Prohibitions in Conjoint Designs: Embrace the Possibilities, Avoid the Peril!

On-demand Webinar

“Oh #*@%! The client wants to prohibit some levels from appearing together!” You’ve probably heard in your first course or book on conjoint analysis to avoid prohibiting combinations of levels because it damages statistical efficiency. While that’s generally good advice, there are many situations that seem to call for prohibitions so we can avoid showing impossible product feature combinations and show both high- and low-end products at more reasonable prices. In this webinar, we’ll discuss how to check to see if modest prohibitions are very damaging from a statistical standpoint. Next, we’ll cover ways to accommodate more aggressive prohibitions in ways that preserve the statistical integrity of the experiment while at the same time making the survey more realistic and the models more accurate. These approaches for “accommodating” prohibitions in an appropriate way include: collapsing attributes, conditional pricing, summed pricing, and alternative-specific designs.

Webinar Details

Presenter(s)

Bryan Orme

President, Sawtooth Software