MCFA - variance of between-level late... PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
 Johnson Song posted on Sunday, February 23, 2014 - 8:46 pm
Dear Dr. Muthen,

A multilevel CFA was run to test the property of a scale. The range of the ICCs for the items are between .10 and .20.
However, the variance of the between-level latent variable is not statistically significant, does it mean that it is not necessary to run a MCFA?

BTW, do you have any recommendations about how to compute the ICCs for categorical items?

Thank you very much!

 Bengt O. Muthen posted on Monday, February 24, 2014 - 11:38 am
Answer to 1st paragraph:

If the between-level model fits well, the answer is in principle yes.

Answer to 2nd paragraph:

Use the same formula, where the within-level variance is 1 for probit and pi-square divided by 3 for logit.
 Bengt O. Muthen posted on Monday, February 24, 2014 - 11:38 am
I should add that my answer to 1 assumes that you have enough between-level observations.
 Johnson Song posted on Wednesday, February 26, 2014 - 2:20 pm
Dr. Muthen,

Thank you so much for your reply!
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message