Accuracy and precision

David R. Reeve & Alan Crozier

The nature of all forms of quantitative analysis is statistical rather than absolute. Regardless of the exactness of the method, the instrumentation or the operator, all analysis involve a numbers of uncertainties. As a consequence the results of quantitative analysis tend to approach, but never equal, the true value of the parameter under investigation, and accordingly are best termed estimates. The magnitude and sign of the deviation of an estimate from the true value is the absolute error which can be expressed as a percentage of the true value to give the percentage relative error. Accuracy is defined as the concordance between estimated and true value […], and it follows that both the absolute erro and the percentage relative error provide a measure of accuracy. […]

With precision defined as the concordance between a series of measurements of the same quantity […], it is apparent that precision will provide a measure of the random or indeterminate error of the analytical system.

Fonte: Reeve, DR & Crozier, A. 1980. Quantitative analysis of plant hormones. In: J MacMillan, ed. Hormonal regulation of development I. Berlin, Springer.