnoms) and/or
Ordinal(_AMords)...
i mean if the answer is no, then won't we miss
something, given in our model estimation we are
treating the nominal as dummy and ordinal as an
indicator of ordered category
say e.g. we are fitting ordered probit, y on (X1 X2 X3
X4 X10) ...
1. Y is (bad/neutral/good) and is also fully observed
2. x1, x2 represents male/female and
employed/unemployed ..they are also fully observed
now, if the answer is Yes ... then it confounds me as
well, since in EM, we assume multivariate normality
... how can we accomodate nominal/categorical
assumption there
thanks and regards, Sanjoy
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive:
http://lists.hmdc.harvard.edu/?info=amelia
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive:
http://lists.hmdc.harvard.edu/?info=amelia