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

# RE: st: Multinomial Logit Model - Cramer Ridder Test

 From Ian Li To Statalist Subject RE: st: Multinomial Logit Model - Cramer Ridder Test Date Mon, 14 Nov 2011 07:55:28 +0000

```Thanks, but that is not it. I initially estimated a model with observations for multiple years in it, and included dummy variables for years, but got the same error message and thus restricted the sample and omitted the year dummies.

For illustration, running the 'full' sample gives me:

. mlogit overed female noneng nonaust if empnat1==1 & overed!= 0

Iteration 0:   log likelihood = -869628.84
Iteration 1:   log likelihood = -864292.36
Iteration 2:   log likelihood = -857382.33
Iteration 3:   log likelihood = -857052.63
Iteration 4:   log likelihood = -857050.54
Iteration 5:   log likelihood = -857050.54

Multinomial logistic regression                   Number of obs   =     566758
LR chi2(18)     =   25156.61
Prob > chi2     =     0.0000
Log likelihood = -857050.54                       Pseudo R2       =     0.0145

------------------------------------------------------------------------------
overed |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
1            |  (base outcome)
-------------+----------------------------------------------------------------
2            |
female |  -.0665102   .0078792    -8.44   0.000    -.0819531   -.0510672
noneng |  -.1927152   .0113891   -16.92   0.000    -.2150374    -.170393
nonaust |  -.3274604   .0285201   -11.48   0.000    -.3833587   -.2715621
_cons |  -.6380651   .0064792   -98.48   0.000     -.650764   -.6253662
-------------+----------------------------------------------------------------
3            |
female |  -.3193849   .0077407   -41.26   0.000    -.3345564   -.3042135
noneng |   .1913581   .0103309    18.52   0.000     .1711098    .2116064
nonaust |   .7069291   .0200759    35.21   0.000      .667581    .7462771
_cons |  -.5751631   .0062579   -91.91   0.000    -.5874283    -.562898
-------------+----------------------------------------------------------------
4            |
female |  -.0578355   .0076683    -7.54   0.000    -.0728651    -.042806
noneng |   .1304178   .0102257    12.75   0.000     .1103758    .1504598
nonaust |    .678697   .0200446    33.86   0.000     .6394102    .7179838
_cons |  -.6413862   .0063754  -100.60   0.000    -.6538817   -.6288906
-------------+----------------------------------------------------------------
5            |
female |  -.1673004   .0137346   -12.18   0.000    -.1942198    -.140381
noneng |   .1705464   .0179613     9.50   0.000     .1353429    .2057499
nonaust |   1.225811   .0279006    43.93   0.000     1.171127    1.280495
_cons |  -2.127142   .0113192  -187.92   0.000    -2.149327   -2.104956
-------------+----------------------------------------------------------------
6            |
female |  -.4087381   .0197191   -20.73   0.000    -.4473868   -.3700895
noneng |   1.223924   .0224788    54.45   0.000     1.179866    1.267981
nonaust |   2.457085   .0262176    93.72   0.000       2.4057    2.508471
_cons |  -3.386475   .0179543  -188.62   0.000    -3.421665   -3.351286
-------------+----------------------------------------------------------------
7            |
female |  -.2350719   .0840274    -2.80   0.005    -.3997626   -.0703812
noneng |   .1498775   .1103836     1.36   0.175    -.0664703    .3662254
nonaust |   1.174268   .1602489     7.33   0.000     .8601858     1.48835
_cons |  -5.829502   .0683024   -85.35   0.000    -5.963373   -5.695632
------------------------------------------------------------------------------

. crtest
factor variables and time-series operators not allowed
r(101);

end of do-file

r(101);

I get the same error when the model includes year dummies.

I have read through the help file for 'crtest', but there appears to be no solution to my problem there.

Regards
Ian

----------------------------------------
> Date: Mon, 14 Nov 2011 12:43:33 +0500
> Subject: Re: st: Multinomial Logit Model - Cramer Ridder Test
> From: anees@aneconomist.com
> To: statalist@hsphsun2.harvard.edu
>
> try without the -if year == 2009- as the help file suggests -crtest-
> "performs the test for pooling states" after the Multinomial logit
> models. The error suggest to not use time operators and factor