Correl. Weights in combination with refl. endog. cons/Q2

[HansDampf3000]

Hello,

I conducted a PLS analysis with four formative (exog.) constructs and calculated the correlation weights using MODE A.
My endog. construct is reflective.
Are there any complications when using MODE A correlation weights in combination with a reflective endog. construct when using blindfolding procedure Q2?
A few days ago I got smartpls3 and the results differ from the results I got with smartpls2 (Smartpls3: 0,3 and SmartPLS2: CV Red.: 0,27)

Thank your for your answer

Kind Regards

HansDampf

[jmbecker]

Generally, there should be no problem with using correlation weights (Mode A) formative constructs as exogenous predictors in the blindfolding, when your endogenous construct is reflective.

Regarding the differing results between SmartPLS 2 and 3:
The blindfolding procedure in SmartPLS 2 is not always reliable. If you search the forum, you will find information on these issues. After all SmartPLS 2 is not a stable version, but only a BETA Version that has some known problems.
The SmartPLS 3 results should be more reliable and has no known issues with blindfolding.