Sampling

Sampling is the selection of a set of elements from a population or product lot. Sampling is frequently used because data on every member of a population are often impossible, impractical or too costly to collect. Sampling lets you draw conclusions or make inferences about the population from which the sample is drawn.

When used in conjunction with randomization, samples provide virtually identical characteristics relative to those of the population from which the sample was drawn.

Beware, however, of three categories of sampling error:

• Bias (lack of accuracy)
• Dispersion (lack of precision)
• Non-reproducibility (lack of consistency)

These are easily accounted for by knowledgeable practitioners.

Determinations of sample sizes for specific situations are readily obtained through the selection and application of the appropriate mathematical equation. All that’s needed to determine the minimum sample size is to specify:

• If the data are continuous (variable) or discrete (attribute).
• If the population is finite or infinite.
• What confidence level is desired/specified.
• The magnitude of the maximum allowable error (due to bias, dispersion and/or non-reproducibility).
• The likelihood of occurrence of a specific event.

Excerpted from Jack B. ReVelle’s Quality Essentials: A Reference Guide from A to Z, ASQ Quality Press, 2004, page 164.