When trying to fit MFA-models with ordinal indicators an error message appears saying that there's not enough memory available, when assuming more than 2 factors on each level. I didn't change the between-level factor-structure up to now since it's reasonable to assume the same number of factors as on the within-level. I already changed the number of intergrations down to 10. Is there anything else I can do to estimate this model?
Thank you for comments.
bmuthen posted on Wednesday, July 06, 2005 - 4:38 pm
You can try using integration = montecarlo instead.
Florian, sometimes I try to get a model estimate by setting integration to 7 (or even 5 or 4) and then bumping it up to the highest reasonable level to get final estimates. Also, what you can do is to set your virtual memory up to its highest level (I think around 4 or 6 Gb) and see if that helps (although this option will not help the problem run fast, it may just help it run).