Sampling Techniques

There are essentially two sampling techniques: Nonprobability and Probability

Nonprobability Sampling – Relies on the personal judgment of the researcher rather than chance to select sample elements.  Estimates obtained are not statistically projectable to the population.

Probability Sampling – Sampling units are selected by chance.

Nonprobability Sampling Techniques

Convenience Sampling:  Respondents are selected because they happen to be in the right place at the right time.  Examples include mall-intercept interviews without qualifying the respondents and pop-up surveys on a website.  Sources of selection bias are present and convenience samples are not representative of any definable population.  Convenience sampling is not appropriate for marketing research projects involving population inferences.

Judgmental Sampling:  Population elements are selected based on the judgment of the researcher.  Example: test markets, bellwether precincts selected in voting behavior research.

Quota Sampling:  Ensures that the composition of the sample is the same as the composition of population with respect to the characteristics of interest.

Snowball Sampling:  Respondents are asked to identify others who belong to the target population of interest.

Probability Sampling Techniques

Simple Random Sampling:  Each element of the population has a known and equal probability of selection.  In SRS the results may be projected to the target population.  However, there are significant limitations.  First, it is difficult to construct a sampling frame that will permit a simple random sample to be drawn.  Second, the result can be samples that are very large or spread over large geographic areas.

Systematic Sampling:  The sample is selected by choosing a random starting point and then picking every ith element is succession form the sampling frame.  It is assumed that the population elements are ordered in some respect.

Stratified Sampling:  A two-step process in which the population is partitioned into subpopulations, or strata.  The elements within a stratum should be as homogeneous as possible, but the elements in different strata should be as heterogeneous as possible.  Variables commonly used for stratification include demographic characteristics, type of customer, or size of firm.  The number of strata to use is a matter of judgment, but experience suggests the use of no more than six.

Stratified sampling can be proportionate or disproportionate.  In proportionate stratified sampling, the size of the sample drawn from each stratum is proportionate to the relative size of the stratum in the total population.  In disproportionate stratified sampling, the size of the sample from each stratum is proportionate to the relative size of that stratum.  For example, large retail stores might be expected to have greater variation in the sales of some products as compared to small stores.  Hence, the number of large stores in a sample may be disproportionately large.

Cluster Sampling:  The target population is first divided into mutually exclusive and collectively exhaustive subpopulations, or clusters.   The key difference between cluster sampling and stratified sampling is that in cluster sampling only a sample of the subpopulations (clusters) are chosen, whereas in stratified sampling, all the subpopulations (strata) are selected for further sampling.  Elements with a cluster should be as heterogeneous as possible, but clusters themselves should be homogeneous as possible.  Ideally, each cluster should be a small-scale representation of the population.  However, it is often difficult to form heterogeneous clusters because, for example, households in a block tend to be similar rather than dissimilar.

Choosing a Technique

If nonsampling errors are likely to be an important factor, then nonprobability sampling may be preferable, as the use of judgment may allow greater control over the sampling process.  A more heterogeneous population would favor probability sampling.   Nonprobability sampling should be used in concept tests, package tests, name tests, and copy tests where projections to the populations are usually not needed.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: