mlogtest

[Robert Kolesar]

Hello,

I am using Stata 14.2 and installed mlogtest (spost13). I am hoping somebody may have experience with mlogtest.

I execute my mlogit regression and then attempt mlogtest, lr/all/etc.. and get the following error message: "data has been altered since mlogit was estimated". I am able to execute mlogitgof just fine. I tried reinstalling the command and that didn't help.

Thanks,

Robert

[Richard Williams]

In Stata 14.2, I can't replicate:

. webuse nhanes2f, clear

. mlogit health weight height i.female

Iteration 0:   log likelihood = -15764.397  
Iteration 1:   log likelihood = -15493.985  
Iteration 2:   log likelihood = -15491.005  
Iteration 3:   log likelihood = -15491.004  

Multinomial logistic regression                 Number of obs     =     10,335
                                                LR chi2(12)       =     546.79
                                                Prob > chi2       =     0.0000
Log likelihood = -15491.004                     Pseudo R2         =     0.0173

------------------------------------------------------------------------------
      health |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
poor         |
      weight |   .0078795   .0029609     2.66   0.008     .0020763    .0136826
      height |  -.0362365   .0064367    -5.63   0.000    -.0488521   -.0236208


[Rest of output deleted]


. mlogtest, lr

LR tests for independent variables (N=10335)

  Ho: All coefficients associated with given variable(s) are 0

                 |      chi2    df   P>chi2
-----------------+-------------------------
          weight |   171.905     4    0.000
          height |   460.599     4    0.000
        1.female |   104.950     4    0.000

I suggest you show your code and output, using code tags like I just did. Better yet, provide a replicable example.

Also, here is my version of mlogtest.

. which mlogtest
c:\ado\plus\m\mlogtest.ado
*! version 3.5.0 2016-10-25 | long freese | problem in Stata 14.2 smhsiao

[Robert Kolesar]

In Stata 14.2, I am attempting to use mlogtest (spost13) immediately following my mlogit command. I am unable to get a result as I get a return message that the "data has been altered since mlogit was estimated."

. mlogit TIMINGANC c.ageM_del c.ageM_del2 ib(1).b1_01 ib(1999).b2_01 ib(1).v106 ib(1).v701 i.v501 ib(5
> ).v190 ib(1).v714 radio c.v201 i(2).v025 [iweight=wgt] if b2_01>=1999

note: 1999.b2_01 identifies no observations in the sample
note: 2008.b2_01 omitted because of collinearity
Iteration 0:   log likelihood = -7133.6382 
Iteration 1:   log likelihood = -6542.1356 
Iteration 2:   log likelihood = -6493.4262 
Iteration 3:   log likelihood = -6491.3481 
Iteration 4:   log likelihood = -6491.3407 
Iteration 5:   log likelihood = -6491.3407 

Multinomial logistic regression                 Number of obs     =      6,369
                                                LR chi2(60)       =    1284.60
                                                Prob > chi2       =     0.0000
Log likelihood = -6491.3407                     Pseudo R2         =     0.0900

