Announcement

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

  • Not allocating certain respondents to certain class in latent class model

    Hi All,

    I'm now trying to do some analysis using stata for a latent class model I'm thinking of.

    I'm using SC data for my analysis. For example:


    pid gid y price contract local wknown tod seasonal
    1 1 0 7 5 0 1 0 0
    1 1 0 9 1 1 0 0 0
    1 1 0 0 0 0 0 0 1
    1 1 1 0 5 0 1 1 0
    1 2 0 7 0 0 1 0 0
    1 2 0 9 5 0 1 0 0
    1 2 1 0 1 1 0 1 0
    1 2 0 0 5 0 0 0 1



    What i want to do is to make certain respondents(distinguished by variable 'pid') not to be allocated in certain classes

    For example, if there are 10 respondents and 3 classes, i want to make some of the respondents not to be considered to be allocated in a certain class (say, class 1) at all.

    what i saw was i could use 'constraint' option for this, however it seems that only coefficient control is available.

    will setting every coefficient for the specific respondents to zero fulfill my wish?

    Or if you have any other recommendations, i'm very glad to follow

    Thank you in advance


    -Hyunhong Choi

  • #2
    One thing that would help in answering your question: What Stata program are you using to perform latent class analysis?

    Regards, Mike

    Comment


    • #3
      How would you know if the respondents could be a member of a particular class? I think it may be helpful to think through some of the underlying assumptions/mechanisms of the methods (e.g., class membership causes the observed response) in order to determine if it is the best method to answer your question.

      Comment


      • #4
        Thank you all for your answers!

        Mike/

        I'm trying to use 'lclogit', however I'm willing to use any program that can do the work


        wbuchanan/

        I'm trying to use so-called confirmatory latent class model, which gives a priori restrictions for each class

        Popular application of this confirmatory LCM is for attribute non-attendance problem, in this case they put the coefficient of certain attributes that consumer does not consider to zero

        In my case, I'm assuming some respondents with specific response pattern would not be a member of certain class

        Comment

        Working...
        X