Do you recommend employing index values (ind. 1* loading 1+....ind.5 *loading 5, ...) in the general model in order to avoid changing factor loadings in subsequent calculations?
Can/should one still employ index values when analyzing interaction effects with Chin/Marcolin/Newsted´s methodology, as "their" interaction variable is based on the multiplied standardized indicator values of the exogenous and moderator variable, not on index values?!
I appreciate your responses.
Regards,
Patrick