Message/Author |
|
Corinna posted on Tuesday, June 23, 2015 - 7:28 am
|
|
|
Dear Linda & Bengt, I have a binary outcome in GMM. I know that with a categorical variable I do not have intercepts but thresholds. However, I do not understand the meaning of variances with a binary outcome. To what does the variance relate? To the probabilities? What is the meaning of a variance with a binary outcome? I would be very thankful for a short answer. With best regards, Corinna |
|
|
Variances of categorical outcomes are not model parameters. |
|
Corinna posted on Tuesday, June 23, 2015 - 9:19 am
|
|
|
Dear Linda, I am not sure if I expressed myself correctly. I am talking about residual variances of the growth parameters (intercept and slope) if you have a binary outcome. If they are not model parameters, does that mean that I do not have to consider them at all in my model analyses? I am trying to find the meaning of these variances in a binary outcome... but if I do not have to consider them then perhaps I can just fix the variances and calculate and LCGA? Thank you very much |
|
|
The growth factors are continuous latent variables not categorical. Their variances/residual variances are interpreted in the regular way. |
|
Back to top |