Creating an effective MaxDiff exercise design is essential for gaining valuable insights. Discover includes a design recommender that generates a default design following best practices. The design can be viewed and customized in the Advanced tab of a MaxDiff.
When the Override recommended exercise design toggle is off, the design is auto-generated based on the number of list items.
Even when disabled, the fields show useful information about the design that will be used during data collection:
The designer distributes items across tasks with three goals:
Turn on the Override recommended exercise design toggle to modify the recommended settings. Turning it off reverts to the default recommendation.
This is one of the most important factors in your MaxDiff design.
For typical studies requiring individual-level precision, show each item three or more times per respondent. For smaller samples (N of 50 to 400), showing each item three or more times improves precision when comparing preferences across groups. For larger samples (N of 600 or more) with fewer segment-level comparisons needed, showing each item one to two times is typically sufficient.
This setting works differently depending on the MaxDiff type:
Several variables impact the MaxDiff design and the number of times each item is shown:
MaxDiff requires at least six items. While it can handle hundreds of items, we recommend limiting your list to around 30 or fewer to reduce fatigue and improve precision. Larger sample sizes can accommodate more items since individual-level comparisons become less critical.
For Relevant items MaxDiff, make sure your dynamic list logic results in a reasonable number of items per respondent.
Show three to five items per task for most exercises. More than five can cause fatigue and response errors. Also avoid showing more than half the total items in a single task — for example, no more than four items per task in an eight-item study.
Controls how many total tasks are shown to each respondent. Keep this within a reasonable range — excessive tasks lead to fatigue and response errors.
If you encounter warnings or errors while configuring your MaxDiff exercise, refer to the following explanations for guidance:
This lowers the precision of individual-level utility scores. You may proceed if you have a large sample or don't need high individual-level precision.
This reduces the ability to distinguish respondents' middle-preference items. Avoid this when possible.
More than seven items increases fatigue and error rates. Five or fewer is recommended for optimal data quality.
Each task needs at least three items for respondents to reliably identify best and worst.
Increase the number of items per task and/or number of tasks to ensure each item is directly or indirectly compared to all others. Connectivity is especially important for individual-level analysis like HB estimation. For aggregate methods like aggregate logit or latent class MNL, some loss of connectivity may be acceptable. as connectivity is achieved when pooling responses across participants.
