Hallo again,
I have another important question concerning the SmartPLS results for my master thesis.
I searched this forum but could not find a solution for the problem. Therefore, I would be really happy if someone could help me.
I have testet a model with several interaction effects and there is one interaction effect that is hard to interpret for me:
I testet the influence of X and B on Y and the interaction effect of X*B as well.
(there are other paths in the model which i leave aside, as they are not significant anyway...)
My results look like this:
Influence of X on Y: 0,084, Insignificant
Influence of B on Y: 0,067, Insignificant
interaction X*B: -0,253, significant (p-level: 0,056)
My question: Is it ok to interpret the interaction effect, even when the direct effects are not significant?
Another problem: The negative interactioneffect does not make any sense at all, in my model, and therefor confuses me.
I just realized, that some values for cronbachs alpha, the AVE and the outer loadings are not good with respect to the interaction effects. Does anybody know whether i have to check those values and if it does not make any sense to interpret my results, when the values are bad?
I would appreciate if someone could help me ;)
Thanks in advance, Katrin
by the way: I have asked another question and nobody has replied to it so far :( So it would be really cool if someone could help me there as well: viewtopic.php?t=1814