Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Is there a way to use the "relative importance" as the dependent variable?

Good afternoon.  I thank you for your feedback and advice.

I plan on using CVA to find the most preferred product bundle along with the relative importance of each attribute.

At the same time, I will measure the respondents' attitudes and behaviors using 7 point likert scale.

I intend to find whether such attitudes and behaviors would correlate to the relative importance respondents put on each attribute.

Would it be possible to examine such influence by:

grouping the people with higher and lower average attitude scores (using Convergent Cluster & Ensemble Analysis) and see whether their relative importance would be significantly different?

Or, is there a better or simpler way to measure the influence of attitude and behavior scores on the utility scores of each attribute?

Thank you.
asked Feb 19 by anonymous

1 Answer

0 votes
I would start more simply and then progress to more complex methods later.

As both the importances and your Likert scale items are arguably metric variables, I would probably run correlations between the Likert items and the importances, to see whether you had significant correlations.  If so, then you might want to look at fancier analyses.
answered Feb 19 by Keith Chrzan Platinum Sawtooth Software, Inc. (95,775 points)