1
$\begingroup$

I am doing a confirmatory fator analysis and I do obtain different results in SPSS AMOS vs. R/lavaan. All parameter estimates, df and standard errors are the same. Yet, standardized residuals and modification indices differ. This makes it hard for me to respecify the model.

In lavaan, I use the "modindices" -command as well as residuals(data, standardized = TRUE).

Can anyone tell me whether there is some major difference in how these indicators are being computed by the different softwares? Does anyone have good literature on which formulas do apply/who I should cite when naming modification indices and standardized covariance residuals as preferred way to decide about model respecification?

I would be really happy about your help!

$\endgroup$

1 Answer 1

1
$\begingroup$

Check out this post. https://www.researchgate.net/post/Mathematical-formula-for-the-Modification-Index-in-AMOS I also have the same question and I searched for some info online. Check the Modification Indicies section in the Lavaan document and you will see they use a different formula to calculate MI, and I believe R and SAS apply the same formula.

$\endgroup$
1
  • 1
    $\begingroup$ Thanks very much, I have been searching for such a long time... Helps a lot! $\endgroup$
    – eventually
    Apr 3, 2021 at 7:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.