Data-driven problem solving is critical for organizations to remain competitive; however, good decisions are built upon accurate measurements. A calibration schedule for instruments must be completed at the right time to ensure minimum bias, the evaluation of as-is condition gives light to serviceability and the adequacy of calibration efforts. This article highlights the use of linearity and bias studies to extract information on the accuracy of measurements, highlighting instruments that may need further investigation.
Measurement devices produce data used to ensure processes are in control and capable of meeting requirements. Measurements include uncertainty, which may interfere with good decision-making. Measurement uncertainty involve three components: accuracy, precision, and random variability. The use of the stopwatch function on a personal device provides a good illustration of each. The difference between the reading indicated and what could be determined with device known to produce robust results denotes the accuracy of the reading. Obtaining the same result with repeated tries or reproduction of the same result by multiple people is considered precision. Nothing can be done about the third component, random variability. Further study into precision involves measurement systems analysis, which is not part of this article and strongly suggested as a follow-up topic for readers. Calibration of measuring devices deals with one of the three sources of uncertainty: accuracy.