How is cluster sampling different from stratified random sampling?

In statistics, both stratified simple random sampling and cluster sampling involve partitioning a population into groups, but they differ significantly in their objectives and execution.1

• Stratified Random Sampling: The population is divided into groups called strata that are internally homogeneous (similar). A simple random sample is then drawn from every stratum, and the results are aggregated. This method is often used when researchers want to ensure representation across specific subgroups or make inferences about those individual subpopulations.12
• Cluster Sampling: The population is divided into groups called clusters. Unlike strata, it is desirable for these clusters to be heterogeneous (diverse) and representative of the population as a whole. In single-stage cluster sampling, researchers randomly select entire clusters and collect data from every unit within them. In two-stage cluster sampling, researchers select clusters first and then take a random sample of units from within those selected clusters.12

While stratification aims to increase precision by grouping similar units, cluster sampling is often employed for practical advantages, such as area sampling based on geographic sections.1