MSA Attribute Data Overview

Many measurement systems deal with qualitative data. For example, parts coming off a production line might be considered as simply good or bad (go/no-go). Although information may be lost when measuring a quality characteristic on an attribute level, it is sometimes simpler to understand and may alleviate the need for expensive precise devices and time-consuming measurement procedures.

The MSA method for attribute data is a straightforward method that can be used to assess the accuracy of appraisers and the types of mistakes they are likely to make. Typically, samples of parts are appraised by operators as good or bad. These classifications are then compared with a reference or standard. The following measures are then calculated:

Effectiveness (E)
The ability of an appraiser to distinguish between defective and nondefective products or parts.
Probability of False Rejects ( FR)
The likelihood of rejecting a good part (i.e. a false alarm).
Probability of False Acceptance (FA)
The likelihood of accepting a bad part (i.e. miss rate).
Bias (B)
A measure of an appraiser’s tendency to falsely classify a part as good or bad.