Hello together,
I am currently writing my Diploma thesis on online brand management and right now evaluating my empirical part. Among other I want to compare the impact of three different clusters, each consisting of a couple of brands, on brand management goals. Therefore I have made an online survey asking participants to answer a 25 item survey on different aspects of a specific brand website, which was randomly assigned to the participant .
As I have understood so far, one possibility of comparing the clusters is to introduce dummy variables to the model and check the significance and change in R² (total effect) in order to see if there are any significant differences in the effect on the relevant LVs by the three clusters.
What I do not understand is how the path-significance of the two dummy variables will give me any information on differences in their impact. In my opinion all you do is compare binary colums of ones and zeros. How can one judge by this the impact of the different characteristics of each cluster on the LV and the whole model?
Is there any other method to compare groups within smartPLS and is there any good literature on it?
Thanks in advance for your help!!!
Cheers,
Ben