Too Few Categories Problem

[Luke Came]

Hi,

I am trying to run a mlogit regression using "Status" as my dependent variable, status has 3 outcomes 1=in school 2=working 3=Not in either, independent variables "wage" and an identifier variable "id" when I try to regress this using "mlogit status i.id wage, baseoutcome (1)", STATA gives me an error r(148) "too few categories", and I can't understand why/how to fix, can anyone please help?

Thanks

[Andrew Musau]

Looks like you only have one category in status. Can you post what you get when you type

 tab status

[Luke Came]

see attached screenshot, it looks ok, there. After playing around if I removed the wage variable it worked, but Im not sure why that would be. Thanks for the help.

[Andrew Musau]

I suppose that wage is observable for only those individuals who work. If you have missing values for wage when status is equal to 1 or 2, change them to zero and re-run the estimation with wage. It should work.

Here is an illustration

. clear

. webuse sysdsn1
(Health insurance data)

. tab insure

     insure |      Freq.     Percent        Cum.
------------+-----------------------------------
  Indemnity |        294       47.73       47.73
    Prepaid |        277       44.97       92.69
   Uninsure |         45        7.31      100.00
------------+-----------------------------------
      Total |        616      100.00

. gen premium= 100 if insure==1
(350 missing values generated)

. mlogit insure age male nonwhite i.site premium

note: premium omitted because of collinearity
too few categories
r(148);


. replace premium=0 if premium==.
(350 real changes made)

. mlogit insure age male nonwhite i.site premium

Iteration 0:   log likelihood = -555.85446  
Iteration 1:   log likelihood = -142.57969  
Iteration 2:   log likelihood =  -126.0916  
Iteration 3:   log likelihood = -123.10222  
Iteration 4:   log likelihood =   -122.375  
Iteration 5:   log likelihood = -122.24066  
Iteration 6:   log likelihood = -122.21004  
Iteration 7:   log likelihood = -122.20259  
Iteration 8:   log likelihood = -122.20106  
Iteration 9:   log likelihood = -122.20073  
Iteration 10:  log likelihood = -122.20066  
Iteration 11:  log likelihood = -122.20064  
Iteration 12:  log likelihood = -122.20064  

Multinomial logistic regression                   Number of obs   =        615
                                                  LR chi2(12)     =     867.31
                                                  Prob > chi2     =     0.0000
Log likelihood = -122.20064                       Pseudo R2       =     0.7802

------------------------------------------------------------------------------
      insure |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Indemnity    |  (base outcome)
-------------+----------------------------------------------------------------
Prepaid      |
         age |     .00079   89.06307     0.00   1.000    -174.5596    174.5612
        male |  -.0121135   2939.865    -0.00   1.000    -5762.041    5762.017
    nonwhite |   .0252292   3378.219     0.00   1.000    -6621.162    6621.212
             |
        site |
          2  |   .1737362   3119.245     0.00   1.000    -6113.433    6113.781
          3  |   .0000592    3220.04     0.00   1.000    -6311.162    6311.162
             |
     premium |  -.4144542   27.25635    -0.02   0.988    -53.83591      53.007
       _cons |    20.4515   4847.597     0.00   0.997    -9480.665    9521.568
-------------+----------------------------------------------------------------
Uninsure     |
         age |   .0017614   89.06307     0.00   1.000    -174.5586    174.5622
        male |  -.0677585   2939.865    -0.00   1.000    -5762.097    5761.961
    nonwhite |  -.7480398   3378.219    -0.00   1.000    -6621.935    6620.439
             |
        site |
          2  |  -1.126736   3119.245    -0.00   1.000    -6114.734     6112.48
          3  |   .4313781    3220.04     0.00   1.000    -6310.731    6311.593
             |
     premium |  -.4109094   44.11962    -0.01   0.993    -86.88377    86.06196
       _cons |    18.9896   4847.597     0.00   0.997    -9482.127    9520.106
------------------------------------------------------------------------------

[vendula_howg]

Hi, having the same issue here. Replacing "." with "0" doesn`t help though. Using Stata 14. Any idea what might be going on anyone? Many thanks in advance!!

mlogit

[Andrew Musau]

Your description is not sufficient to know whether your problem is similar to #1. Please review the FAQs and post exactly what you typed and the resulting Stata output in its entirety. Enclose this within CODE delimiters.