Estimating and testing factor correlations with CFA

preview_player
Показать описание

Рекомендации по теме
Комментарии
Автор

Thanks for the Video. Just wondering if it's correct to use bootstrapping in Amos to get the CI?

ambercy.t
Автор

Hi Mikko. I just finished reading Online Supplement 4. But you were not clear how to conduct a discriminant validity test using your suggested way.
I would appreciate if you upload a video showing exactly what you do with photos not just by words

alexweringo
Автор

Dear Mikko
Can you please help me regarding a problem I have experienced?
To develop a new scale for a construct, EFA suggested two factors solution.
I ran a CFA with a 200 sample size to test convergent and discriminant validities. After removing 4 items, convergent validity seems to be established as all standardised factor loadings are above 0.5 and fit indices are acceptables. The components have good reliability.
However, when running a Chi-square test for discriminant validity, the difference is 3.05 and the CFA says the correlation between the components is 0.91.
If I add 4 error covariances (3 are within-component and 1 between component), the correlation becomes 0.88 and Chi-square difference becomes 6.5.
What do you suggest?
Kind regards

jasonkuavi