Dear Statlisters,
I was wondering if anyone knows how to set up a constraint for a full equation when estimating a multiple equation models.
Lets take for example a Multinomial model:
What i would like to do its to estimate the second model, keeping, say all the coefficients corresponding to the option "prepaid" the same as the ones observed in the first model when estimating the second model.
I know i could do this using single constraint for each coefficient, but i was wondering how to do it for all coefficients within the "prepaid" equation.
Thank you.
Fernando
I was wondering if anyone knows how to set up a constraint for a full equation when estimating a multiple equation models.
Lets take for example a Multinomial model:
Code:
webuse sysdsn1, clear
mlogit insure age male nonwhite i.site, base(1)
est sto m1
set seed 1
* Now I randomly reclassify and create a new category:
replace insure=4 if insure>2 & runiform()<.3
mlogit insure age male nonwhite i.site, base(1)
est sto m2
est tab m1 m2
----------------------------------------
Variable | m1 m2
-------------+--------------------------
Prepaid |
age | -.01174499 -.01153409
male | .56169343 .56202261
nonwhite | .97477679 .96815683
|
site |
2 | .11303586 .10606888
3 | -.58798794 -.58740439
|
_cons | .2697127 .2641631
-------------+--------------------------
Uninsure |
age | -.00779614 -.00589985
male | .45184964 .3545223
nonwhite | .21705895 .3373406
|
site |
2 | -1.2115627 -1.1830257
3 | -.20781232 -.29666224
|
_cons | -1.286943 -1.5329003
-------------+--------------------------
4 |
age | -.01930015
male | .85839536
nonwhite | .08429027
|
site |
2 | -1.4816411
3 | -.74083635
|
_cons | -1.9029739
----------------------------------------
I know i could do this using single constraint for each coefficient, but i was wondering how to do it for all coefficients within the "prepaid" equation.
Thank you.
Fernando