# Cluster Sampling

Cluster sampling (also known as one-stage cluster sampling) is a technique in which clusters of participants that represent the population are identified and included in the sample[1]. Cluster involves cluster of participants that represent the population are identified and included in the sample.  This is popular in conducting marketing researches.

The main aim of cluster sampling can be specified as cost reduction and increasing the levels of efficiency of sampling. This specific technique can also be applied in integration with multi-stage sampling.

A major difference between cluster and stratified sampling relates to the fact that in cluster sampling a cluster is perceived as a sampling unit, whereas in stratified only specific elements of strata are accepted as sampling unit.

Accordingly, in cluster sampling a complete list of clusters represent the sampling frame. Then, a few clusters are chosen randomly as the source of primary data.

Area or geographical sampling can be specified as the most popular version of cluster sampling. Specifically, a specific area can be divided into clusters and primary data can be collected from each cluster to represent the viewpoint of the whole area.

The nature of cluster analysis depends on comparative size of separate clusters. If there are not major differences between sizes of clusters, then analysis can be facilitated by combining clusters. Alternatively, if there are vast differences in sizes of clusters probability proportionate to sample size can be applied to conduct the analysis.

## Application of Cluster Sampling: an Example

Imagine you want to evaluate consumer spending on various modes of transportation in Greater London. Since Greater London is a large area, we need to sample from only 6 boroughs out of total 32 boroughs of the Greater London.

There are three stages for the application of cluster sampling:

1. Select a cluster grouping as a sampling frame. In example above, all 32 boroughs of the Greater London represent the sampling frame for the study
1. Mark each cluster with a unique number. We can easily number each borough from 1 to 32.
1. Choose a sample of clusters applying probability sampling. Using systematic random sampling (or any other probability sampling), we can choose 6 boroughs from the total 32 boroughs. Households residing in 6 boroughs will represent samples for the study.

1. It is the most time-efficient and cost-efficient probability design for large geographical areas
2. This method is easy to be used from practicality viewpoint
3. Larger sample size can be used due to increased level of accessibility of perspective sample group members