Four ways to edit the same questionnaire.
| Editor | When it helps |
|---|---|
| Text | Bring a written question list and turn it into structured questions. |
| Script | Define questions and options with a structured survey script. |
| Visual | Add and arrange questions and sections in a visual workspace. |
| Flow | Inspect conditions and paths to understand who sees each question. |
You can move between views while working on the same survey. AI-assisted drafting gives you another starting point; it does not replace review in the editor.
Make the path relevant to the participant.
Show a follow-up only when someone has the relevant experience. A person who did not use a support service should not be asked to rate it. Conditional logic helps distinguish that skipped question from a question that was shown and left unanswered.
Moving between stages → The same condition logic applies between stages in a project.
Use earlier answers in later question text when the study calls for it. For example, reciprocal questions about two communities can use the appropriate group name without requiring separate questionnaires. You can also vary question order where the research design requires it.
Keep a record of the survey that went into the field.
A published study keeps its questions and logic fixed to the deployed version. Editing a template afterwards does not silently change what participants in that study are answering. The response data records the study version.
If the research needs a new version, plan how you will distinguish or compare the results. A stable question code helps you recognise the same variable when you export data or repeat a measurement.
Test more than the wording.
- Walk through every branch, including refusal and ineligible paths.
- Check mobile screens, required questions and long answer options.
- Review each language version and its participant information.
- Confirm the variable names, codes and intended analysis before publication.