In the case of -logit- and -lstat-, choosing
to divide predicted probabilities at p = 0.5
is a little arbitrary, but appealing on
symmetry grounds.
Can you say what people do, or what would
you justify, for reducing predictions from
-oprobit- to integer categories? Any
notion of percent correct seems a bit more
elusive in the ordinal case. Why not
just -predict, xb- and calculate a
concordance correlation? Or even
draw a graph?
Nick
[email protected]
Patricia Sourdin
> can anyone help me to calculate percent correctly predicted
> after oprobit. I
> guess what I'd like is something like -lstat-. I'm currently
> overseas and all
> my Stata manuals are at home!
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
