~When using cluster sampling, the population being sampled is separated into sections (Clusters), then randomly select some of these clusters for your sample. All members of the selected clusters are used in the sample. This approach saves much time and money associated with sampling a very dispersed population.
~Let us assume you wanted to conduct interviews with apartment managers in a major city about some aspect of their apartment complex. You would first separate the entire city into sections, then randomly select some of these sections (clusters) for your sample. You would then interview all apartment managers in these clusters.
~This is referred to as "one-stage cluster sampling".
~If all persons to be interviewed are employees within the selected clusters, it is call "two-stage cluster sampling".
~You could also combine cluster sampling with stratified sampling.
~For example, in the above example, you might want to stratify the employees based on some attribute most relevant to your study (i.e., salary, job function, seniority, etc.) and then select the employees from each of these strata. This type of sampling is referred to as "two-stage sampling". In more sophisticated studies, there could be many stages (i.e., multi-stage sampling).
~Cluster sampling is used quite extensively in sampling the results of elections of all sorts. (randomly selecting precincts across a state or county & noting how a random number of people voted). This again, can be combined with a stratified technique by separating voters in different income levels, political parties, or gender.