- When population of interest is spread over large area. It may not be feasible to make up a list of every person living within the area from the list.
- Subdivide large area into smaller units (state to counties or city to precincts of school boundaries).
- It is important that clusters be similar with equal heterogeneous mix of individuals
For example, a city can be divided into 12 clusters, we randomly selected 4 clusters and then all their members become the sample