Start with the decision the survey will inform
Before writing questions, decide whose answers would help you make the decision at hand. A survey about improving a grocery delivery service might need recent customers who have placed an order, rather than everyone who has heard of the service. A survey about barriers to trying it may need people who have not ordered. Both groups are useful, but they answer different questions.
Write a one-sentence description of the population you want to understand. Include the relevant behavior, context, and time frame: for example, “People who personally placed at least one grocery delivery order in the past three months in the areas we serve.” Then ask whether each restriction is necessary. An overly broad definition can make answers hard to interpret; an overly narrow one can exclude people whose experiences matter.
Turn the definition into clear eligibility rules
Eligibility criteria are the conditions someone must meet to take part. They might cover experience with a product, responsibility for a purchase, location, or age when age is relevant to the research. Distinguish these requirements from characteristics you merely want to compare. If you want to learn whether occasional and frequent users differ, both groups should usually be eligible; usage frequency can be recorded for analysis rather than used to screen one group out.
Make each rule specific enough to apply consistently. “Recently used the service” leaves room for interpretation, while “placed an order in the past three months” gives respondents a clearer basis for answering. Choose a time frame people can reasonably remember. If someone shares an account with a household, specify whether placing an order, receiving one, or making the purchase decision qualifies them.
Write screeners that are easy to answer honestly
Screening questions establish eligibility before the main survey begins. Ask about observable facts rather than asking respondents to judge whether they fit your target audience. “Which of these have you done in the past three months?” is generally clearer than “Are you an active customer?” Provide response options that cover plausible situations, including “None of these” and, where appropriate, “Not sure.”
Avoid signaling a desired answer if you can. A question such as “You have purchased this product recently, correct?” may encourage agreement. Neutral wording and a logical order help respondents report their experience accurately. Keep screeners brief: people who do not qualify should not have to complete a long questionnaire to find out. If you use an incentive, describe it accurately without suggesting that a particular response guarantees entry.
Plan what happens when an answer is missing or ambiguous. Someone who cannot remember the timing of their last purchase may not meet a rule based on a specific period. That does not make their experience unimportant, but including them without a consistent rule could change what your results mean.
Separate eligibility from recruitment and quotas
How you reach people also shapes who responds. An email to existing customers cannot, on its own, tell you much about people who have never bought from you. A public link may attract a wider range of respondents, but you may have less confidence that they belong to the intended group. Choose recruitment channels with the target population in mind, and do not treat a screening question as a complete solution to recruitment bias.
Quotas serve a different purpose from eligibility rules. You might accept both new and long-standing customers but set a target number for each group so neither is overlooked. State in advance whether a quota is intended to support comparison or to reflect the audience’s composition. Reaching a quota does not make a group’s answers unimportant; it means you have stopped collecting more of them under your sampling plan.
Test the rules and describe the limits
Try the screener with a few people before launch. Ask them how they interpreted each question and check whether someone who should qualify can get through. Test edge cases, such as a person who used the product through a family member’s account. If the survey software ends participation based on an answer, verify that the routing matches your written rules.
Keep a record of the final eligibility definition, screening questions, recruitment channels, and any changes made after launch. When sharing findings, describe the people who were invited and those who were eligible, rather than referring loosely to “customers” or “consumers.” A clear eligibility boundary makes a survey more useful: readers can see whose experiences informed the findings and where further research may be needed.
