ABSTRACT

Sampling effectiveness is based on the theory of probability. If a batch of parts is obtained from a controlled process where process control or another prevention-based system is employed, then a very large sample will show favorable results. However, if the process is not so managed, then small samples could indicate noncomplying parts. The presence of many parts "hugging" a specification, for example, might signal sorted material or a process controlled to a specific region of a tolerance for some specific reason (engineering fit preferences, readiness for a thick or heavy finish, tool size, etc.).