+ Reply to Thread
Results 1 to 2 of 2

Thread: Variance components on logit scale

  1. #1
    Points: 8,343, Level: 61
    Level completed: 65%, Points required for next Level: 107

    Posts
    278
    Thanks
    14
    Thanked 7 Times in 7 Posts

    Variance components on logit scale




    Hello all,

    I have data with a binary response (0,1), a categorical independent variable with 3 levels (treatment i.e. an experimental treatment and two controls) and a random effect (can be a center from multicenter study or any other random effect).

    I ran a GLMM (generalized linear mixed model) using SAS proc Glimmix.

    On the output, one of the results is the variance estimate of the random effect, under the table of "Covariance parameter estimates". The number I got was 0.325. I understand from reading on the procedure, that this number is the variance estimate of the random effect, on the logit scale.

    What I want to do, is to try and make a meaningful interpretation of this number, and I need your help with it. I read somewhere, that if I call this number V, I can calculate the "intraclass correlation coefficient" using the formula:

    ICC=V/(V+3.29)

    I can't say why...what I want to ask you, is:

    1. Do you know this formula ?
    2. Independently to the formula or to ICC, how would you make an interpretation, a conclusion without the words "logit scale" ?

    Thanks !

  2. #2
    Points: 8,343, Level: 61
    Level completed: 65%, Points required for next Level: 107

    Posts
    278
    Thanks
    14
    Thanked 7 Times in 7 Posts

    Re: Variance components on logit scale


    I want to correct myself in one point, the measure I described is the variance of the random factor intercepts on the logit scale

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats