ABSTRACT
If you are planning to study a population that is divided into a number of naturally existing groups with leaders, you might consider using cluster sampling. For instance, all students in a high school typically have a homeroom with a homeroom teacher (the leader). Rather than drawing individual names at random and then trying to contact individual students who will be scattered throughout the school, you might find it more convenient to draw a random sample of homerooms (i.e., clusters) and ask the homeroom teachers to gather the data for you (i.e., ask them to distribute questionnaires). Of course, you will want to draw a number of the homerooms. For instance, drawing only one homeroom with 30 students would not give you a good sample if it happened to be a homeroom for the academically talented. Example 1 shows how cluster sampling might be proposed. Example 1
The use of cluster sampling is not limited to school settings. It can be used for sampling from any population that is already separated into groups, such as Girl Scouts (who are in troops) and members of the Southern Baptist Convention (who are in congregations).
