Probabilities are calculated here given predictor values, and generating
probabilities subject by subject is what makes sense. What you are
suggesting is only going to work well in very simple cases.
It won't work at all for continuous predictors If all of your
predictors are binary, categorical, or ordinal, I'd still use predict,
and then collapse across your predictor categories. If you have 1 or 2
binary predictors, you could use nlcom, inserting predictor values into
the equations.
If
Kanter, Rebecca wrote:
Hi,
I would like to get the logit output in stata in a form of "probabilities" (via transforming the log odds through the formula e^x / (1 + e^x)). Does anyone know if there is a command(s) to do this? (I know you can get the "predicted" line in the probability format, but am looking here, to be able to see the coefficients in the form of probabilities).
Thanks,
Rebecca
___________________________________________
Rebecca M. Kanter
PhD Candidate
Johns Hopkins Bloomberg School of Public Health
Department of International Health
Center for Human Nutrition
*
* 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/
--
E. Paul Wileyto, Ph.D.
Assistant Professor of Biostatistics
Tobacco Use Research Center
School of Medicine, U. of Pennsylvania
3535 Market Street, Suite 4100
Philadelphia, PA 19104-3309
215-746-7147
Fax: 215-746-7140
[email protected]
*
* 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/
