# Cluster Sampling

-

Cluster Sampling is a sampling technique where a population is divided into roughly equal units called clusters and then a simple random sample of these clusters is drawn. Note that once a cluster is chosen we include every individual in that cluster as part of our sample.

When applying this technique, the clusters should be broadly similar to each other and of approximately same size. There can be internal variation between the units in a particular cluster.

Consider the following example. Suppose a city has 200 districts of roughly the same population size. We want to find a sample to study the electricity consumption pattern of the city dwellers using cluster sampling technique. We will first randomly choose say around 10 cities using the simple random sampling technique. All of the people living in these 10 cities are then included in our sample.

How to carry our cluster sampling?

1. Divide the population into clusters which are broadly similar and of equal size.
2. Randomly choose which of the clusters you want to include in your sample.
3. Include all units in the chosen clusters as part of our sample.

Types of Cluster Sampling:

1. Single Stage Sampling – In this method, we randomly select our clusters and choose all units in them. So we are done in a single step. This is the example that we considered above,
2. Two Stage Sampling – In this method, there are sub-clusters within the clusters. We first randomly choose our clusters and then randomly choose the sub-clusters among them. As an example suppose there is a state which is divided into cities which are further divided into districts. We first randomly choose our cities and then within each city we randomly choose certain districts. All the people living in those districts become part of our sample.
3. Multi-Stage Sampling – Here the sub-clusters are further divided into even smaller groups and the same procedure is repeated. As an example a country can be divided into many states further subdivided to cities which are in turn divided into districts. We first randomly choose some states and then randomly choose cities in them. The randomly chosen districts in those cities become part of our sample.

Difference between Stratified Sampling and Simple Random Sampling:

The main difference between simple random sampling and stratified random sampling is that when we do stratified sampling we do not include all units in the cluster as part of our sample. From each cluster, we randomly choose some units in proportion to the size of the cluster. On the other hand in cluster sampling, all units in a chosen cluster are included in the sample.

Also in stratified sampling, all clusters are represented whereas, in cluster sampling, all clusters are not represented in our sample.