Discussion on the Iterative Process in Robust Algorithm A

Abstract

ISO 13528-2005 “Statistical methods for use in proficiency testing by interlaboratory comparisons” gives a robust method for calculating mean and standard deviation—Algorithm A. The algorithm gives a robust estimate of population mean and population standard deviation through continuous iterative calculations. In each step, the original value compared with the predetermined range ends. But in practical applications, some experimenters use the values from last iteration instead of the original value, and believe that it brings better results to the experiment. Through simulation, this paper shows that it is unreasonable to use the last iteration results, and the original data should be used for calculation.