MaxDiff

MaxDiff (Maximum Difference Scaling) shows respondents a small set of items at a time and asks them to identify the BEST and the WORST option in that set. The exercise repeats across multiple rounds with different item combinations. Analysing which items are consistently chosen as best and consistently avoided as worst produces a statistically robust preference ranking across all items.

MaxDiff is far more accurate than asking respondents to "rate each item on a scale of 1–10" — rating questions compress toward the middle and suffer from scale use bias. MaxDiff forces genuine discrimination.

When to use it

Reach for MaxDiff when:

• you have 8–30 items to prioritise and need accurate relative preference data
• you are doing feature prioritisation, message testing, or benefit ranking
• previous rating surveys have given you "everything scores 4 out of 5" results

How to configure it

Open the question in the builder and expand Settings:

• Items — the full list of items to be evaluated (8–30 recommended).
• Items per set — how many items appear in each round (4–5 is standard).
• Number of rounds — how many sets each respondent evaluates (typically 8–12).
• Best label — the label for the best-choice selection, e.g. "Most important".
• Worst label — the label for the worst-choice selection, e.g. "Least important".

Tips

Tip — SurveyKar generates the rotation design automatically so that each item appears roughly the same number of times and is paired with each other item fairly. You do not need to design this manually.

Tip — MaxDiff requires a larger sample than a simple rating question to produce stable scores — aim for at least 150 respondents for a 15-item study.