This is called cluster sampling, where the researcher divides the population into groups (clusters) then takes a random sample from among the different clusters.
9.5m questions
12.2m answers