Is missing data "automatically" dealt with in mirt?

[Elia Emmers]

Hello Phil (or any other mirt expert who can help me).

I have to estimate a relatively easy Exploratory Factor Analysis (EFA) with ordered categorical data, but there is some missing data in it. I'm used to the psych packages for EFAs but it does not provide options for missing data and lavaan's FIML estimator assumes continuous data. semTools only provides multiple imputation support for CFAs.

If I remember correctly (and please tell me if my understanding is wrong), the maximum likelihood estimators are "full information" types when doing IRT. Would it be OK to just simply estimate my model as an IRT model in mirt knowing that the estimator will use all available data and hence "take care" of the missingness? Or is there something else I'd need to do or specify?

[Phil Chalmers]

Correct, IRT/MML estimation used in mirt is full information. So missing data causes no immediate problem (so long as the missing data mechanism is MAR or MCAR). That's one particularly nice feature of MML estimation compared to the limited information variants in the SEM approach for ordinal data, and makes for some interesting multiple-group planned missingness designs for linking and equating. Cheers.

Phil

--
You received this message because you are subscribed to the Google Groups "mirt-package" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mirt-package+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.