|
|
 |
| Assessing internal validity of a LCGA... |
 
|
| Message/Author |
|
|
|
|
Dear Muthen and team,
I have used two substudies within a large longitudinal cohort to derive trajectories using the LCGA. To assess whether the results were robust, I subsequently repeated the same analysis in each one of the two studies separately. I was asked to provide the confusion matrix of study A vs B classes to support that my main results were robust. The information I have got was to use the study A-LCGA classifier to group the trajectories of the study B.
That was not very clear to me and I could not find any additional information to help me with how to do that in Mplus. Your advice would be greatly appreciated. |
|
|
|
|
| I would recommend looking at the classifications for each individual: study A v.s. study (A&B) and study B v.s study (A&B). You can use the most likely class or the posterior membership probability. |
|
|
| Back to top |
|
|
