Theory suggests that these latent variables (SEFM, SEFI, SIF, and SER) variables are correlated.
Only the residual (disturbance, error) terms can be correlated for endogenous (dependent) variables in a path/SEM model. The reason is that the covariance of endogenous variables is not a free parameter but a function of other free model parameters. You should be able to include covariances among the residual/error/disturbance (e) terms.
Thank you very much for the response Christian Geiser. I am attaching a picture of a separate model of SEFM, SEFI, SIF, and SER based on theory. In this model, the same latent variables are exogenous variables that are correlated. Is the interpretation of the correlations the same when there is correlation between error terms in the first model?
No, the interpretation is different. Endogenous variable residual correlations are partial correlations, whereas the correlations between exogenous variables are zero-order correlations.
Thank you for the responses so far Christian Geiser ! One last question: If theory have shown that SEFM, SEFI, SIF, and SER are correlated when analysed in a separate model (see model 2 above), should I correlate the attached error terms e62, e63, e64, and e65 in model 1, even if it's not suggested by the modification indices in SPSS AMOS?
- Badges
- Science topic
- Similar topics
- Books