Importance of the rational subgroups in designing control charts

Sampling is useful only if the sample data are representative of an entire population at the time they are taken. A systemʹs common and special causes have to be represented by each sample. If special causes are present, then the sample should reflect those. Here, Rational Subgrouping comes to the stage. Because detecting shifts in the mean requires that data is taken in so called rational subgroups. Data sets should be set up and taken in such a way that variation within a subgroup reflects common variation only, while any significant variation between subgroups reflects special causes. Rational subgrouping must be done beforehand, because setting up the Shewhart Control chart is based on the Rational Subgroup. This paper provides an overview how to choose rational subgroups while using control charts to control processes.