Spread the Word . . .
-
Tag Archives: Correlated Component Regression
Embracing Multicollinearity in Key Drivers Analysis
Contributed By Robert Mills
One of the most common goals of survey research is to identify those aspects of your brand/product that most strongly relate to overall consumer satisfaction with your brand/product. Sounds simple, right? The fact that there are almost as many opinions about how to perform this analysis as there are people who perform it suggests not. One of the key issues surrounding this type of analysis, often referred to as “key drivers analysis,” is the fact that many of the attributes on which respondents rate your brand are highly interrelated (ormulticollinear) – meaning that a change in one item produces a change in other items.
