Research Methods in Tourism, Hospitality and Events Management

4.4.3 Cluster sampling

Where surveys cover very large areas, such as whole countries, the types of sampling

methods described to this point may not be very cost-effective. Consider, for

example, a company wishing to conduct a nationwide survey of a sample of its

customers using personal interviews. Random sampling may well yield a list of

potential respondents spread widely across the country but thinly in some areas, thus

interviewers may have to travel very long distances to conduct a small number of

interviews. To overcome this type of problem, the researcher identifies groupings or

clusters of respondents in a particular area. A range of clusters are selected randomly

and then, within each cluster, respondents are again selected on a random basis. This

method of sampling is called ‘cluster sampling’ (some older texts may refer to this as

area sampling). This method can also be useful where the extent of the survey

population is not fully known, as shown in the following example.