Dear Statalist,

Happy new year!


Here is my question:

Code:
clear
set seed 2020
set obs 400
set more off
seq time, f(0) t(3) b(1)
seq id, f(1) t(100) b(4)
gene depression = round(runiform())
gene drug = cond(id>50,1,0)


meprobit depression drug##time  || id: , nolog
contrast drug@time, nowald effects
margins drug#time
The results looks like as follows:

Code:
 margins drug#time

Adjusted predictions                            Number of obs     =        400
Model VCE    : OIM

Expression   : Marginal predicted mean, predict()

------------------------------------------------------------------------------
             |            Delta-method
             |     Margin   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   drug#time |
        0 0  |   .4600545   .0704564     6.53   0.000     .3219624    .5981465
        0 1  |   .4200706   .0697705     6.02   0.000     .2833228    .5568183
        0 2  |   .4399289   .0702464     6.26   0.000     .3022484    .5776094
        0 3  |   .4800251   .0706285     6.80   0.000     .3415959    .6184543
        1 0  |   .4197953   .0698913     6.01   0.000     .2828109    .5567796
        1 1  |   .4398788    .070273     6.26   0.000     .3021462    .5776114
        1 2  |   .5598365   .0701128     7.98   0.000      .422418    .6972551
        1 3  |   .5000074   .0706692     7.08   0.000     .3614984    .6385165
------------------------------------------------------------------------------

How do I test the difference between the marginal probability .5000074 - .4800251 ? Is it correct to estimate the statistical significance by?

Code:
 lincom 1.drug#3.time-0.drug#3.time
Thank you for any tip.


All the best,

Tiago