forum.smartpls.com

One problem could be the datasettings: is your decimal separator ("." or ",") correct? Etc.

How large is your data file? Are all rows imported?

Honestly, you should think of good reasons that justifies the use of PLS. Independent of PLS, a reviewer expects a well informed choice of the method and some good reasons why one chooses the one over the other. It all depends on your model and the context of your research. Is your model quite compl...

Generally, Cronbach’s Alpha is the lower bound to reliability, while composite reliability is the upper bound. Both are not the best estimates. The first is underestimating, the second overestimating reliability. In addition, Cronbach’s Alpha depends heavily on the number of indicators. With few ind...

With the release of the new SmartPLS 3 we changed the registration procedures. The new SmartPLS 3 release can be downloaded from http://www.smartpls.com/ and there are different licensing options available. If you want to get a new activation code for the old SmartPLS 2 software, please register aga...

Yes, if your model is as simple as you describe: 1 IV, 1MV, and 1 DV, you can use the indirect and total effect from the SmartPLS output and do not have to calculate the indirect effect yourself. In fact, calculating the indirect effect yourself, should give you the same result as from the outputs o...

It is current state of the art to assess the indirect effect (from IV to DV via MV) using the bootstrapping for significance. The framework of Zhao et al. (2010) is quite nice for interpretation. Thus, you can (and should) assess the mediating role of the MV for each IV separately. Each IV will have...

It always depends on the content domain of the predictor. You should be aware that investment sometimes also means decreasing a certain variables/construct. For example, price has a negative relation to many outcome variables (e.g., Satisfaction, Purchase Intention, etc.). If you increase price, the...

It is the usual way that you get one R² for each inner model regression. As it is common for multiple regression. It is the explained variance when considering all independent (exogenous) variables together. You can calculate the R² contribution of each construct by omitting the construct and calcul...

You must differentiate between two different ways to calculate the f2 effect size. One is to delete the construct from your PLS path model. This can also change the outer weight and leads to different latent variables scores and correlations between the latent variables. In this setup, it is not unu...

Actually, you should not use both. Current state of the art is to assess the bootstrapped indirect effect for significance. The framework of Zhao et al. (2010) is quite nice for interpretation. Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in sim...

Because you are analysing different things. In SPSS you analyse the item mean and standard deviation. In the SmartPLS output that you provide you analyse the bootstrapp distribution of the outer loading of your items on the latent variable (i.e., the bivariate correlation of the item and latent vari...

If you also expect heterogeneity in the formative measures, then you should use PLS-POS. The recommendation regarding the necessity of an underlying theory, however, is not limited to PLS-POS, but applies to all segmentation algorithms. The recommendation is meant as a warning not to use segmentatio...

I would usually not use the VAF as a criterion. It is outdated. It can only be used if you have complementary mediation (i.e., the indirect and direct effect have the same sign). However, there are also many reasonable scenarios where competitive mediation can occur (i.e., indirect and direct effect...

Reliability and validity assessment of interaction term constructs are meaningless. The cross-product indicators do not stem from one specific conceptual domain and are just auxiliary measurement to facilitate the interaction term generation. Hence, don't care about low AVE, Composite Reliability or...

If you want to do a two-step approach, you can use the latent variable scores from the first step of a normal PLS and then use PLSc only for the second step. That should work quite well.

However, I don't understand why PLSc should be advantageous when you have multicollinearity in your model.