Research TV : Max-Diff Scaling

May 14, 2010 By 5 Comments

Research TV : Max-Diff Scaling

Dr. Nico Peruzzi (@NicoPeruzziPhD) from Outsource Research Consulting has put together a very informative video on Max-Diff Scaling. Max-Diff Scaling is a simple (yet relatively powerful) technique for users to determine what is “truly” important, and for that matter what is NOT important for a user.

For too long, we’ve relied on 5Pt likert Scales – and this has to change. 5Pt Likert scales do not provide data that is actionable – Just ask NetFlix – they spend a million dollars and the fundamental problem is that 5pt rating scales do not “discriminate” enough.

Learn more about Max-Diff Scaling by watching this video from Nico.

Additional Links:
Filed Under: News, Video Tagged With: , , ,
About Vivek Bhaskaran

Vivek Bhaskaran is the President and CEO of Survey Analytics.

  • Pingback: Max-Diff Scaling is a Great New Way to Really Measure Importance « SurveyAnalytics Blog

  • http://www.intellitics.com Tim Bonnemann

    Nice intro, thanks.

    For an online group brainstorming tool we’re developing, we’re currently looking at pairwise comparison as a method for ranking a (potentially long) list of items.

    With MassDiff, is there a standard system for “counting” the respondents’ answers? How do the answers get translated into the MaxDiff score?

    Thanks!

  • Pingback: Ditch Your Likert Scale for the Max-Diff « QuestionPro Blog

  • http://www.orconsulting.com Nico Peruzzi

    Hi Tim,

    I’ve got the perfect technical paper for you to check out to answer this question. It’s called “MaxDiff Analysis: Simple Counting, Individual-Level Logit, and HB (2009)”. I am personally a fan of using Hierarchical Bayes (HB) to get the scores. You can access the paper here:
    http://www.sawtoothsoftware.com/education/techpap.shtml.

    Here’s a summary:

    “This paper compares different methods of obtaining individual-level scores for MaxDiff surveys at the individual level: Simple counting, individual-level logit, and HB. Key to the success for all these methods was having enough information available for each respondent to estimate stable scores.

    The author (Orme) finds that counting analysis provides reasonable population estimates of scores, but that the individual-level scores can lack precision. Precision is better under the logit model estimation methods: either individual-level logit or HB, which “borrows” information across the sample to improve the individual-level logit scores for individuals.

    Despite the simplicity of the counting approach and its weaknesses, it tends to do quite well in predicting responses to holdout choices. But, across-respondent variance (heterogeneity) tends to be weaker than the other methods studied.”

    Good luck!

  • http://www.intellitics.com Tim Bonnemann

    That’s great, thanks!