Rationale

A significant rewrite of Section 12, Precision and Bias, is proposed. It is based on combined data from ILS Report #1863 and all CCRL PSP ASR samples 1 through 16. This represents an order of magnitude more data than is represented by either the ILS alone or the ILS combined with the original study (whose results were used for the current P&B statement). Based on this much larger data set, a more complete representation of the Coefficient of Variance (CoV) than a simple constant is justified, and was found to differ significantly from the current guidance around the levels of expansion in the test that are typically used for specification purposes. The consequences of this more complicated CoV relationship will be presented graphically in the P&B statement, and although more complicated that using a constant CoV, it will actually be simpler for the user of the method to comprehend the numerical precision of the result compared to interpreting (and subsequently calculating from) the verbal description of precision that currently exists in the method. These results, and this approach, was presented at the December 2024 meeting in Orlando to both a C09.50 Task Group and the C09.94 Evaluation of Data subcommittee.