forum.smartpls.com

Hi

I have generated a weighting vector based on census using SPSS RAKE WEIGHTS. I then turned the weight on and ran regression models in SPSS. Meanwhile, I also imported the same dataset to SmartPLS and ran the models with the weighting vector.

However, the results from SmartPLS are different from those of SPSS.

Could someone explain why it happens and possible solutions? Thank you.

Stella

Which results are you actually comparing? PLS algorithms routinely use standardized input data. Thus you will get standardized coefficients.