Although the Shewhart chart serves well as the fundamental tool for statistical process control (SPC) applications, its assumptions are challenged by many modern manufacturing environments. For example, when standard control limits are used in applications where the process is sampled frequently, autocorrelation in the measurements can result in too many out-of-control signals. This section also considers process control applications involving multiple components of variation, short production runs, nonnormal process data, and multivariate process data.

These questions are subjects of current research and debate. It is not the goal of this section to provide definitive solutions but rather to illustrate some basic approaches that have been proposed and indicate how they can be implemented with short SAS programs. The examples in this section use the SHEWHART procedure in conjunction with various SAS procedures for statistical modeling, as summarized by the following table:

Process Control Application	Modeling Procedure
Diagnosing and modeling autocorrelation in process data	ARIMA
Developing control limits for processes involving multiple components of variation	MIXED
Establishing control with short production runs and checking for constant variance	GLM
Developing control limits for nonnormal individual measurements	CAPABILITY
Creating control charts for multivariate process data	PRINCOMP