Estimation and testing

A researcher divides schools into urban and rural strata, then randomly samples students from both urban schools and rural schools. How does this differ from selecting several schools and surveying every student in only those schools?