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

# st: cmp issue

 From "Dr. Bartus" To statalist@hsphsun2.harvard.edu Subject st: cmp issue Date Mon, 07 May 2012 21:47:53 +0200

Hi David,

I encountered a problem I cannot solve: I estimated an interval regression model, but then cmp refuses
to estimate the same model saying the second dependent variable is collinear with some of the predictors.
(I got error message 459)

Here is the log file:

. intreg bdurlo bdurhi ib3.pedu ib3.edu c.bage##c.bage ib1.cohort if sample==1 & nchild==1

Fitting constant-only model:

Iteration 0:   log likelihood = -2116.2241
Iteration 1:   log likelihood = -1271.5175
Iteration 2:   log likelihood = -1253.0027
Iteration 3:   log likelihood =   -1252.86
Iteration 4:   log likelihood =   -1252.86

Fitting full model:

Iteration 0:   log likelihood =  -1882.655
Iteration 1:   log likelihood =   -1057.84
Iteration 2:   log likelihood = -1015.1725
Iteration 3:   log likelihood = -1009.9485
Iteration 4:   log likelihood = -1009.1258
Iteration 5:   log likelihood =  -1008.972
Iteration 6:   log likelihood = -1008.9421
Iteration 7:   log likelihood = -1008.9358
Iteration 8:   log likelihood = -1008.9346
Iteration 9:   log likelihood = -1008.9343
Iteration 10:  log likelihood = -1008.9343
Iteration 11:  log likelihood = -1008.9343

Interval regression                               Number of obs   =       1422
LR chi2(15)     =     487.85
Log likelihood = -1008.9343                       Prob > chi2     =     0.0000

------------------------------------------------------------------------------
|      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
pedu |
0  |   .3249509    .173096     1.88   0.060     -.014311    .6642128
1  |  -.3082846   .2430748    -1.27   0.205    -.7847026    .1681333
2  |  -.0748098   .1601104    -0.47   0.640    -.3886203    .2390008
4  |  -.4076649   .1912106    -2.13   0.033    -.7824309    -.032899
|
edu |
1  |  -.1103826   .1958593    -0.56   0.573    -.4942597    .2734945
2  |   .0840579   .1508878     0.56   0.577    -.2116767    .3797925
4  |  -.1562882   .1580749    -0.99   0.323    -.4661093    .1535328
|
bage |  -.0473836   .1381719    -0.34   0.732    -.3181955    .2234282
|
c.bage#|
c.bage |  -.0007055   .0025894    -0.27   0.785    -.0057808    .0043697
|
cohort |
2  |  -.1636366    429.618    -0.00   1.000    -842.1995    841.8722
3  |  -.4902144   465.0304    -0.00   0.999    -911.9331    910.9527
4  |  -5.065201   306.9131    -0.02   0.987    -606.6037    596.4733
5  |  -6.595311    306.913    -0.02   0.983    -608.1337    594.9431
6  |  -7.341457    306.913    -0.02   0.981    -608.8799     594.197
7  |  -8.193675    306.913    -0.03   0.979    -609.7321    593.3448
|
_cons |   14.52831   306.9176     0.05   0.962    -587.0192    616.0758
-------------+----------------------------------------------------------------
/lnsigma |   .0798116   .0632361     1.26   0.207    -.0441289    .2037521
-------------+----------------------------------------------------------------
sigma |   1.083083     .06849                      .9568306    1.225994
------------------------------------------------------------------------------

Observation summary:         0  left-censored observations
0     uncensored observations
1269 right-censored observations
153       interval observations

. cmp ( bdurlo bdurhi = ib3.pedu ib3.edu c.bage##c.bage ib1.cohort ) if sample==1 & nchild==1 , ind(
> 7)

Fitting individual models as starting point for full model fit.
Note: For programming reasons, these initial estimates may deviate from your specification.
For exact fits of each equation alone, run cmp separately on each.
bdurhi collinear with 0.pedu 1.pedu 2.pedu 4.pedu 1.edu 2.edu 4.edu bage c.bage#c.bage 4.cohort
5.cohort 6.cohort _cons
r(459);

Could you give some hints what went wrong?

Tamas

--------------------------------------------

Tamas Bartus, PhD
Associate Professor, deputy director
Institute of Sociology and Social Policy
Corvinus University, Budapest
1093 Budapest, Közraktár utca 4-6.
Room 423.
Phone: +36-1-482-7301
Fax: +36-1-482-7348
Homepage: http://web.uni-corvinus.hu/bartus

• Follow-Ups:
• st: apology
• From: "Dr. Bartus" <tamas.bartus@uni-corvinus.hu>