Sampling

Sampling is used to make an estimate of the characteristics of the population as a whole. Sampling overcomes the impossibility in almost every market of asking all members of a population their opinion

The sampling process. Wilson (2003) outlines a six-stage sampling process:

n Define the population of interest

n Determine whether to sample or census

n Select the sampling frame

n Choose a sampling method

n Determine the sample size

n Implement the sampling procedure

There are two broad sampling methods:

The key characteristic is that every member of the population of interest has a known chance of being selected, independent of any subjective selection by the researcher

Four commonly used methods of probability sampling:

n Simple random sampling

n Systematic sampling

n Stratified random sampling

n Cluster sampling

Each member of the population has an equal chance of being selected for the survey. Members are randomly selected by a computerized random number generator or tables until the required sample size is filled

This method divides the population into two or more mutually exclusive groups, e.g., men or women, users or non-users of a product, and takes random samples from within them, using either of the methods above

Cluster sampling is a procedure in which clusters of population units are selected at random and then all or some of the units in the chosen clusters are studied. The technique works when a population can easily be divided into representative clusters, e.g. in membership directories

Examples of non-probability sampling are:

Based on the convenience of the researcher

The researcher consciously selects a sample considered appropriate for the study. This may be based on certain companies representing a sector. The market might include all major department stores in the sample, as well as a random selection of other outlets

Is defined by ESOMAR as ‘A type of non-probability sample where the required number of units with particular characteristics are specified’. This is based on the idea that, if known characteristics of the population are reproduced in the same proportion in the sample, it is representative of that population; e.g. age, sex and social class can be used to select quotas

n Speed and cost

n Allows sampling to take place where a sample frame may not be available but key characteristics of the population are known – e.g. in overseas b2b research

n Interviewers do not have to interview named individuals, they are screened in or out via a small number of classification questions

n The data, when compared to random methods which are perhaps double the cost, has been proven to be acceptable, provided that the research is managed effectively

n Cost savings may be used to improve the quality of research, through increasing sample sizes or using a different method in support of the survey

n Its popularity shows that it works!