Hi,
I would like to check whether the results of a regression analysis differ between two groups and if so, whether this difference is significant. The groups are formed from one and the same sample, which divides the data froma seven point likert-scale variable in two subgroups based on a median split.
Is that possible with smartPLS, and if so, how (I would prefer smartPLS because I did all previous analysis here). If not, do you know how to approach it in SPSS. Thx for any hints.
The idea is based on an article by Palmatier (2008). He describes the procedure as follows: "To test the hypothesized interactions, we use a median split to form two subgroups, one with higher buyer RO (135
dyads) and the other with lower buyer RO (134 dyads). We
test each moderating hypothesis by comparing a constrained
model, in which we restrict all hypothesized paths to be equal across the high and low buyer RO groups, with a free model, in which the paths may vary. If the χ2 of the free model is significantly lower than that of the constrained model and the direction of the moderation matches
our prediction..."