Importance Performance Matrix Using Max Diff Scores

Question

I am running a max diff study and following up with a satisfaction rating on those same items.  Our client is interested in looking at an Importance- Performance matrix to be able to prioritize those highly important attributes that are currently not performing well.   I wondered if it is suitable to use the max diff scores for the importance component.  Is this done?  Any issues to consider or another analysis that would be better suited for max diff scores?

Answer 1

If the labels on your MaxDiff question are "Most Important" and "Least Important", then I think MaxDiff would be an excellent way to achieve importance scores for items.  Using a standard rating scale would lead to lower discrimination and lower precision as well.

Answer 2

This is an excellent idea.  The halo effect is particularly severe for customer sat reatings, which adds a lot of noise to importances you might "derive" via regression or similar analyses.  

In addition to the benefits Bryan cites for MaxDiff over importance ratings, another important one is that MaxDiff utilities have greater predictive validity than importance ratings.

Comments

Great - thank you both for the quick response!