Multigroup Analysis with SmartPLS 3 - parametric?

[Anna_Mueller]

Hello everybody,

I run a multigroup analysis and for using the parametric test, I have to fulfill the assumption of normally distributed variables.
But, how do I check this?
I know I can use SPSS to calculate the K-S, but I have 2 questions:

1. Do I calculate K-S for the whole sample (Group 1 + 2 that I compare)? Or do I have to split the data (Group 1 vs. Group 2)?
2. As I use multi-item constructs for my latent variables: Do I use all items to calculate K-S? In my opinion it makes sense to use the factor scores for the latent variables as I compare the path coefficient of the constructs (latent variables).