Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • The independence of irrelevant alternatives assumpition test of multinominal logit has an error

    Dear statalist users:
    The independence of irrelevant alternatives assumpition test using "mlogtest, hausman base" command has an error after running the multinominal logit regression.
    My dependent variable is a categorical variables with 5 outcomes. My independent variables are dummy variables and continuous variables. For example, education which is one of the independent variables is has 5 levels- illiteracy, primary school, junior high school, high school, university. I choose illiteracy as the reference group and set 4 dummy variables Edu1, Edu2, Edu3, Edu4.
    I run the regression by "
    Code:
     mlogit Y X1  X2 Edu1 Edu2 Edu3 Edu4 ...
    ". I use "
    Code:
    mlogtest, hausman base
    " to test the independence of irrelevant alternatives assumpition. However, it shows the error as follow:
    Code:
    mlogtest, hausman base
    Problem determining number of categories.
    **** Hausman tests of IIA assumption
     Ho: Odds(Outcome-J vs Outcome-K) are independent of other alternatives.
    You used the old syntax of hausman. Click here to learn about the new syntax.
    (storing estimation results as _HAUSMAN)
    factor variables and time-series operators not allowed
    r(101);
    Why is this happening? How to solve it? I am looking forward to your help. Thank you very much.

  • #2
    Welcome to Statalist.

    Apparently the mlogtest command is part of the user-written spost13 package described by

    Code:
    net describe spost13_ado, from(http://www.indiana.edu/~jslsoc/stata)
    That description gives an email address for support, and you may perhaps be able to get help by contacting the authors directly.

    The various messages you show in your output from the command are not immediately obvious. But, I don't think you have shown us the actual mlogit command you ran, because I don't think it is possible to have X1 (with a subscript) as a Stata variable name.

    I am concerned about the message that factor variables are not allowed. That suggests you may either have an old version of the software or an old version of Stata.

    On my fully updated system
    Code:
    . which mlogtest
    /Users/lisowskiw/Library/Application Support/Stata/ado/plus/m/mlogtest.ado
    *! version 3.5.0 2016-10-25 | long freese | problem in Stata 14.2 smhsiao
    so if on your system mlogtest is an earlier version, run
    Code:
    adoupdate spost13_ado
    Or perhaps you are running an old version of Stata. Stata 15.1 is the current version.


    Comment


    • #3
      Dear William, thank you very much. I solved the problem according to your advice. I updated the Stata14.1 and downloaded the spost13_ado in the author's website.

      Comment

      Working...
      X