Forms · Glossary
What is a sampling frame?
A sampling frame is the actual list, map or source from which a survey sample is drawn, such as a customer database, a staff directory, an electoral roll or a file of postal addresses. It stands in for the target population, and any gap between the two, people missing from the frame or on it who should not be, becomes coverage error.
Every probability survey depends on one, and most survey problems that look like bad questions start here instead. If a group is not on the list, no amount of random selection or weighting will put it back.
Indunil Asanka · Co-founder
4 min read · Published
| Survey | Target population | Sampling frame | Common coverage gap |
|---|---|---|---|
| Staff engagement | All current employees | HR system export | Casuals and contractors missing, leavers not yet removed |
| Customer satisfaction | Everyone who bought in the last year | Customer email list | Guest checkouts and in store cash buyers missing |
| Household survey | All residential households | A national address file | New builds and informal dwellings missing |
| Member survey | Current club members | Membership register | Lapsed members still listed, family members sharing one email |
Population, frame and sample
Survey design names three things. The target population is who you want to describe. The sampling frame is the list you can actually select from. The sample is the people you select. The ABS lists defining the population, frame and units as the first step of sample design, and notes that probability samples require a frame, while non probability samples do not. Pew Research Center's American Trends Panel is a well documented example: since 2018 it has recruited by sending invitations to a random sample of households drawn from the US Postal Service's master list of residential addresses, which serves as the frame.
Coverage error
When the frame leaves out part of the population, results describe only the part it covers. Undercoverage, the usual problem, happens when lists are out of date or exclude whole groups, such as households with no landline in a phone survey. Overcoverage happens when the frame includes people outside the population, such as former customers, or lists the same person twice, which gives them a higher chance of selection. Statistics Canada notes that the characteristics of the survey frame can limit which probability design is possible, because stratifying or sizing a sample needs information that is on the frame.
Checking a frame before you sample
Compare its total with an independent count of the population. Look for duplicates by email, phone or name. Remove people who left, closed or died. Check that the variables you will stratify on are filled in for everyone. Ask who would be missing and whether they might answer differently; casual staff, recent customers or people without email often do. If a group is missing, either add a second frame for it or describe the survey's population more narrowly so the results are honest.
Common mistakes
Treating a mailing list built from people who opted in to marketing as a frame for all customers. Drawing a sample from an export that is months old. Sending a survey to everyone on a frame and calling it a census without accounting for non response. Merging two lists without removing people on both. And describing results as representative of a population that the frame never covered.
Using a frame with the form builder
The form builder does not hold sampling frames or draw samples, so the frame is your own list, kept wherever it lives. Draw the sample there, then share the form. The login required access type can restrict the form to the email addresses or whole domains you allow, which keeps responses inside the frame instead of letting a forwarded link pull in outsiders. The check happens when the form opens and the account is not saved with the answers. Responses export to CSV with one column per question and the submission time first, ready to compare with the frame.
Questions people ask
What is an example of a sampling frame?
A university's list of enrolled students used to sample students for a wellbeing survey, a council's rates register used to sample property owners, or a company's HR export used to sample staff. In each case the list stands in for the population, and its gaps decide who the survey can describe.
Is a sampling frame the same as the population?
No. The population is everyone you want to describe; the frame is the list you can reach. The ideal frame matches the population exactly, but real frames miss some people, include some who do not belong, or list some people twice. The gap between them is coverage error.
Can a survey have more than one sampling frame?
Yes. Dual frame designs combine two lists, such as landline and mobile numbers, or a customer database plus an in store intercept, to cover more of the population. People who appear on both have a higher chance of selection, so the design must identify overlaps and adjust weights for them.
Do online surveys have a sampling frame?
Only if you sample from a list, such as a customer email list or a panel. An open link posted on social media has no frame, because anyone can find and answer it and nobody's chance of selection is known. That makes it a convenience sample, whatever its size.
How do I reduce coverage error?
Use the most complete and current list available, combine lists to fill known gaps, remove duplicates and people outside the population, and record what the frame misses. Where a group cannot be reached, say so in the results rather than weighting other respondents to stand in for them.
Make one with forms
The button opens the generator with this use case already described. Change the wording to match your own.
Create a form with OneCraftRelated questions
- What is convenience sampling?Convenience sampling surveys whoever is easiest to reach, such as passers by or a shared link. Why it is quick, where its bias comes from, when it is enough.
- What is stratified sampling?Stratified sampling splits a population into groups, called strata, and samples each one. A worked staff survey example, proportionate and disproportionate.
- What is a representative sample?What is a representative sample? One that mirrors the population on the traits that shape the answers. Random, stratified, convenience and quota compared.
- What is a good survey response rate?What is a good survey response rate? It depends on channel and audience. Dated benchmarks from Pew, the US Census Bureau and SurveyMonkey, read with care.
Step by step in the builder: Control who can fill in your form, then Create a form with AI.
Written and checked by the OneCraft team. Last checked .