How does the QDE Toolkit fit with the current drive for “simplicity” in the analysis of comparisons?
2002-06-26 (Edited 2002-08-14.) We think that it is quite easy to use the QDE Toolkit, yet we would be the first to admit that full mastery of the QDE Toolkit is not “simple” – we would not have written a 185-page user’s guide to something that was trivially simple to master fully! There is a trade-off between simplicity and the rigor-and-accuracy of a statistical treatment. If a factor-of-two accuracy were thought to be sufficient for establishing equivalence, then everything could be much simpler. If Lab A is equivalent to the KCRV, and Lab B is equivalent to the KCRV, then within a factor of two Lab A would be equivalent to Lab B. Unfortunately, if one wants accuracy better than a factor of two, this simple picture breaks down. The ISO Guide to the Expression of Uncertianty in Measurement emphasizes the importance of not unnecessarily inflating uncertainties and thereby “cheating” clients of the precision for which they have paid. Fortunately, only rarely will one really want better