--------------------------------------------------------------------------------------
           TIMINGANC |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
---------------------+----------------------------------------------------------------
no_ANC               |
            ageM_del |  -.0052072   .0029076    -1.79   0.073     -.010906    .0004917
           ageM_del2 |   5.34e-06   3.92e-06     1.36   0.173    -2.34e-06     .000013
                     |
               b1_01 |
                  2  |   .1358732    .128943     1.05   0.292    -.1168504    .3885968
                  3  |  -.0481839   .1272047    -0.38   0.705    -.2975005    .2011327
                  4  |  -.5314377   .1326889    -4.01   0.000     -.791503   -.2713723
                  5  |  -.1642282   .1259953    -1.30   0.192    -.4111744    .0827179
                  6  |  -.1408556   .1294425    -1.09   0.277    -.3945582    .1128471
                  7  |  -.5968642   .1361429    -4.38   0.000    -.8636995    -.330029
                  8  |  -.2778816   .1320377    -2.10   0.035    -.5366708   -.0190925
                  9  |  -.1397301   .1385683    -1.01   0.313     -.411319    .1318587
                 10  |  -.4688562   .1438285    -3.26   0.001     -.750755   -.1869575
                 11  |   -.478483   .1461263    -3.27   0.001    -.7648853   -.1920807
                 12  |  -.3625287   .1525743    -2.38   0.017    -.6615688   -.0634886
                     |
               b2_01 |
               1999  |          0  (empty)
               2004  |   .4835833   .1116735     4.33   0.000     .2647072    .7024593
               2005  |   .0571927   .1016433     0.56   0.574    -.1420246    .2564099
               2006  |   .3193181   .0938224     3.40   0.001     .1354295    .5032067
               2007  |   .0027357   .0880125     0.03   0.975    -.1697657    .1752371
               2008  |          0  (omitted)
                     |
                v106 |
       no education  |   .3711169   .0751965     4.94   0.000     .2237345    .5184993
          secondary  |  -.9216353   .2230225    -4.13   0.000    -1.358751   -.4845192
                     |
                v701 |
       no education  |   .2082547   .0650121     3.20   0.001     .0808333    .3356761
          secondary  |  -.1095411   .1233797    -0.89   0.375    -.3513609    .1322787
                     |
                v501 |
living with partner  |   .5377412   .2732295     1.97   0.049     .0022212    1.073261
                     |
                v190 |
            poorest  |   .5206591   .1361256     3.82   0.000     .2538578    .7874603
             poorer  |   .1537373   .1348572     1.14   0.254     -.110578    .4180526
             middle  |  -.0392743   .1350085    -0.29   0.771     -.303886    .2253375
             richer  |   -.088918   .1345031    -0.66   0.509    -.3525391    .1747032
                     |
                v714 |
                 no  |   .2556752   .0677776     3.77   0.000     .1228336    .3885168
                     |
               radio |  -.5095089   .0727285    -7.01   0.000    -.6520541   -.3669637
                v201 |   .1178838   .0195737     6.02   0.000       .07952    .1562476
                     |
                v025 |
              rural  |    .792258   .1635846     4.84   0.000     .4716382    1.112878
               _cons |  -.7273508   .5452793    -1.33   0.182    -1.796079    .3413769
---------------------+----------------------------------------------------------------
early_1st_vis        |
            ageM_del |  -.0051246   .0033835    -1.51   0.130    -.0117561    .0015069
           ageM_del2 |   .0000105   4.71e-06     2.22   0.026     1.24e-06    .0000197
                     |
               b1_01 |
                  2  |   .1488715   .1674503     0.89   0.374     -.179325     .477068
                  3  |   .0859704   .1663663     0.52   0.605    -.2401015    .4120424
                  4  |   .1981116   .1579686     1.25   0.210    -.1115013    .5077244
                  5  |   .2666855   .1570177     1.70   0.089    -.0410635    .5744346
                  6  |   .0631782   .1675003     0.38   0.706    -.2651164    .3914728
                  7  |   .3131532   .1612383     1.94   0.052     -.002868    .6291744
                  8  |   .2145203   .1643556     1.31   0.192    -.1076108    .5366514
                  9  |   .4846292   .1678566     2.89   0.004     .1556364     .813622
                 10  |   .1358122    .175157     0.78   0.438    -.2074892    .4791136
                 11  |   .5668669   .1696695     3.34   0.001     .2343208     .899413
                 12  |   .2968105   .1779236     1.67   0.095    -.0519132    .6455343
                     |
               b2_01 |
               1999  |          0  (empty)
               2004  |  -.1587066   .1366357    -1.16   0.245    -.4265076    .1090944
               2005  |  -.0932128   .1165017    -0.80   0.424     -.321552    .1351264
               2006  |   .0154549   .1099811     0.14   0.888     -.200104    .2310139
               2007  |   -.163443   .1018473    -1.60   0.109      -.36306    .0361739
               2008  |          0  (omitted)
                     |
                v106 |
       no education  |  -.1467074   .0865719    -1.69   0.090    -.3163852    .0229704
          secondary  |   .1314817   .1289474     1.02   0.308    -.1212505     .384214
                     |
                v701 |
       no education  |   .1763101   .0813121     2.17   0.030     .0169413    .3356788
          secondary  |  -.0960861   .1126574    -0.85   0.394    -.3168905    .1247183
                     |
                v501 |
living with partner  |  -.1161659   .3191297    -0.36   0.716    -.7416485    .5093168
                     |
                v190 |
            poorest  |  -.2402368   .1496471    -1.61   0.108    -.5335398    .0530662
             poorer  |  -.3298236   .1450933    -2.27   0.023    -.6142013    -.045446
             middle  |  -.1360701   .1408515    -0.97   0.334    -.4121339    .1399938
             richer  |  -.1927966   .1354894    -1.42   0.155    -.4583509    .0727578
                     |
                v714 |
                 no  |   .0285673   .0754126     0.38   0.705    -.1192387    .1763733
                     |
               radio |  -.0837297   .0769078    -1.09   0.276    -.2344662    .0670068
                v201 |  -.1290656   .0243993    -5.29   0.000    -.1768874   -.0812438
                     |
                v025 |
              rural  |  -.5841399   .1325214    -4.41   0.000    -.8438771   -.3244027
               _cons |   .7108815    .614104     1.16   0.247    -.4927402    1.914503
---------------------+----------------------------------------------------------------
late_1st_vis         |  (base outcome)
--------------------------------------------------------------------------------------

