How to use the Survey Bias Checker.
Bias checking should explain the problem, not merely label a question bad. Look for assumptions, loaded wording and questions that ask two things at once.
Make the workflow fit your task.
Inspect assumptions, emotional wording, response options and questions that combine different concepts. Explain the likely interpretation problem and propose a neutral alternative that still measures the intended subject.
- What you provide
- Questionnaire and research goal.
- What you get
- Leading, double-barreled and ambiguous questions with fixes.
See the input and the result.
Illustrative input and output · a teaching example, not a live WebAct run
Example input
Review this survey question: How helpful and easy was our excellent support service? Produce separate neutral questions with response options.
Completed example
Issue: excellent suggests a positive answer, while helpfulness and ease are different concepts. Question 1: How helpful was the support you received? Options: not at all, slightly, moderately, very, extremely; not applicable. Question 2: How easy was it to obtain the support you needed? Options: very difficult, difficult, neither, easy, very easy; not applicable.
Load this input into the prompt, then copy it to WebAct to try the task. Your result may differ from the illustration.
Decisions and troubleshooting.
Is a positive adjective always inappropriate in a survey?
It becomes a problem when it suggests the expected answer or presupposes an evaluation. Use neutral wording when measuring that evaluation.
Why does removing one loaded word fail to fix the question?
Check for a remaining assumption or double-barreled structure. Bias can come from what the question requires the respondent to accept.
Try it with your own source.
Replace the example with your material in the task prompt. Keep the requirements you need, then copy the task into WebAct.
Customize and copy the task ↑