Hello, I have built a SEM model in Amos, containing interaction effects between latent construct and a dummy variable. I think I have made everything by the book:
(1) I have centered the latent constructs
(2) I have multiplied the centered constructs with the dummy variable and introduced the result in the model, as the interaction variable.
However, I get very high standard errors (huge, actually), and all the path coefficients are NOT significant (t