MaxDiff: How to Measure Importance

What Is MaxDiff?

MaxDiff (Maximum Difference Scaling) is an approach for obtaining preference/importance scores for multiple items (brand preferences, brand images, product features, messages, advertising claims, etc.). MaxDiff is also known as “best-worst scaling”.

When To Use MaxDiff

If you’ve ever been asked to measure preferences for things such as brands, product features, messages, benefits, packages, images, etc, MaxDiff is for you.

MaxDiff estimates the relative importance or preference for the items.

Why not use standard rating scales? Because research has shown that MaxDiff scores demonstrate greater discrimination among items and between respondents on the items. How many times have you tried to interpret the difference between a mean of 4.2 and 4.4 on a 5-point rating scale? You know the futility of that exercise.

MaxDiff scores are easy to interpret. They are placed on a 0 to 100 common scale and sum to 100. Thus, when you see a “10” it has twice as much value as a “5” – something you can’t do with rating scale results.

Why Not Use Conjoint?

MaxDiff is related to conjoint, however, conjoint is more appropriate when you want to estimate how multiple attributes taken together affect overall preference. MaxDiff looks at a lot of “items” (levels) within a single attribute (variable) (e.g., product features), whereas conjoint looks at multiple attributes, each with some number of levels.

Features Of MaxDiff

• Typically 8 to 40 or more items can be examined at once. Advanced methods allow examination of over 100 items.

• Items are presented 4 or 5 at a time (see graphic below)

• The algorithm creates a design that avoids bias in presentation

• Respondents typically see a dozen or so sets

• The total exercise can take as little as 2 or 3 minutes to complete

• Individual level data is available for analysis

• The task is very simple and reliable for a wide range of ages, cultures and education levels. Different interpretations and biases associated with rating scales do not exist with MaxDiff.

Respondents see a number of screens like the one below. Each time, they simply choose the most important (or preferred) and least important item. By doing this task a number of times, extremely valid data is obtained that help you differentiate the value of all items.

Conclusion

MaxDiff is the best ways to measure preference or importance of a list of items – from product features to benefit messages, images, brands, etc. For more information about how MaxDiff might help you, contact us at contactus@elucidatenow.com, or check out our video on MaxDiff in the Resources section.