|
![](http://www.statmodel.com/discussion/icons/csc/1px.gif) |
GGMM equations for ordered & continuo... |
![Previous](http://www.statmodel.com/discussion/icons/csc/navgen_arrow_left.gif) ![](http://www.statmodel.com/discussion/icons/csc/1px.gif)
|
Message/Author |
|
|
The equation for binary distal outcomes (for GGMM) is provided in several sources, but I cannot find the equations for ordered and continuous distal variables (with hopefully at least brief explanations). Can someone point me in the direction of these? Much thanks. |
|
|
For continuous and ordinal distals predicted by a latent class variable I don't think we have equations written out per se in a publication. But it is simple to describe that they would be. This is based on a distal outcome being just like another latent class indicator. For a continuous distal you have mu_k where k indexes the different classes (so the mean varies across classes). Variances can also vary across classes. For an ordinal distal you have tau_{ck} where tau refers to the set of thresholds that vary both across the C categories of the distal (c=0, 1, ...C-1; thresholds ordered) and across the classes k of the latent class variable. The relationship between thresholds and probabilities for various categories of the distal is shown in the appendices to the UG (e.g. appendix 1). |
|
Back to top |
|
|