TABLE OF CONTENTS An analyst needs to select a suitable method for carrying out an analysis. The method needs to be validated and its measurement uncertainty determined.
Terms associated with this topic include:
Combined standard uncertainty
Coverage factor
Error
Expanded uncertainty
Measurement uncertainty (Uncertainty of measurement)
Method validation
Standard uncertainty
Uncertainty component
All laboratories need to demonstrate that their methods are fit for the purpose for which they are intended. The process of obtaining data to demonstrate this fitness for purpose and the resulting documented evidence is called method validation. The purpose of method validation is to determine that one or more of the performance characteristics of the method is satisfactory so that the analytical results obtained will be fit for their intended purpose.
Table 1 contains examples of what needs to be checked in the process of method validation and the procedures that could be followed. This list is not meant to be comprehensive.
| Performance parameter | Procedure that could be used |
|---|---|
| Limit of detection | Analyse blank sample matrix |
| Limit of quantitation | Analyse spiked (low level) blank sample matrix |
| Bias estimate | Use CRM (see Section 3.1) or spiked samples |
| Linearity range | Plot response of a set of solutions of known concentration against their concentration |
| Selectivity | Analyse CRM |
| Precision
| Repeat analysis of either a CRM or a homogenous sample |
| Ruggedness | Identify components of the procedure... |
