CFA on JASP
Hello,
I'm preparing to run a CFA for my Master Thesis, and I would like to know if I can run a first-order CFA with 23 factors on JASP. Is there a maximum?
Thanks!
Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Categories
Supported by
Comments
I don't think there's a set maximum, have you tried setting up the analysis? I don't have data to test but I successfully made the way for 23 factors:
Hi! Thank you for your reply, I don't have the data yet, I was just curious if it was possible, and I didn't know I could try it without a dataset! Thanks, that's great to know that I will be able to do it!
Valentina
There is no simple software maximum, but “the interface accepts 23 factors” is not evidence that the model is estimable or appropriate. The important quantities are the number of indicators per factor, cross-loadings and residual covariances, factor correlations, sample size, missingness, estimator and response distributions.
Before collecting data, write the full measurement model and count the free parameters. Check whether all factors are conceptually distinct and whether each has enough defensible indicators. A 23-factor CFA can be computationally possible yet produce weak identification, correlations near one, Heywood cases or unstable standard errors.
Do not select N from a universal cases-per-item rule. Use plausible loadings, factor correlations, item distributions and the planned estimator in a Monte Carlo simulation. If the model is intended to represent several higher-order domains, test that theory explicitly rather than fitting 23 correlated first-order factors only because the software permits it.