A survey is useful when the research question can be answered through standardised reports of people’s characteristics, experiences, behaviours or views. It is not automatically the quickest method, and a long list of questions does not guarantee useful data.
Research involving people may require ethics approval, consent, secure storage and a clear data-retention plan. Follow the university process before recruitment or data collection; an online form does not make a project exempt.
Design the questionnaire and the analysis together. A question that looks clear on screen can still produce ambiguous, biased or unusable responses when its audience, scale or coding has not been tested.
Confirm that a survey fits the question
Decide whether standardised responses can address the objective. If the study needs detailed accounts, observation or document analysis, another method or a mixed design may be more suitable.
Define the target population
State who should be represented and who can realistically be reached. Distinguish the target population, sampling frame and final respondents. Do not describe a convenience sample as representative without evidence.
Plan recruitment ethically
Use an approved invitation, avoid undue pressure and make participation genuinely voluntary. Consider power relationships when a teacher, manager or clinician recruits people they supervise or serve.
Map objectives to variables
For each research objective, list the information needed and how it could be measured. Remove questions with no clear purpose. Collecting personal details “just in case” creates privacy risk without improving the study.
Choose suitable question types
Multiple choice, rating scales, rankings and open responses serve different purposes. Select a format that participants can answer accurately and that matches the planned analysis. Provide an appropriate option when listed responses may not fit everyone.
Write one idea per question
Avoid double-barrelled wording such as asking whether a service is fast and affordable in one item. Keep language concrete, neutral and appropriate for the audience. Define time periods rather than relying on words such as often or recently.
Avoid leading assumptions
Do not imply a preferred answer or assume an experience occurred. Sensitive questions need particular care, justification and positioning. Explain why information is collected when that supports informed participation.
Design response scales carefully
Use balanced, clearly ordered options and keep scale direction consistent. Ensure points are meaningful and do not overlap. Decide how “not applicable”, “prefer not to answer” and missing responses will be treated.
Order the questionnaire logically
Begin with accessible, relevant questions and group related items. Use routing only when it has been tested. Place sensitive or detailed demographic questions where they will not create an unnecessary barrier to participation.
Keep the survey as short as the objective permits
Estimate completion time using real pilot participants, not the writer’s speed. Remove duplication and low-value items. Burden can reduce completion and encourage rushed responses.
Plan analysis before launch
Create a table linking each item to the research question, coding and intended analysis. Check that response options will produce usable distinctions. Do not collect open text if there is no plan or permission to analyse and protect it.
Pilot with the target audience
Ask people similar to the intended participants to complete the survey and explain confusing points. Test the same device and mode planned for the study. Record changes and recheck routing after every revision.
Prepare consent and privacy information
State purpose, participation, risks, withdrawal limits, storage, access and contact routes as required. Do not promise anonymity if the platform, identifiers or small sample could allow responses to be linked to individuals.
Monitor without changing the study silently
During collection, watch technical completion patterns and approved recruitment targets. Material changes to wording, eligibility or consent may require review. Preserve versions so the final report can explain what participants actually received.
Plan for accessibility and language
Check that the form works with keyboard navigation, screen readers and the devices likely to be used. Plain language can improve comprehension without making a research concept inaccurate. Translation or multilingual administration needs an approved process, documented adaptation and suitable testing; automatic translation alone may change the meaning of questions or response scales.
Report response and missingness honestly
Record how many people were invited where known, how many began, how many completed and how incomplete responses were handled. A response rate needs a defined denominator. Do not assume non-respondents resemble respondents, and avoid presenting a large number of clicks as a representative sample when eligibility or duplicate participation was not established.
Freeze the approved instrument
Give the launched questionnaire a version number and save the exact participant information, consent wording, questions, routing and response options. If a technical fault requires a change, document when it occurred, which responses were affected and whether approval is needed. Analysis cannot treat materially different instruments as identical without explanation.
Final checklist
- A survey is justified by the research objective and target population.
- Every item has a clear analytical purpose.
- Wording, response options, routing and completion time were piloted.
- Consent, privacy, storage and ethics requirements are satisfied.
- The analysis plan exists before responses are collected.
Students can review study programmes on Univs, compare listed institutes and learn how Univs works. Survey collection must wait until the university’s ethics and data-protection requirements have been confirmed.
Editorial note
Requirements can change and may differ by institution, programme and applicant. Recheck current university and government guidance before paying or travelling.
