|
Message/Author |
|
Jon Heron posted on Friday, April 22, 2016 - 5:38 am
|
|
|
Hi Bengt, I'm currently pondering the final model from your 2011 STAR-D paper (the one resembling an Olsen+Schafer model). Do you think it's in any way important that there is symmetry in the use of growth factors for the two processes? Admittedly the figure looks more pleasing if both Y and M are modelled using I/S/Q but in our example we will probably need to forego the polynomial altogether for missingness and stick with a simple longitudinal latent class. Any thoughts? cheers, Jon |
|
|
I don't see why one would impose the restriction to have the same number of growth factors for the two processes. |
|
Jon Heron posted on Monday, April 25, 2016 - 2:26 am
|
|
|
that's good, I couldn't think of a reason either best, Jon |
|
Back to top |
|
|