Capability indices are key measures in the context of never-ending improvement in quality. Confidence bounds are derived for the common measures of process capability. The process measures are estimated based on a single random sample of observations from the normally distributed process, which is in statistical control. In practice, and in much of the quality control literature, process data are collected over time in subsamples representing rational subgroups. In this paper, therefore, we use the Patnaik’s (1950) approximation to construct the estimation and capability testing of Cpp based on multiple samples. An example is also given to demonstrate this simple approximate procedure for judging whether a stable process meets the present capability requirement.
