Sampling

Abstract

In order to measure the quality characteristics of a population — be it a process, a product, or a lot — it is usually desirable to select a sample from the population and examine it. Rarely is it desirable (or even possible) to examine 100% of a lot to determine its quality; therefore, some assumptions must be made so that a rational sample can be selected which represents the quality of the underlying population. Needless to say, if the sample selected does not represent the underlying population, the results of examining that sample are meaningless and can produce incorrect and perhaps costly conclusions. If a sample of cherries is selected from trees on the southern side of an orchard, a study of the degree of maturity of the sample may result in the erroneous conclusion that the orchard is ready for harvesting. The conclusion should have been that only the southern side of the orchard was mature.