SamplingFrame

A sampling frame is a practical representation of the population from which a sample is drawn. It is usually a list, database, registry, or other device that contains the units eligible for selection and serves as the basis for implementing a probability sampling design. Ideally, the frame should include all elements of the target population and exclude elements outside it; in practice, coverage errors occur when some target elements are missing (undercoverage) or nonmembers are included (overcoverage).

Common examples include electoral rolls, voter or customer databases, telephone directories, address registries, and business lists. For online studies, sampling frames may be panels, user databases, or collections of cookies or IP addresses. The choice of frame depends on the population definition and the sampling unit, which may be an individual, household, organization, or other entity.

Frame quality directly affects survey accuracy. Coverage error, duplication, outdated contact information, and misclassification reduce representativeness and can increase bias or variance. Frame management involves updating records, de-duplicating entries, and sometimes linking multiple frames to improve coverage. When the population is dynamic, such as ongoing customer activity, the frame must be periodically refreshed to maintain relevance.

Use of a frame is central to many probabilistic sampling methods, including simple random sampling, systematic sampling, stratified sampling, and cluster sampling. When a reliable frame is not available, researchers may adopt nonprobability methods or construct composite frames from multiple sources, acknowledging potential biases and reporting frame-related limitations.