Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

st: Post-estimation predicted probability and predicted probability calculated by hand do not match after running cloglog

 From Urmi Bhattacharya To statalist@hsphsun2.harvard.edu Subject st: Post-estimation predicted probability and predicted probability calculated by hand do not match after running cloglog Date Fri, 27 May 2011 15:12:13 -0400

```Dear Statalisters,
I am estimating a discrete proportional hazard model using clogclog
model. I use a fully non parametric specification for the baseline
hazard. My variables include continuous as well as categorical ones.
Of the categorical variables, three of them (dad_edu, mom_edu and
caste) have three categories each.
I run the following specification(durat1-11 are time dummies):

cloglog school_left childage i.child_female i.urban i.caste
wage_bey
> on~9 distance_p~l distance_m~l distance_h~l  month_percap_cons durat1 durat2 durat3 durat4 durat5 durat6 durat7 durat8 durat9 durat10
> durat11, nocons nolog
Complementary log-log regression                Number of obs     =      47569
Zero outcomes     =      40967
Nonzero outcomes  =       6602
Wald chi2(28)     =   17650.53
Log likelihood = -14452.672                     Prob > chi2       =     0.0000
------------------------------------------------------------------------------
school_left |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
childage |  -.1256983   .0052831   -23.79   0.000    -.1360531   -.1153435
1.child_fe~e |   .0038576   .0257962     0.15   0.881     -.046702    .0544172
1.urban |   .0133847   .0309483     0.43   0.665    -.0472728    .0740421
|
caste |
2  |   1.822595      .1166    15.63   0.000     1.594063    2.051126
3  |   1.743429    .118398    14.73   0.000     1.511373    1.975484
|
1  |   .5320739   .0372011    14.30   0.000     .4591611    .6049868
2  |   .2982183   .0428693     6.96   0.000      .214196    .3822406
|
mom_edu |
1  |   1.104712   .0750107    14.73   0.000     .9576939    1.251731
2  |   .7798729   .0818774     9.52   0.000     .6193962    .9403496
|
wage_less_~y |  -.0207552   .0052278    -3.97   0.000    -.0310015   -.0105088
wage_compl~5 |  -.0049688   .0038456    -1.29   0.196     -.012506    .0025684
wage_compl~8 |   .0170525   .0036375     4.69   0.000     .0099231    .0241818
wage_beyon~9 |    .011931   .0018516     6.44   0.000     .0083018    .0155601
distance_p~l |  -.0294569   .0135674    -2.17   0.030    -.0560485   -.0028654
distance_m~l |  -.0046905   .0093872    -0.50   0.617     -.023089    .0137081
distance_h~l |   .0163331   .0033153     4.93   0.000     .0098353    .0228309
month_perc~s |  -.0001359   .0000245    -5.55   0.000    -.0001838   -.0000879
durat1 |  -5.423936   .1248589   -43.44   0.000    -5.668655   -5.179218
durat2 |  -4.619776   .1035593   -44.61   0.000    -4.822748   -4.416803
durat3 |  -4.371627   .1000453   -43.70   0.000    -4.567713   -4.175542
durat4 |  -3.766473   .0925507   -40.70   0.000    -3.947869   -3.585077
durat5 |  -2.848097   .0862321   -33.03   0.000    -3.017108   -2.679085
durat6 |  -3.315598     .09182   -36.11   0.000    -3.495562   -3.135634
durat7 |  -2.654227   .0876545   -30.28   0.000    -2.826027   -2.482428
durat8 |  -2.156138   .0864863   -24.93   0.000    -2.325648   -1.986628
durat9 |   -1.53362   .0858737   -17.86   0.000     -1.70193   -1.365311
durat10 |   -.831892   .0873459    -9.52   0.000    -1.003087   -.6606973
durat11 |  -2.114238   .1222815   -17.29   0.000    -2.353905   -1.874571
------------------------------------------------------------------------------

Since I want the predicted hazards, I type the following
predict h,p
However, when I run the check as calculate the p by hand, using the formula

p=1-exp(-exp(b'x+duration parameters))
I get very different values from h.
But they should be the same. I would appreciate any help in figuring
out what am I missing here.