How can I build a sampling frame for a small survey without quietly leaving people out?

A sampling frame is the list or other workable way you identify and select the people—or other units—you may ask. It is not automatically the same as the population your survey is meant to describe. First define that target population in plain language; then compare it with the frame you have. A good frame ideally covers all members of the target population and gives you a clear way to identify and select them. If groups are left out or overrepresented, the people selected may not represent the population well, limiting what you can generalize. [S1]

For example, suppose your intended population is “people who live in Cedar Ward and are 16 or older,” and your question concerns participation in local activities. These boundaries are hypothetical; choose definitions that fit your question. A list of people who attended last year’s activities would not, by itself, cover residents who never attended. It might suit a survey specifically about past attendees, but that is a different target population. Write your population definition before choosing the list.

Review the frame in both directions. Ask who belongs in your target population but might not appear, and who appears on the list but does not belong. Check for duplicate entries, unclear eligibility, and records that do not identify a distinct person or unit. Keep a simple gap log rather than assuming the list is complete.

For the hypothetical Cedar Ward survey, imagine your frame is a community-center sign-up list. Ask whether it includes residents who do not use that center, whether sign-ups are current, and whether nonresidents are listed. These are questions to check, not claims about a real center. The key issue is whether the list’s coverage matches the question’s target population. A review cannot guarantee that every missing person will be found. [S1]

You can consider more than one frame or recruitment route, but document this as a coverage decision, not proof that everyone is reachable. For example, in the hypothetical survey, you might compare an activity sign-up list with another list you are permitted to use that identifies residents. Before combining them, decide how you will recognize duplicates and confirm eligibility; otherwise, someone could appear more than once or an out-of-scope person could be invited. If you add open invitations, record that this route differs from selecting people from a defined list.

Describe which lists or routes you used, who they appear to cover, and known gaps. If the frame omits parts of the intended population, say so and avoid presenting results as though they necessarily describe everyone. A frame review cannot guarantee a representative sample or repair every coverage gap. [S1]

Use this copyable note, replacing the bracketed parts. The Cedar Ward details are examples only:

“Target population: [who the survey is intended to describe, including boundaries]. Survey unit: [person or other unit]. Frame(s): [list or invitation route]. Eligibility rule: [who may be selected]. Likely omissions: [groups or units the frame may miss]. Possible out-of-scope entries or duplicates: [what to check]. Coverage limitation: [what the frame cannot establish]. Selection plan: [how eligible units will be chosen].”

Hypothetical opening: “Target population: Cedar Ward residents aged 16 or older. Frame: community-center activity sign-up list. Likely omission to check: eligible residents who have not signed up for center activities.” This is a warning to investigate, not evidence that anyone is actually missing or a count of how many people are affected. Keep the population definition, frame description, and limitation together so readers can see what your survey can—and cannot—speak for.