missing data & lavaan

[chris distefano]

Hello,

I am running a dataset with incomplete data (MAR). The data are ordinal, so I am interested in using WLSMVS or ULSMVS.

When I run these methods with 'missing=default'  and  'missing = pairwise', I get different results.  (Not greatly different, but still - different).

This leads me to the following questions:

-What is the difference in the options default & pairwise for these estimators?

-From my understanding, the pairwise option is similar to FIML (just with WLSMVS -using an ordinal estimator)--this is also similar to Mplus, correct?

Thanks for any assistance,

Chris

[Terrence Jorgensen]

When I run these methods with 'missing=default'  and  'missing = pairwise', I get different results.  (Not greatly different, but still - different).

This leads me to the following questions:

-What is the difference in the options default & pairwise for these estimators?

I think the default is listwise deletion, whether data are categorical or continuous.  You can verify this by comparing the sample sizes using the two methods.

-From my understanding, the pairwise option is similar to FIML (just with WLSMVS -using an ordinal estimator)--this is also similar to Mplus, correct?

Only in the sense that no rows of data are thrown away as long as they are observed on at least one (FIML) or two (WLS) variables.

Terry

[christine distefano]

Thanks, Terry,

I guess my question was that the N's are the same (and the full dataset N) but there is a difference if the default or the pairwise are used...  I think they are similar, but just wanted to know what the difference was....

Thanks again! 

Chris

--
You received this message because you are subscribed to a topic in the Google Groups "lavaan" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/lavaan/3BgYv1n5GDU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to lavaan+un...@googlegroups.com.
To post to this group, send email to lav...@googlegroups.com.
Visit this group at https://groups.google.com/group/lavaan.
For more options, visit https://groups.google.com/d/optout.