|
|
Interpreting H1 Model Non-convergence |
|
Message/Author |
|
Ted Schwaba posted on Wednesday, September 14, 2016 - 6:20 pm
|
|
|
Hello, I am estimating a multiple-groups latent growth curve model across seven time points. The model estimation terminates normally, but the H1 model estimation does not converge. I have set H1iterations to 30,000 and starts to 20. Because the H1 model estimation does not converge, chi-square model fit is not computed, but all other parameters and fit statistics are estimated. Substantively, what does it mean that the H1 model estimation does not converge? Should I take caution in interpreting the model? |
|
|
Perhaps you have low or zero coverage for some pairs of variables. If it is zero you can try the multiple-group approach to multiple cohort data as in UG ex 6.18. |
|
Back to top |
|
|