Hello Murali,

Any variable in a proposed model should have appropriate justification for its inclusion, regardless of whether it is a latent or manifest variable. So, the issue is really whether you can make an appropriate case for each specific variable.

Good luck with your work.

Thanks David. Can you help me with some reference articles where new variables were added (beyond the ones identified from EFA) while performing SEM? I need to quote them in my work. Thanks again.

@Murali Any variables, while testing the hypothesized model, can be only be added if it makes sense theoretically. I agree with @David that you need to provide justification for the inclusion of that variable be it latent or observed.

Thank Imran and David.