Chapter 9: Control Charts for Attributes

In this chapter, we are going to discuss a selected amount and types of charts used for attribute data. Our selection is based on tradition, efficiency, and value.

OVERVIEW

As we discussed in the last chapter, control charts, like data, are divided into two categories. Variables data are usually studied with Xbar and R charts or with Xbar and s charts. Attribute data are often studied with p, np, c, and u control charts.

Attribute data are numbers that result from counts of visual observations. They portray characteristics classified as acceptable-reject, go-no-go, or conforming-non-conforming. Attribute data are categorical and are written as whole numbers (integers).

This broad category of data is divided into two types: defects and defectives. A defect or nonconformity is a specific characteristic that does not meet an engineering specification or inspection standard. An individual part may have more than one nonconformity (e.g., open welds per armature, errors per invoice, wrong diagnosis per 100 patients , scratches per pressure tube, dents per primary coil, and so on). Systems that produce defects are studied with c or u control charts.

A defective or nonconforming part does not meet an engineering specification or inspection standard. Each defective part has at least one defect. An individual part is classified as defective only once (e.g., motor with a soft armature shaft, ignition coil that produces low energy levels, motor with a shorted frame, an invoice with the date wrong). Systems that produce nonconforming parts (defectives) are studied with np or p control charts.