Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: negative predicted values after -xtlogit-


From   "Verkuilen, Jay" <JVerkuilen@gc.cuny.edu>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: negative predicted values after -xtlogit-
Date   Tue, 16 Sep 2008 14:59:41 -0400

Campbell, Richard T. wrote:

>You are asking for the linear predictor when you
>use the xb option. These can be negative, since they are
>fitted log odds. You want to use the default which will give you fitted
>probabilities

Or gen predprob = exp(linpred)/(1 + exp(linpred))

The latter might be useful if you can't get -predict- to make some
linear combination you want, e.g., adjusting for the random effect in a
particular way. 

Sometimes Stata has a nice option that's about 50% of the way to what
you want, in which case a little additional calculation gets you all the
way there. 

JV

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index