Hi there,
For the below mlogit output, I am trying to calculate the predicted probabilities for my subpopulation of interest. The margins command returns me what I want. I am trying to reproduce the predicted probability but what I get it different than margins returns. What is wrong? Thanks for your help.
Best,
NM

Code:
margins ,  predict (outcome (2)) at (SEX=0 AGE=65 SRH_B=0  RACE=0 ) atmeans   // .0087285
Code:
display invlogit([2_Y1]_cons+(0*[2_Y1]1.SEX)+(65*[2_Y1]AGE)+(0*[2_Y1]1.SRH_B)+(0*[2_Y1]1.RACE)) // .00917686
------------------------------------------------------------------------------
| Linearized
SRH_E_new | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
0_Y0. | (base outcome)
-------------+----------------------------------------------------------------
1_Y1. |
SEX |
1.female | .2962368 .0723987 4.09 0.000 .1541516 .4383219
AGE | .0773073 .0039719 19.46 0.000 .0695124 .0851023
|
SRH_B |
1.X. | 2.596534 .0731401 35.50 0.000 2.452994 2.740074
RACE |
1.Z1 | .2720347 .108607 2.50 0.012 .0588894 .4851801
_cons | -7.984602 .3139163 -25.44 0.000 -8.600675 -7.36853
-------------+----------------------------------------------------------------
2_Y2 |
SEX |
1.female | -.3784588 .0985947 -3.84 0.000 -.5719546 -.184963
AGE | .1267682 .0065075 19.48 0.000 .1139971 .1395394
SRH_B |
1.X. | 2.471062 .1093944 22.59 0.000 2.256372 2.685753
RACE |
1.Z. | .0514034 .1754268 0.29 0.770 -.2928782 .395685
_cons | -12.92179 .5251256 -24.61 0.000 -13.95237 -11.89121
------------------------------------------------------------------------------