Statistics
Probability sampling technique in which each sampling unit is a cluster, a group of population elements, rather than a single element.
Cluster sampling.
Card 1 of 259. Answer: Cluster sampling.
Cluster sampling.
Card 1 of 259. Answer: Cluster sampling.
About · Quizzes · History of AI · To suggest new questions or report an error send me a dm.
Cluster sampling. This probability sampling method divides the population into naturally occurring groups, called clusters, like schools or villages, and randomly selects whole clusters. One-stage designs measure every element in a chosen cluster; two-stage designs take a random subsample within each one.
It is cheap: you only need a list of clusters, and fieldwork is concentrated. To test a state's fifth graders, pick 50 schools at random and test all their fifth graders, about 30 per school. The price is precision, because pupils in one school resemble each other. With intraclass correlation ρ = 0.1 and m = 30, the design effect is 1 + (m − 1)ρ = 3.9, so the 1,500 pupils carry about as much information as 385 sampled at random.
The classic follow-up is stratified sampling, which samples within every group and improves precision; cluster sampling selects only some groups and usually lowers it. Analyses must account for the clustering, for example with cluster-robust standard errors, or they overstate precision.