Different cross loadings in SmartPLS and SPSS

pmartin · Post by **pmartin** » Mon Jul 23, 2007 1:47 am

I am a first time user of SmartPLS. I started with purifying my measurement model with SPSS and then used them as input to SmartPLS. When I run my model I noticed SmartPLS generates a completely different set of factor loadings for each construct. I traced my data, it is exactly the same as I used in SPSS. I do not have any missing values. Even one has a negative loading (0.93 in SPSS vs. -0.54 in SmartPLS)

To cross-check I calculated cross loadings (using LV scores from SmartPLS) with the excel template posted by Dr. Behrens a couple years ago: it generated the same cross loadings as SPSS.

I have checked previous posts on this problem and it was suggested that it is a user induced problem and these data should be the same. I seem to eliminate user-related errors. Does anyone have an idea on what could be the problem?

cringle · Post by **cringle** » Sun Sep 16, 2007 12:08 pm

Hi,

please be more specific what you are computing in SPSS and comparing to SmartPLS outcomes. Why should SPSS factor loadings (please explain which method you apply) similar to PLS path modeling results.

Best
Christian