Design and validate original questions for an answer-ranking or survey-style game.
A strong survey-game question is immediately understood by the intended audience and broad enough to produce several plausible answers, but specific enough to code consistently. The quality comes from real responses, transparent grouping, safe topics, and a pilot—not from guessing what “most people” would say.
Define the audience and play context
State who answers the source survey, who will play, language, region, age range where appropriate, tone, team format, and whether the game rewards predicting a group or recalling facts. Avoid claiming general public opinion from a convenience sample.
Write a scoreable prompt
- One concrete situation and one requested kind of answer
- Neutral wording without a preferred or shaming response
- Vocabulary and cultural references familiar to the intended group
- Enough breadth for several answers but not unlimited interpretation
- No unnecessary health, identity, trauma, finance, sex, politics, or private disclosure
- Documented answer-coding, other category, ties, and challenge rule
Choose a useful prompt structure
| Structure | Example shape | Watch for |
|---|---|---|
| Common association | Name something people keep by a door | Multiple valid synonyms |
| Situational action | What do people do when a call drops? | Order and context differences |
| Preference with boundary | Pick a snack for a short road trip | Diet and region assumptions |
Develop a question set
- Draft more prompts than needed.
- Bias-screen and clarity-review them.
- Collect independent responses.
- Code synonyms using written rules.
- Pilot timing, distribution, and ambiguity.
- Revise, retire, and version the set.
Avoid fake survey authority
- Inventing percentages from intuition
- Changing answer groups after seeing a player response
- Using a tiny friend group as universal evidence
- Writing questions with only one obvious answer
Maintain a question record
For each question store exact wording, audience and collection method, response count, date, raw answers, coding rules, grouped answers, counts, ties, exclusions, pilot notes, version, and facilitator clarification. That record supports consistent scoring and better revisions.