Announcement

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

  • GEE (working correlation, multiple imputation)

    Hello! This is my first post.

    I am Stata beginner.





    Using Program: Stata 13.0.

    Method: GEE

    Data: 5waves (unbalnaced)
    Obs: 27,246


    Situation:

    1) My data has missing values (0.01%~2.96%)
    2) I did Little’s MCAR test => result : not MCAR
    3) So I did multiple imputations. => I made it.

    mi set mlong
    mi register imputed health1stclf i.caretran health1stclb i.edub i.genderb agerealb sposatb i.workb i.religionynb i.catime_w_binab lifehabitb i.yearf
    mi impute mvn health1stclf i.caretran health1stclb i.edub i.genderb agerealb sposatb i.workb i.religionynb i.catime_w_binab lifehabitb i.yearf




    Problem:

    1) I tried to do QIC after multiple imputations to find good working correlation.

    * DV: health1stclf

    xi: qic health1stclf i.caretran health1stclb i.edub i.genderb agerealb sposatb i.workb i.religionynb i.catime_w_binab lifehabitb i.yearf, i(pid) j(time) family(gaussian) link(identity) corr(independent)

    * I got error message


    no; data are mi set
    Use mi xtset to set or query these data; mi xtset has the same syntax as xtset.

    Perhaps you did not type xtset. Some commands call xtset to obtain information about the settings. In that case, that
    command is not appropriate for running directly on mi data. Use mi extract to select the data on which you want to run the
    command, which is probably m=0.

    r(119)



    ///////

    2) So, I tried again in this way

    mi estimate: qic health1stclf i.caretran health1stclb i.edub i.genderb agerealb sposatb i.workb i.religionynb i.catime_w_binab lifehabitb i.yearf, i(pid) j(time) family(gaussian) link(identity) corr(independent)


    But I got an error message again.


    mi estimate: command not supported
    qic is not officially supported by mi estimate; see mi estimation for a list of Stata estimation commands that are
    supported by mi estimate. You can use option cmdok to allow estimation anyway.
    r(198);

    /////////////////


    Are there any ways to use QIC for multiple imputations-gee?

    Or how can I find "goodness of fit" for working correlation for imputated dataset?



    Many thanks for your advice.



  • #2
    Hello! This is my first post too. And I'm visiting this thread to follow up on this subject.

    I have a particular question that I've been trying to figure out in the last couple of days:
    How do I perform a model selection in mi estimate with xtgee if mi estimate does not support qic command?

    It doesn't work with cmdok option either.

    I already consulted and it seems like qic is an rclass command.

    If anyone can give me any input on this, I would very much appreciate it.

    Thank you,

    Natalia.

    Comment

    Working...
    X