.
end of do-file

. do "C:\Users\Robert\AppData\Local\Temp\STD06000000.tmp"

. mlogtest, lr

data has been altered since mlogit was estimated

[Richard Williams]

What happens if you drop iweight? Or the if? Let’s try to isolate what is causing the problem and then figure out how to solve it.

[Richard Williams]

Why are you using iweights anyway? Pweights and fweights are much more common.

[Robert Kolesar]

Thanks Richard. The issue seems to be the iweights as when I remove it I am able to get a results. I am using iweights as I am working with survey data.

[Richard Williams]

If you are working with survey data, you should probably use pweights or svyset the data and use the svy prefix.

[Robert Kolesar]

Thanks Richard,

I think the command cannot handle survey data. I svyset the data and used the prefit (removing the weight from the regression equation) and got:

. mlogtest, lr
LR tests cannot be run with svy estimation

Thanks anyway.

Robert

[Richard Williams]

LR definitely won't work with svy. Try the Wald option. And if it doesn't work you can do it with test commands. Long and Freese's book is pretty good at showing how different things could be done if you didn't have the spost commands. https://www.stata.com/bookstore/regr...ent-variables/

[Robert Kolesar]

Ok, thanks much for the resource.

[Richard Williams]

mlogtest seems to work, and you don't even need to specify the wald option.

webuse nhanes2f, clear
svy: mlogit health weight height i.female
mlogtest

. mlogtest

Wald tests for independent variables (N=10335)

  Ho: All coefficients associated with given variable(s) are 0

                 |         F    df   df_r      P>F
-----------------+--------------------------------
          weight |    23.317     4      4    0.000
          height |    53.000     4      4    0.000
        1.female |    10.297     4      4    0.000

[Richard Williams]

Incidentally, in your earlier code you were using if to select cases. Don't do that with svy. Use subpop instead. See

https://stats.idre.ucla.edu/stata/fa...data-in-stata/

[Nawazuddin Ahmed]

In Stata 14.2 after successfully running my mlogit command, I want to run mlogtest with some options like iia, lr, lrcomb, combine etc.
But, I am getting this:
. mlogtest, iia

Hausman tests of IIA assumption (N=27585)

Ho: Odds(Outcome-J vs Outcome-K) are independent of other alternatives
invalid syntax
_______

after running which mlogtest, I getting this


. which mlogtest
c:\ado\plus\m\mlogtest.ado
*! version 3.5.0 2016-10-25 | long freese | problem in Stata 14.2 smhsiao
____

Where is the problem?

thank you