Dear all,
I am facing several problems with regards to the analysis of a multiplicative model.
I am testing a set of empirical data based on the Expectancy (also called the VIE) model. The model suggests that Motivation, M is a combined additive-multiplicative function of V, I, and E, such that:
M = V + I + E + VI + VE + IE + VIE
This means that I need to account for the interaction effect of V*I*E.
According to Chin et al (2003) Information Systems Research, I should first standardize all the indicators for V, I, and E. Then take the indicators for each LV and multiply them.
There are two problems:
1. Multicollinearity: Standardization of indicators eliminate (or hide??) this problem when there are only two LVs involved. Once I, say, multiply E to (V*I), the VIF would go beyond the suggested limit (i.e. > 5).
2. Performing the interaction effect of V*I*E on SmartPLS. I do not know how to do it!!!
Can anyone help?
Thanks much!
Cheers,
Awie