CFA Lavaan module - how does it handle ordinal data?

Discuss the jamovi platform, possible improvements, etc.

by scroon » Wed Sep 09, 2020 1:17 am

Hi all

I love, love, love the CFA module in Jamovi, it's great for teaching. I see that it is using the Lavaan r code.

Does anybody know what correlation matrix it uses if you feed it ordinal vs continuous data? It will make a difference to fit and will determine if people need to go into mplus or not....

thanks!
scroon
 
Posts: 17
Joined: Tue Feb 27, 2018 3:52 am

by seol » Wed Sep 09, 2020 2:47 am

If your data is dichotomous data, Tetrachoric correlation matrix should be used as a input data matrix. Tetrachoric correlation matrix can be calculated using "seolmatrix" module, which is available in the jamovi library.
seol
 
Posts: 39
Joined: Mon Dec 04, 2017 8:40 am

by scroon » Wed Sep 09, 2020 11:38 pm

thank you Seol!

On inspection of the jamovi output, it appears that regardless of declaring variables as ordinal in jamovi the data is treated as continuous for analysis in the lavaan call. (so it looks like my question is solved)
scroon
 
Posts: 17
Joined: Tue Feb 27, 2018 3:52 am


Return to General