When should I use sampling techniques?

When training programmes involve a large number of learners you may wish to evaluate using a smaller, carefully chosen ‘sample’ to represent the group as a whole.

Some basic sampling techniques include:

Random sampling

Each member of the total target group is given a unique number. A sample size is chosen and then a corresponding number of respondents is chosen randomly using, for example, a lottery method or an on online random number generator.

Systematic sampling

This method is often used instead of random sampling. First the total target group number (X) is divided by the required sample size number (Y) to create the ‘sampling interval’ (N). A starting point between 1 and N is then randomly selected, and the sample is then formed from every Nth person.

Example: 100 target group learners/sample size of 20 = sampling interval of 5. Random starting point between 1 and 5 = eg 4. The sample is then formed from the. 4th, 9th, 14th, 19th etc person on the list.

Stratified sampling

This method relies on identifying relevant ‘strata’, which are subsets of the total target group that share at least one common characteristic. Examples of strata include gender, age groups, ethnic groups, managers, non-managers, departments. Random sampling is then used to select a ‘sufficient’ number of subjects from each stratum, ie a sample size large enough to be reasonably confident that the stratum represents the total target group.

Judgment sampling

The sample is selected on the basis of judgment. For example, you may decide to draw the entire sample from one representative department, even though the total target group includes a number of departments. When using this method, you should be confident that the chosen sample is broadly representative of the total target group.

Quota sampling

As with stratified sampling, first relevant strata are identified along with their proportions as they are represented in the total target group. Then judgment sampling is used to select the required number of respondents from each stratum. This differs from stratified sampling, where the respondents are selected by random sampling.

Snowball sampling

This method is used when the desired sample characteristic is rare or when it is difficult to get people to participate. The method involves getting referrals from initial respondents to generate further respondents. While this can be a useful method in the above circumstances it should be noted that the sample produced is unlikely to be representative of the total target group.



Article Quick Jump :