Survey Question Auditor
Quality-check a draft survey before launch: leading language, double-barreled items, scale mismatches, bias, order effects, and more. Pairs with Survey Question Generator for write-then-audit workflows.
Prompt
Act as a Principal UX Researcher and expert in survey methodology and questionnaire design.
Please audit the following survey questions for quality issues. After your audit, rewrite flawed questions and provide a corrected version of the full survey.
[Paste your survey questions here - include question text, response options, and any instructions]
For each question, please assess and flag the following dimensions:
- Leading language
- Double-barreled questions (two things at once)
- Loaded or assumptive language
- Ambiguous wording or jargon
- Response scale mismatches (e.g., Likert vs. frequency)
- Acquiescence bias risk ("agree" or "yes" as socially easy)
- Recall burden (memory too specific or too far back)
- Order effects that may bias later questions
For each flagged issue:
- Quote the problematic text
- Name the issue type
- Explain why it is a problem in one sentence
- Provide a rewritten version
At the end, provide:
- An overall quality score (1-10) for the survey
- A prioritized list of the top 3 fixes with the biggest impact on data quality
- A recommendation on whether the survey is ready to deploy or needs another revision pass
Important: If I tell you that certain items are validated scales (e.g., SUS, NPS, UMUX-Lite), flag them as pre-validated and do not rewrite those items - only note any surrounding context issues.
Optional second pass: "Audit again focusing only on response scales and labels."How to use
- 1Paste the full instrument including scales and instructions.
- 2Run a second pass focused only on scales if the first pass was wording-heavy.
- 3Add audience context (e.g., non-native English, older adults) and ask for reading-level checks if needed.
Pro Tips
- • Still run Survey Cognitive Pretest with 5 people before full deployment. An AI audit is not a think-aloud.
- • Not every flag is a mistake; forced-choice items may be intentional.
- • Use the quality score to justify revision time with stakeholders.
