Design a Bias-Checked Survey Where Every Question Informs a Decision

You get a short questionnaire in a table, with each question tied to a decision, a bias check for leading, double-barreled or loaded wording, and a list of cut questions.

Prompt

Prompt
Act as a survey methodologist who designs short, unbiased questionnaires.

Your job: build a survey for me, check every item for bias, and justify why each question is there.

Before writing anything, ask me these questions in one message and wait for my answers:
1. What decision will this survey help you make, and by when?
2. Who will answer it, and how will you reach them?
3. What is the most time you can ask of them (minutes or number of questions)?
4. Paste any draft questions you already have, or say none.

Rules:
- Tie every question to a specific decision. Cut any question that informs nothing.
- Check each item for leading wording, double-barreled phrasing, loaded terms and unbalanced answer scales. Name the problem and give the fix.
- Use balanced scales, and add a neutral, none or prefer-not-to-say option where it fits. Keep each question under 20 words.
- Order questions from general to specific, with sensitive and demographic items last.
- If I pasted draft questions, quote the original next to your rewrite.

Output format:
1. Survey goal (one sentence)
2. Table with columns: #, Question, Answer options, Decision it informs, Bias check (Pass, or Issue + fix)
3. Questions cut and why
4. Estimated completion time

End with one line on your assumptions and what this survey cannot tell me.

Example output

At a glanceMembership survey at a glanceHollow Oak Books survey for 1,200 newsletter subscribers
Questions in final survey
7
6 closed, 1 optional open
Biased drafts fixed
2
One leading and loaded, one double-barreled
Questions cut
3
Loaded, off-decision or sensitive
Estimated completion time
About 3 minutes
Under the 5-minute limit

Example for: an independent bookstore deciding whether to launch a $60/year membership, surveying 1,200 newsletter subscribers, 5-minute limit, 3 draft questions supplied.

1. Survey goal
Decide by June whether Hollow Oak Books launches a paid membership and which two perks to include.

2. Questionnaire

#QuestionAnswer optionsDecision it informsBias check
1In the last 3 months, how many times did you buy a book from us?0 / 1 / 2-3 / 4-6 / 7+Segment results by purchase frequencyPass
2How do you usually buy from us?In store / Website / Phone order / Haven't bought yetWhether perks must work onlinePass
3How interested are you in an annual membership costing $60?Not at all / Slightly / Moderately / Very / ExtremelyGo or no-go on launchIssue: original "Wouldn't a membership be a great way to support local books?" was leading and loaded. Fix: neutral wording with the price stated.
4Which perks would matter most to you? Choose up to 2.10% off books / Free shipping / Early event tickets / Monthly staff pick / Members-only evening / NoneWhich 2 perks to includePass. "None" added so uninterested readers aren't forced to pick.
5How often would you attend an evening event at the store?Never / Once a year / Every few months / Monthly / More than monthlySize of events calendarIssue: original "How often would you attend and recommend our events?" was double-barreled. Fix: split, recommendation part dropped.
6What, if anything, would stop you from joining?Open text, optionalPrice or perk changesPass
7Which age group are you in?Under 25 / 25-39 / 40-59 / 60+ / Prefer not to sayEvent timing and targetingPass. Placed last.

3. Questions cut and why

  • "How satisfied are you with our excellent selection?" Loaded ("excellent"), and no membership decision depends on it.
  • "Would you recommend our events?" Removed when splitting Q5; it does not change the launch or perk decision.
  • Household income. Sensitive, and Q3 already tests willingness to pay at the real price.

4. Estimated completion time
About 3 minutes: 6 closed questions and 1 optional open question.

Assumptions: newsletter subscribers stand in for all customers, and stated interest usually overstates real sign-ups, so treat "Very" and "Extremely" as an upper bound and confirm with a small pre-sale.

Written by AI from this prompt. Your results will vary.

Tips and common edits

Why this works

  • Asking for the decision first means every question gets a job, and filler questions are cut instead of padding the survey.
  • Naming the four bias types gives the model a checklist, so problems are flagged by type with a concrete fix.
  • Quoting originals beside rewrites shows exactly what changed and teaches you to spot bias in future drafts.

Use before sending any customer, member or staff survey. Have your decision, audience and draft questions ready. Works in ChatGPT, Claude or Gemini. Paste drafts to get side-by-side rewrites.

Common edits

  • Add a column for which chart or metric each answer will feed in your report.
  • Ask for a version formatted for Google Forms or Typeform with question types labelled.

Make it repeatable

Keep using this prompt without pasting it in every time.

Prompt details

By
Prompts-GPT editors
Published
How prompts are written and checked

Did this prompt work?

Keep exploring

More in Research