Setting Item Count and Items Per Set in a MaxDiff Study

Overview

Size your MaxDiff so every item gets seen enough times to rank reliably, without exhausting respondents. Three settings interact: total Items, Items per set, and Number of sets.

Step-by-step

1. Open the MaxDiff editor — select your MaxDiff question.

2. List your Items — 10–20 items is the comfortable range. MaxDiff handles more than a ranking question can, but the list still needs to be finite.

3. Set Items per set — 4–5 is the standard. Fewer gives too little information per screen; more makes the "most vs least" judgment slow and noisy.

4. Set Number of sets — enough that each item appears at least 3 times across a respondent's sets. That's the rule of thumb for a stable score.

Sizing rule of thumb

Each respondent sees sets × items-per-set item slots. Divide by your item count to get appearances per item — aim for 3+.

Example: 12 items, 4 per set, 12 sets → 48 slots ÷ 12 items = 4 appearances each. Comfortable.

Tips

Tip: If you can't reach 3 appearances per item without a punishing number of sets, your item list is too long. Cut it — or accept that the tail of the ranking will be noisy and only trust the top and bottom.

Note: Every respondent doesn't need to see every item. The design rotates items across people, so the aggregate ranking covers the full list even though each individual sees a subset.

Related articles

• Getting Started with MaxDiff Best-Worst Scaling in Surveti — the basics
• Interpreting Best-Worst Score as a MaxDiff Preference Output — the output