Fundamentals of Statistics contains material of various lectures and courses of H. Lohninger on statistics, data analysis and chemometrics......click here for more.


Determination Limit

Every measurement (e.g. the determination of the sodium content of a solution) is subject to an uncertainty that can arise from various sources and that is indicated by the confidence interval of the analytical result (this is also referred to as analytical precision). Since the analytical precision is roughly constant over the entire measuring range (due to the required homoscedastic residuals in the calibration), the relative precision naturally deteriorates the lower the value to be measured is. In general, a relative (im)precision of 33% is accepted, which leads to a lower limit xB of

xB = 3 * confidence interval

This limit is called the limit of quantification. The limit of quantification is the concentration from which the measurement meets a specified requirement for precision. In general:

decision limit < detection capability < limit of quantification

As a rule of thumb for checking the determined limits, one can assume that the detection capability is about twice as high as the decision limit, and the limit of quantification at 33% relative precision at four times the decision limit. However, the exact relationships are influenced by a large number of parameters; a more in-depth discussion can be found at Funk et. al.