Measurement Systems Analysis

Abstract

The EAS bases most of his or her decisions about design parameters on measurements, so it is imperative that those measurements be trustworthy. There are two categories of measurement quality we will discuss, namely accuracy and precision. Accuracy is the degree to which the measurement differs from the truth, on the average. Precision is the degree to which the measurements vary from instance to instance of measuring the same unit or item. These definitions would probably be better termed “inaccuracy” and “imprecision”, but we will use the terms “accuracy” and “precision” to more or less mean inaccuracy and imprecision. In order to assess accuracy, the true value of the dimension or performance parameter for each part (or item being measured) in the sample must be known. In lieu of knowing the true value, a more trustworthy “reference method” might be used to provide a more accurate and precise measurement against which the measurement system to be evaluated will be compared. This chapter will treat both the case where only precision can be evaluated (no reference) and where accuracy may also be evaluated (in comparison to a reference method result).