Setting Up a Max-Diff Experiment
A Max-Diff experiment involves getting respondents to provide rankings. That is, in each set, one item is ranked first, one last and the rest are assigned a middle rank. As such, this data is most readily analyzed in Q using the Question Type of Ranking.
The most straightforward way of setting up a max-diff experiment is by selecting Marketing - Max-Diff - Max-Diff Setup from an Experimental Design in the Online Library.
Alternatively, if your data is not in this standard format, you can instead:
- Create variables according to the instructions in Max-Diff Specifications.
- Manually set these variables to a Question Type of Ranking (see Max-Diff Case Study).
In general, this manual process is more time consuming and is best used only more exotic applications, such as anchored max-diff.
Max-diff can also be analyzed using the Experiment question type. This will lead to identical analyses as those created with the Ranking question type. However, the benefit of the Experiment question type is that it permits the alternatives in the max-diff experiment to be constructed from attributes (this is sometimes referred to as best-worst choice experiments). See Experiments Specifications for details on how to set up experiments.