Significant Path - no R2 change: Interpretation?

[redlake]

Hi all,

I have a question regarding interpretation of PLS results: I have a model with one construct explaining 68% of variance in another construct. This is rather good. When I add more latent constructs, these at times show significant indirect paths to the dependent construct via the major explanatory construct, bootstrapping often results in significant total effects, but the r2 does not change at all. How can I interpret such a construct? Am I right assuming that this variable has no effect on the DV over and about the main explanatory variable? Furthermore, if I try a number of different constructs, with the same results again, does that mean that the "explainable variance" in the DV is somehow "exhausted"? Or could this hint at problems in the distribution of the DV?

This may sound stupid questions. Above all, I would be interested in all kinds of practical experiences of similar situations.

thanks

matthias rothensee

[viswadatta]

when you write construct equations you never use indirect influences. RSq value only uses the immediate construct influences. The beta values are used to find out the indirect influences of each cnstruct on another construct.