ABSTRACT
To paraphrase the English poet Alexander Pope (1688-1744): To err is human; to quantify, divine. Measured values differ from their actual values; they contain errors. Quantification of the errors in any measurement should be made, so that the certainty with which a statement about the results of any analysis is known. Errors can arise during four stages of an analysis process, they are:
• calibration • acquisition • data analysis • data combination
The process of calibration does not eliminate errors but should help keep them to a minimum. Errors owing to calibration occur because a perfect standard for calibration is not feasible; for example, with a motion analysis system the control point positions, used for calibration, are not known to infinite accuracy. Calibration exploits a model to relate a known input to a measurement instrument and, therefore, a change in some aspect of the instrument, to give a desired output. These models are not perfect representations of the system.
