Calibration in routine analysis

The previous section described how the conditions for postulating a linear regression function as a calibration function are checked, and how the parameters of an analytical method are determined. In routine analysis, the scientist now has the task of proving that the procedural characteristic data which he has obtained are not significantly different from the prescribed data. For this purpose the information values iy[) and the procedural standard deviation Sxo(R) are determined from 10 equally spaced concentration levels (xj). 

A simple variance test (F test) shows whether the procedural standard deviation Sxo(R) determined in routine analysis differs significantly from the procedural standard deviation Sxo prescribed in the standard. 

If the first calibration function appears non-linear, the simplest solution is to attempt to position the working range so that the calibration function is linear within the working range. 

If this is not possible, parameters of the calibration function aQ, aj[ and a2 can be determined according to the rules of polynomial approximation of the 2nd degree  

Only the sensitivity in the middle of the working range can be given as a procedural parameter (x)  

---