9.4 : データ検証

Method validation is the process employed to ensure a method produces replicable results with acceptable quality.

It is necessary when the method is routinely used to analyze a different sample matrix or via a modified procedure.

This solicits various method validation parameters like method specificity, linearity, accuracy, precision, range, the limit of detection, the limit of quantitation, and robustness.

To estimate these parameters, it is necessary to distinguish the contribution of random and systematic errors from the method and the analyst.

For this, the standard deviation of the replicate measurements provides information on the random error and precision of the method.

Analysis using standard reference, blank, and a larger or smaller sample size gives insight into the systematic error and the method's accuracy.

While independent analysis of a sample using alternative methods yields information on the systematic errors from the method, analysis of the same sample using the same method at different laboratories sheds light on the systematic errors from the analyst.