Example: Cluster Sample

• If we want to survey people from around the country by telephone we can use a cluster sample to get a representative sample.
• Out of all of the area codes in the United States, you randomly select 11 (or whatever you first sampling ratio is). This is your first cluster.

• Within these clusters, you do another random sample of exchanges within those area codes using a random number table. This is you second cluster.

...

• Now from those exchanges you randomly dial numbers to get you sample. You lack the ability to further randomize your sample because of the bias of who picks up the phone, however you still have a random sample.