Factors affecting the performance of a system

The factors that affect a measurement system can be categorised generally into those that affect the result and those that affect the variability of the measurements.

Factors related directly to the result can be assessed using stability and linearity; variability factors are assessed by repeatability and reproducibility.

Stability and linearity

Assessment of stability provides a means to determine the sensitivity of the measurement system to change and the rate of recurrence of the change. If the measuring process is changing over time, the value of the acquired data is reduced. Knowledge of the stability gives confidence that the system can be used to reliably and consistently make measurements. Calibrations provide some information about changes to the measuring instruments, and by maintaining calibration histories; it is possible to study the long-term stability of parts of the measurement system.

Stability is determined through repeated measurements of a single characteristic over time. The variation of the measurements is a direct indication of the stability. When analysing the measurement system it is important to consider the appropriate measurement time interval. Knowing when and how the system will be used can help to identify situations when the system may be unstable. The greater the likelihood of change to the measurement system, the shorter the interval should be between measurements.

Linearity can be determined by assessing measurements over the entire operating range of the system. The same relationship between input and output to the system is expected to hold true over the operating range. The linearity over the input level range can be found by plotting the measurement result from the lowest level to the highest level. The linearity of the system is represented by the slope of a least squares best fit line through these points.

Repeatability and reproducibility

Repeatability refers to the variation of measurements made with one system. Repeatability conditions are therefore those that exist for repeated application of the same measurement method on the same measurand within a short interval of time under the same conditions (same laboratory, same operator, same apparatus).

Reproducibility refers to the variation of measurements made with one system, under different conditions. Reproducibility conditions are therefore those that exist for repeated application of the same measurement method on the same measurand at different times and under different conditions (different laboratory, different operator, different apparatus). The standard deviation of reproducibility, therefore, includes the standard deviation of repeatability.

More about repeatability and reproducibility may be found in: BS ISO 5725 : Part 1 : 1994 'Accuracy (trueness and precision) of measurement methods and results – Part 1: General principles and definitions' and BS ISO 3534 : Part 1 : 1993 'Statistics - Vocabulary and symbols - Part 1: Probability and general statistical terms'

Uncertainties

Uncertainty is part of any measurement and results presented without a statement of the associated uncertainty should be regarded as incomplete. The estimation of uncertainty is thus an essential part of any measurement procedure. There are various theoretical arguments about the method of combining the different types of uncertainty to produce one overall value. UKAS Publication M 3003 recommends a procedure for doing this. It makes best use of the information available and results in a single ± figure to apply to the measured value. This method is adopted at NPL to calculate uncertainties. More information can be found in NPL Publication, "A Beginner's Guide to Uncertainty of Measurement".

Acoustical measurement systems >>