Dear all,
I have fitted a simplex model with six time points. It's a threshold model with 2 thresholds. I fixed the thresholds and freely estimated the means. I used the CSOLNP optimizer.
I get stable estimates, they are in accordance with what I would expect based on the univariate results, and the gradients look ok.
As a next step, I wanted to calculate confidence intervals, but I only get this:
confidence intervals:
lbound estimate ubound note
atm21 NA 0.4871751 NA !!!
I had many problems with fitting threshold models in openMx before, so I was wondering whether calculating those CIs is a realistic plan anyway, and whether there are any tricks to get the model with CIs run properly?
Many thanks in advance!
Kind regards
Charlotte