I ran a bivariate probit regression and then tried to calculate
marginal effects with the command -mfx, predict(pmarg1)-. This
resulted in the following message:
warning: derivative missing; try rescaling variable age
The variable age (measured in years) is one of the independent
variables in my model. I read -help j_mfxscale- and the document at
http://www.stata.com/support/faqs/stat/mfx_scale.html and then began
to rescale the variable age. I multiplied and divided age by values
between 10 and 1,000,000,000,000 and still received the error
message. When I added the option -tracelvl(1)- to the -mfx- command I
found that the error message about age occurred while another
independent variable was processed. Dividing that second variable by
10 made it possible to calculate the marginal effects and standard
errors for all independent variables in my model.
I had the same problem with another regression. The warning message
again recommended rescaling the variable age, which turned out to be
misleading. In that case, rescaling a third variable solved the
problem.
I have two questions: (1) Why does the error message recommend
rescaling a variable that is not the cause of the problem? (2) Can
Stata/SE handle marginal effects better than Intercooled Stata 8.2,
which I am using now?
Many thanks,
Friedrich Huebler
__________________________________
Do you Yahoo!?
Y! Messenger - Communicate in real time. Download now.
http://messenger.yahoo.com
*
* 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/
