Stratification is one of critical from the seven basic Quality Control Tools. It is defined as a method of sampling to bifurcate total population into smaller, homogeneous subgroups or layers known as strata. Generally, stratified sampling allows for more precision than simple random sampling. Alternate names for stratification is Stratified sampling, proportional and disproportional random sampling etc. Also, some lists of seven quality tools do not include stratification, and instead of stratification, that list contains either flow-charts or run charts.
Some of the applications are:
• When the overall population is too big and costly to analyze in its entirety.• When the sample is too broad or general to derive useful conclusions.• To reduce sample error and increase sample precision.• Stratification is usually used in processes with a large number of sub-steps.
When to use:
• Before using stratified sampling, it should be determined whether it will be advantageous.• Before collecting data, strata should be determined.• When collecting data from a number of different sources such as employees of different ranks/tasks, shifts, branches, educational background etc.• When forming homogeneous or separate categories from the bulk of data.
One of the most common uses of stratification is observed in political surveying.