Post by Sizwe Mfanafuthi

As I was analysing the NLA-SA Round 2 proficiency testing results, I noticed that some laboratories obtained extremely high Z-scores. In one case, the Z-score was approximately 19, while the highest was as much as 21 045. This raised a question for me: why does the statistical program not identify and classify such extreme values as obvious blunders in order to protect the integrity of the assigned value and standard deviation? My concern is that, in some cases, these extreme results may influence the statistics sufficiently to cause another laboratory's result to fall just outside the acceptable Z-score limits. When conducting a root cause analysis and implementing corrective actions, one would naturally want to understand whether the issue originated within their own laboratory or whether it was influenced by the presence of extreme results submitted by other participants. I would appreciate your perspective on this matter. Alternatively, am I already sufficiently protected against such effects through the use of robust statistical methods, such as the H15 robust mean and associated robust standard deviation calculations? Or is there any other Technical Justification that i need to be aware of?, i am just a lame man navigating technical world.