The papers winning “best presentation” award at the last two Sawtooth Software conferences (Steve Cohen 2003, Keith Chrzan 2004) had one thing in common: they utilized a relatively new scaling technique for multiple items called MaxDiff (also known as Best-Worst Scaling).
We are pleased to announce that we are working very hard on a new software system called the MaxDiff System. It will be an integrated component within the SSI Web system. Users may conduct MaxDiff studies over the Web or via CAPI (computers not connected to the web) using this system. For paper-and-pencil implementation, we have been offering an experimental design program that includes a tool that prepares the data for estimation using CBC/HB software.
MaxDiff scaling is a trade-off method for measuring the importance or preference for multiple items, such as brands, product features, political platforms, advertising claims, etc. Any time you are considering using a rating scale, ranking scale, or constant sum scale for multiple items, you can consider using MaxDiff.
Here’s an example MaxDiff question, involving importance of server features:
Many researchers also have favored the Method of Paired Comparisons (MPC) for these kinds of problems. With MPC, we present respondents two items at a time and ask which is more important/preferred. The new MaxDiff Software system can also be used for conducting MPC experiments.
Even though we haven’t offered a complete solution for MaxDiff studies yet, many of our users have been applying the technique using a combination of separate tools: our Best/Worst Experimental Designer (for generating questionnaire plans), and the Latent Class Module or CBC/HB Module for estimating item scores. The new MaxDiff system will integrate design of experiments, questionnaire programming, and quick estimation using hierarchical Bayes (HB) analysis all within SSI Web’s menu system.
We expect to complete this software in the fourth quarter.