Closed Pip-griff closed 4 years ago
Thanks, Pip. This is actually by design since bfactor()
provides a specific data-reduction integration approach that doesn't generalize to everything else that mirt supports; hence, this function has been isolated from the more central mirt()
function. That said, you can reproduce the fitted bfactor()
models in mirt()
just by specifying a suitable syntax and using the 'MHRM'
algorithm as you've requested.
Thanks for this Phil,
I have now found the google groups and I will post there in future.
Now I am off to try and build the model in mirt()
Hi Phil, I may be wrong here (and I am not sure this is an issue as such), but it seems that the bfactor model cannot be set up to use model = "MHRM". I have tried running a longitudinal IRT model with between 3 and 7 timepoints in mirt and the quadrature is an issue. Either quadrature is is too low to reach a solution, or too high for memory allocation.
Could bfactor be set up to include MHRM, or is there something which would prevent stochastic methods here?
thanks! Pip