Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Hausman IIA test after asclogit in Stata12

From   "Subramanian, Hemang" <>
To   "" <>
Subject   st: Hausman IIA test after asclogit in Stata12
Date   Mon, 25 Feb 2013 02:09:14 +0000

Dear Stata-listers.

   I need to run the Hausman IIA test after obtaining the results from the "asclogit" command.  Any help with respect to commands that could be followed for the Hausman test would greatly help me clear this one last hurdle in proving that my choice model has independent alternatives, and is valid. 
I have followed the threads on the same topic prior to this - but didn't find appropriate solutions for the same. On the outset I apologize if I'm reposting on the same topic, but I am facing similar issues as the previous 2 posts on the same topic. (dated 2011 & 2010) have had. Is there any user generated program that  could run the Hausman IIA after asclogit. I am aware of similar commands after xtreg, mlogit, and so on...but not in the context of IIA based testing.

 The description of my runs is given below: 
 my main asclogit commands are as follows:
1.   eststo est12:asclogit achoice ratio_simulcast buyer_distance supply  meanPriceValueRatio pctsold price_variability interactac , case(caseid) alt(ac)
2. eststo est13:asclogit achoice ratio_simulcast buyer_distance supply  meanPriceValueRatio pctsold price_variability interactac if ac != 3, case(caseid) alt(ac)

 The variables descriptions & data types are as follows:
 achoice - is the choice variable 1 or 0,  ratio_simulcast  - is a  real number between 0 - 1, , buyer_distance - a real number between 0 & 4000, supply - number between 1 & 10000 , meanPriceValueRatio - a real number between 0 & 2, pctsold - number between 0 & 1, price_variability - between 1 & 7000 , interactac  - real number between 0 & 65000:  There are 85 alternatives of the variable ac, each indicated by a unique number between 1 & 85 in the model. 
 There are approximately 1.25 M total observations with about 0.105 cases in the asclogit model - each case indicated by a caseid.

 I tried the I) hausman test and II) the suest test to obtain the desired results for testing IIA, but couldn't get past the following errors.
I) I tried using hausman est12 est13 -but got the error
 hausman est12 est14

Note: the rank of the differenced variance matrix (4) does not equal the number of coefficients being tested
        (7); be sure this is what you expect, or there may be problems computing the test.  Examine the output
        of your estimators for anything unexpected and possibly consider scaling your variables so that the
        coefficients are on a similar scale.

                 ---- Coefficients ----
             |      (b)          (B)            (b-B)     sqrt(diag(V_b-V_B))
             |     est12        est14        Difference          S.E.
ratio_simu~t |    -.360078    -.3472927       -.0127852               .
buyer_dist~e |   -.0019122    -.0023039        .0003917               .
      supply |    .0017486     .0017187        .0000299               .
meanPriceV~o |    4.13e-08     4.18e-08       -5.90e-10        1.18e-08
     pctsold |   -1.925903    -1.978081        .0521788               .
price_vari~y |    1.38e-08    -2.32e-08        3.69e-08               .
  interactac |   -.0000571    -.0000186       -.0000385               .
                        b = consistent under Ho and Ha; obtained from asclogit
         B = inconsistent under Ha, efficient under Ho; obtained from asclogit

    Test:  Ho:  difference in coefficients not systematic

                  chi2(4) = (b-B)'[(V_b-V_B)^(-1)](b-B)
                          = -1406.95    chi2<0 ==> model fitted on these
                                        data fails to meet the asymptotic
                                        assumptions of the Hausman test;
                                        see suest for a generalized test


II) Then I tried suest - 
The number of score variables does not match number of equations in model est12

I didn't get a solution for either I) or II) above. Any help would be greatly appreciated. what can I do to correct these errors,

Thanks in advance.
Hemang C S

*   For searches and help try:

© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index