-4

So I'm in a bit of a trouble. I need to put constraint on my data in state to run a mlogit, but it keeps stating that the equation have not been found. Can someone help?

enter image description here

dimitriy
  • 9,077
  • 2
  • 25
  • 50
Emma
  • 1
  • 2
    Welcome to Stack Overflow! Please read [ask] and how to create a [mcve]. Then edit your question accordingly. –  Jul 06 '18 at 13:34

1 Answers1

1

It is hard to determine exactly what is going wrong since you don't provide complete output or an MCVE. Error 303 means:

You referred to a coefficient or stored result corresponding to an equation or outcome that cannot be found.

This means you probably are defining the constraint incorrectly by referring to the numeric values rather than the corresponding value label that Stata used to label each equation.

Instead of 

constraint 1 [insure=3]:age =-1*[insure=2]:age

try something like this:

. webuse sysdsn1
(Health insurance data)

. label list insure 
insure:
           1 Indemnity
           2 Prepaid
           3 Uninsure

. constraint 1 [Uninsure]:age =-1*[Prepaid]:age

. mlogit insure age male i.site, constraints(1) nolog

Multinomial logistic regression                 Number of obs     =        615
                                                Wald chi2(7)      =      20.43
Log likelihood = -544.32915                     Prob > chi2       =     0.0047

 ( 1)  [Prepaid]age + [Uninsure]age = 0
------------------------------------------------------------------------------
      insure |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
Indemnity    |  (base outcome)
-------------+----------------------------------------------------------------
Prepaid      |
         age |  -.0054932   .0046759    -1.17   0.240    -.0146579    .0036714
        male |   .4673414   .1980637     2.36   0.018     .0791436    .8555392
             |
        site |
          2  |  -.0030463   .2045162    -0.01   0.988    -.4038907     .397798
          3  |  -.4001761   .2179144    -1.84   0.066    -.8272804    .0269282
             |
       _cons |   .1940359   .2672973     0.73   0.468    -.3298572     .717929
-------------+----------------------------------------------------------------
Uninsure     |
         age |   .0054932   .0046759     1.17   0.240    -.0036714    .0146579
        male |   .4015065   .3642457     1.10   0.270    -.3124019    1.115415
             |
        site |
          2  |  -1.192478   .4670081    -2.55   0.011    -2.107797   -.2771591
          3  |  -.1393011   .3592712    -0.39   0.698    -.8434596    .5648575
             |
       _cons |  -1.868834   .3596923    -5.20   0.000    -2.573818    -1.16385
------------------------------------------------------------------------------
dimitriy
  • 9,077
  • 2
  • 25
  • 50
  • Nice catch. Although i suspect we will never know if that was the problem indeed. –  Jul 10 '18 at 21:39
  • @PearlySpencer Using the first version of the constraint will yield the same 303 error, but I agree this is not definitive. – dimitriy Jul 10 '18 at 21:41