Survey questionnaire bias check before it goes out
Reviews a draft survey question by question for leading wording, double questions, missing answer options, unbalanced scales and order effects, and returns a corrected version ready to send. Use it before a customer, employee or research survey goes live, when bad questions can no longer be fixed.
The output this prompt produced, unedited.
The prompt
Paste
The draft survey: [paste every question with its answer options]
What the results will be used for: [e.g. decide whether to open on Sundays]
Who will answer: [e.g. existing customers of a dental clinic, mostly over 40]
How it is sent: [e.g. WhatsApp link, email, or a paper form at the counter]
Check every question for
Leading or loaded wording, such as "How much did you enjoy".
Two questions in one, such as "fast and friendly".
Answer options that overlap or leave people out; add "Other" or "Not applicable" where needed.
Scales with more positive than negative points, or labels only at the ends.
Jargon or terms the audience may not know.
Questions that do not serve the stated purpose.
Order effects, where an early question shapes later answers.
Length: flag the survey if it takes more than five minutes.
Return
| No. | Original | Problem | Rewritten question | Answer options |
Questions to remove, with one reason each.
The final survey in sending order.
Two checks to run on the first twenty responses.
Keep to
The survey's own language and tone; do not add new topics.
Neutral wording that a person on either side could accept.
Copy the prompt and paste it into Gemini, then set the aspect ratio to 3:4. No reference photo is needed for this one.
The prompt is written in Markdown with each part under its own heading, which is what the copy action hands over. Change one section at a time, keeping the rest intact, so you can tell which change did the work.