Cluster sampling is a sampling method in which a population is partitioned into groups called clusters, a sample of clusters is selected, and all or a further sample of the units in the selected clusters are observed. Observing every unit in the selected clusters is called one-stage cluster sampling; sampling units within the selected clusters is called two-stage cluster sampling. Unlike stratified sampling, clusters are normally chosen to resemble small versions of the whole population, whereas strata are chosen to be internally homogeneous, meaning that units within the same stratum are similar with respect to the variables used to form it.

See also

Cluster, Population, Sample, Sampling, Simple Random Sample, Stratified Sampling

Explore with Wolfram|Alpha

References

Cochran, W. G. Sampling Techniques, 3rd ed. New York: Wiley, 1977.

Cite this as:

Weisstein, Eric W. "Cluster Sampling." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/ClusterSampling.html

Subject classifications