This is another type of restricted random sampling which the whole universe is divided into certain prominent subgroups called clusters, and certain such clusters are selected random to provide for the sample. All the units belonging to these selected clusters constitute the sample desired.
For example, to conduct an opinion poll on the reservation issue, the whole country may be divided into certain clusters say 500 blocks and out of the 5000 blocks, 50 blocks may be selected at random. Now all the inhabitants of these 50 blocks (clusters) will constitute the sample units to provide with the desired information. This method is invariably used in conducting an enquiry relating to some common problem of a population. However in using this method of sampling attention must be paid to the following important points:
(i) The clusters should be as small as possible in their size.
(ii) The numbers of sample units in each cluster should be more, or less the same.