{"id":"pm-survey-questions-neutral","audience":"product-managers","category":"research","track":"research","technique":"constraints","title":"Rewrite biased survey questions so they're neutral","free":false,"task":"Your draft customer survey has leading and double-barreled questions. Rewrite them as exactly 4 numbered questions, each asking about one thing only, worded neutrally, and ending with its answer scale in parentheses, plain text, under 90 words. Drop the question the research lead cut.","materialLabel":"Draft survey and comments (attached to your prompt automatically)","material":"Draft survey - Fieldnote mobile app users (sent after 30 days)\n\n1. How much do you love our new, faster note editor?\n2. How satisfied are you with search and tagging?\n3. Don't you agree that offline mode would make Fieldnote better?\n4. How easy or hard was it to set up your first project and invite your team?\n5. How likely are you to recommend Fieldnote to a friend or colleague?\n6. Would you pay more for premium features?\n\nComments from Joy (research lead):\n- Q1 assumes they love it and that it's faster. Make it neutral.\n- Q2 and Q4 each ask about two things. Split them? No: we have room for only 4 questions total. Keep Q2 about search only, and Q4 about setting up the first project only.\n- Q3 is leading. Ask how often they're without internet when using the app.\n- Q5 is our standard recommend question. Drop it from this survey; it's in the quarterly NPS email.\n- Q6: drop, pricing research is separate.\n- Scales: use 1-5 for satisfaction/ease (\"1 = very dissatisfied, 5 = very satisfied\" style) and Never/Sometimes/Often for frequency.","rules":[{"type":"numberedItems","count":4},{"type":"maxWords","max":90},{"type":"plainText"}],"mistakes":[{"title":"Asking to 'improve the survey'","why":"The AI rewrites all 6 questions with headings and design tips.","fix":"Be exact: \"Exactly 4 numbered questions. Plain text.\"","checks":["numberedItems","plainText"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting. Ask for a total word limit and it often lands over it.","fix":"Cap each question instead: \"each question under 20 words including the scale.\"","checks":["maxWords"]},{"title":"Keeping double-barreled questions","why":"The AI rewords but still asks about search and tagging together.","fix":"Add: \"One topic per question: search only; first project setup only.\"","checks":["r0","r1"]},{"title":"Leading wording","why":"AI softens 'love' to 'enjoy', which still assumes a positive answer.","fix":"Add: \"Neutral wording, like 'How satisfied or dissatisfied are you with...'.\"","checks":["r1"]},{"title":"Forgetting the scale","why":"Without a scale, every respondent answers differently.","fix":"Add: \"End each question with its scale in parentheses.\"","checks":["r2"]}],"promptHints":[{"label":"Asks for neutral wording","patterns":["neutral","leading","unbiased","don'?t assume"]},{"label":"One topic per question","patterns":["one (thing|topic)","double.barrel","split"]},{"label":"Adds answer scales","patterns":["scale","1.?5","never.{0,10}sometimes"]},{"label":"Sets the number of questions","patterns":["\\b(4|four)\\b.{0,15}question"]},{"label":"Sets a word limit","patterns":["\\d+\\s*words","word (limit|count)"]}],"hint":"Biased questions get biased answers. Tell the AI to keep one topic per question, use neutral wording, add the scale, and follow the research lead's cuts.","rubricLabels":["The 4 questions cover the note editor (neutral, not assuming love or speed), search only, how often they lack internet while using the app, and setting up the first project only","No question is leading or double-barreled (no 'love', 'faster', 'don't you agree', no 'and' joining two topics like search and tagging)","Each question ends with an answer scale in parentheses (1-5 for satisfaction or ease, Never/Sometimes/Often for frequency), and the recommend and pay-more questions are gone, with no intro"]}