Define the population before writing the screener
A screener decides whether someone can enter a study. It should apply a rule that exists before the answers arrive. Start with a sentence that defines the population: “Adults in the United States who bought a home appliance for their household in the past six months.” Then list what each phrase means, how it will be measured and which answer will end the interview.
Separate genuine eligibility from the characteristics you would merely like to balance. A respondent's region might be a quota variable rather than a reason to exclude them. If the study is about all recent buyers, screening out people who purchased online because they are harder to recruit changes the population. Record that decision explicitly if the scope really is store buyers.
Document the recruitment source too. A customer list, an opt-in panel and an open web link reach different people. AAPOR's disclosure standards call for researchers to state the population, recruitment method, eligibility requirements and quotas. A well-written screener cannot repair a recruitment source that misses part of the population.
Ask only what is needed to decide eligibility
Put the least sensitive qualifying questions first. Ask about actual behavior or role rather than a conclusion the respondent may not share. For a software buyer study, “Which of these tasks have you done for your organization in the past 12 months?” is more concrete than “Are you a decision-maker?” Define the tasks that qualify in the study plan, not in a hint shown to respondents.
Use a specific recall period and exhaustive, non-overlapping answer choices. “In the past six months, which of the following have you personally purchased?” needs a clear “None of these” choice and a rule for people who cannot remember. If several products qualify, decide whether the question accepts one answer or multiple answers. Pew Research Center's questionnaire guidance explains why wording, response options and question order can change what a survey measures.
Avoid collecting a full birth date when an age band suffices. Avoid asking for a name or employer when neither is needed for eligibility or a clearly stated panel process. If an external panel already supplies an approved variable, verify its definition and age before reusing it; do not silently substitute stale panel data for a screener answer.
Keep qualifying answers out of the invitation
An invitation can describe the broad topic, expected time and who is conducting the study. It should not teach people which answer lets them in. “Seeking people who bought Brand X last week” may be necessary for some recruitments, but “Select Brand X to qualify” is not a neutral invitation. Use plain language about the intended audience without revealing unnecessary routing thresholds.
Do not turn every screener into a trap. Honest respondents should be able to answer “Not sure” or “Prefer not to say” when appropriate. Decide in advance whether those responses are eligible, ineligible or require a manual review. An ambiguous answer should not quietly be treated as a “no” in one report and as missing in another.
For sensitive topics, explain why a question is asked and how answers will be used. If an identified list is used to manage invitations, tell respondents whether their survey responses are linked to their identity. The participation guide covers invitations that are clear without pressuring people.
Build an explicit routing and quota map
Make a small decision table before implementing logic. Each row should show the question, qualifying response, nonqualifying response, missing-response treatment and destination. For example, a purchase question may send “Yes, for my household” to the main survey, “No” to a polite screen-out page, and “Not sure” to a short clarification question. The logic should have one defined outcome for every possible answer combination.
Keep quota closures separate from true ineligibility. A person can meet all criteria while their region is already full. Label that as “eligible, quota closed” in internal records. This distinction matters when calculating screening yield and describing the final sample. Our quota guide explains why the achieved sample and target population should be reported separately.
If a panel supplies a respondent ID and completion redirect, test each terminal path: eligible complete, ineligible screen-out, eligible quota closure and technical failure. Return only the agreed status and identifier to the panel. Do not include sensitive screener answers in a redirect URL. Survey Studio's Help Center is the starting point for configuring and checking the current collection workflow.
Run a small pilot that covers every exit
Before launch, use fictional test respondents to take every route. Test people who qualify on the first path, who fail each criterion, who choose missing options and who arrive after a quota closes. Verify the visible message and the stored status for each path. Check on a phone as well as a desktop, including large text and keyboard navigation.
Inspect the saved records, not only the preview. A screen-out should not appear as a completed main questionnaire, and a person who completed should not be counted twice. If a respondent goes back and changes an answer, the final route and retained answer must still agree. Our pilot checklist provides a broader launch test across logic, saved answers and reports.
With a real pilot, review where people pause or choose “Other” and whether they interpret the question as intended. Revise wording before the full field period if possible. Keep a version record; changing a screener midway can change who enters the study and make early and late completes less comparable.
Read the screening funnel without overstating it
Track invitations or entries, starts, answers to each screener, ineligible exits, quota closures and qualified completes. Use counts with percentages and name the base. “120 of 200 people who answered the eligibility question qualified” is a different statement from “120 of 500 invitations qualified.” Do not fold people who never opened the survey into a question-level denominator.
Fictional illustration: 500 invitations produce 260 starts. Of those starts, 240 answer the purchase screener. Ninety report a qualifying purchase, 130 report none and 20 are unsure. If all 90 qualified respondents then complete, the observed qualification rate among screener answerers is 90/240, or 37.5%; the completed-survey share of starts is 90/260, or 34.6% to one decimal. Neither figure estimates the share of qualifying people in the invitation list, because 260 invitees did not answer that screener. Report the 20 unsure responses explicitly.
Compare funnel counts by recruitment source only when definitions match. If one source pre-screened its members and another did not, their qualification rates are not directly comparable. The denominator guide shows how to display bases and missing answers clearly.
Check for unfair exclusion and alternative explanations
After a pilot, ask which eligible people the wording may wrongly exclude. Does “your household” make shared-household purchasers unsure? Does a job title list omit people doing the same work under different titles? Does an English-only screener exclude part of the population the research claims to cover? These are measurement and coverage questions, not just interface questions.
Look at screen-out reasons by source and, where privacy and sample size permit, by relevant subgroup. A high screen-out rate may mean poor targeting, an overly narrow rule, confused wording or fraudulent traffic. The funnel alone cannot distinguish these explanations. Review the question text, recruitment promise, timing and any quality flags before changing criteria.
Do not loosen eligibility merely to hit a deadline while still describing the original population. If the study scope changes, record when and why, report results separately when needed, and update the study description. Conversely, excluding respondents after seeing their substantive answers can bias results; quality decisions need a documented rule and review. See the response quality guide.
Report the rule and the limits of the resulting sample
A methods note should include the target population, recruitment sources, exact screener wording and answer options, logic and quota rules, dates, counts at each exit, any mid-field changes, and what was done with missing or ambiguous answers. AAPOR asks for enough method detail to let readers independently assess the research claim. Keep an internal routing table even when a shorter public summary is appropriate.
Make one final check: does the headline still describe only the people who were eligible and reachable through this recruitment method? If the sample is an opt-in panel of recent buyers, do not present it as a probability sample of all adults. If the study covered only one country or purchasing channel, put that boundary near the result, not buried in a footnote.
Survey Studio can help build and review surveys and their results. The research team still owns eligibility, recruitment and reporting decisions. To explore the current workflow, use the interactive sample or create a survey, then test every route before inviting real participants.