Announcement

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

  • Best command for multilevel multinomial logistic regression models

    Dear Statalist users,
    I have Stata/SE 14.2.
    I have a pooled cross-sectional time-series dataset. Basically monthly surveys are conducted on different nationally representative samples of individuals who are nested in provinces--so not a panel data.
    I only have 7 waves of surveys.
    My dependent variable is vote choice and is a nominal variable listing four political parties.
    I would like to use mixed models as individuals from the same provinces may have intra-class correlations, though I am not sure how to handle the survey waves as the number is too few, only 7 waves.
    I was planning to add the wave a variable for fixed effects.

    Two questions:
    1) If I am to use a multinomial choice model, what would be the best command for me to use? In Stata 14, there is no multilevel mixed effects model for nominal variables (such as memlogit), and I have seen clogit, pomlogit, or fomlogit as options. The manual suggests gsem though I am having trouble applying the example given in the manual to my case.
    What would be the best command for my case?
    2) How would you recommend I handle the errors from intra-wave correlations?

    Thanks.

  • #2
    You might see if this helps.

    https://www.stata.com/stata-news/news29-2/xtmlogit/
    -------------------------------------------
    Richard Williams, Notre Dame Dept of Sociology
    StataNow Version: 19.5 MP (2 processor)

    EMAIL: [email protected]
    WWW: https://www3.nd.edu/~rwilliam

    Comment


    • #3
      Thanks much, Richard. It does help.

      Comment

      Working...
      X