Effects of sample sizes in phases I and II on p control chart performance

This study expands on the work of Braun (1999) to study the effects of the initial reference sample size, m, in Phase I and the on-line sampling size, ni, in Phase II on the statistical performance of a conventional p chart. The conditional and marginal distributions of run length are derived. The oscillation of ARL with known p
_ο
is investigated. The recommended values of sample sizes to achieve required control chart performance are provided.