The purpose of sampling is to determine the characteristics of a population using a relatively small sample taken from the population.
If a sample is to give meaningful results it must be representative of the population. There are various ways of attempting to achieve this.
Convenience Sampling is a sampling method in which units are selected based on easy access/availability. The disadvantage of convenience sampling is that the units that are easiest to obtain may not be representative of the population:
- products on top of a box of parts may be a different quality from those at the bottom
- people who are at home when the market researcher calls may not be representative of the entire population
Random Sampling involves taking items for the sample in a way that ensures that any unit in the population has an equal chance of being selected. This is intended to ensure that the sample is not biased by the method of collection. Bias may be inadvertently introduced by selecting items that are easy to collect, or accessible.
Stratified Sampling is a sampling method in which the population is split into several categories that share common characteristics. Items are collected at random from each category, in proportion to the size of the category relative to the population. Stratified sampling may give more reliable results than pure random sampling because it ensures that all categories are fairly represented.
Systematic Sampling involves taking a sample from the available data in a set pattern, rather than at random. A telephone survey that called every hundredth caller listed in the phone book would be a systematic sample. Systematic sampling is often convenient and economical, but carries the risk that there will be an unsuspected systematic pattern in the data.
For more information on sampling refer to MiC Quality Six Sigma Glossary.