Re: st: Testing the extent of difference between two coefficients in the same model

[David Hoaglin <dchoaglin@gmail.com>]

You could change the predictors in the model.  Instead of "old" and
"new" you could use "new MINUS old" and "old"; the fit will be the
same as with "old" and "new".  Then the estimate that you seek will be
the coefficient of "new MINUS old".

In either form of the model, those coefficients tell you about the
contributions of those predictors after adjusting for the
contributions of the other variables in the model.  (I prefer
"adjusting for" because those other variables are probably not under
any actual "control.")

David Hoaglin


> Is there a way to estimate the extent of the difference between two coefficients obtained in the same xtlogit model.
>
> I have one variable "old" and one variable "new" plus some controls. My hypothesis is that "new" will have a larger positive influence y than will "old".
*
*   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/