Abstract
A procedure for the evaluation of results of interlaboratory comparisons of trace multi-component determinations is presented. The distribution of interlaboratory comparison data may differ significantly from the normal distribution and, therefore, various types of median and modal values approximate their central values better than the mean. The simple median and its confidence interval, calculated on the basis of a nonparametric distribution after the rejection of outlying results, is shown to be the optimal indicator of the central values. The total procedure consists of three parts: statistical evaluation of the data, classification of the results according to a special set of analytical criteria, and graphic presentation of the data. The application of the results of such comparisons for certification of analytical reference materials is discussed.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
