Hi all,

I am encountering the following issue: I want to compare two groups with each other in using smartPLS since I just have 32 observations per group and my model is really complex (see the attachement). I compare Vinyl (treatment) to CD (control group) and I intend to ascertain whether a causal effect is significantly stronger or weaker for vinyl as opposed to CD. The question is whether I can use a categorical variable as the first exogenous variable and use dummy coding (1= Vinyl, 0= CD for the reference category) in order to compare the causal effects with each other. Since PLS uses standardized coefficients I am uncertain if it is the right approach.

Alternatively I was thinking about omitting the categorical variable and test the first exogenous constructs via t-testing and then using the PLS procedure for the other causal relationships of each model seperately. Which means I want to set up a seperate model for vinyl (N=32) and CD (N=32) and compare the significant path coefficients with each other. Is it a valid approach?

Thanks for your help!