Performance Criteria

DOE-STD-1112-98

5.2.3 Performance Criteria. The performance criteria used by the testing laboratory are relative

bias and relative precision. See ANSI N13.30-1996 sections 3.42, 3.43, 4.32 and 4.33 for

greater detail on the bias and precision statistics development.

5.2.3.1 A relative bias statistic is defined in this document for the purposes of performance

testing of a finite number of measurements in each category of analysis. The relative bias is

individually calculated for each radionuclide in the category being tested. The individual

relative bias statistic for the ith measurement in a category with respect to the correct value is

defined as

Bri ' (Ai & Aai) / Aai

(6.1)

where Ai is the value of the ith measurement in a category being tested, not necessarily a

replicate, but possibly a different quantity of spike for each measurement, and Aai is the actual

quantity in the test sample, as defined by the spike.

5.2.3.2 In order to avoid the expense of a large number of replicates at each radioactivity level

in each category, the relative bias (which may be obtained at different quantity levels) for that

test category is calculated from the individual relative biases (Bri). The test category relative

bias (Br) is defined as:

Bri

(6.2)

i'1

Br ' Bri '

where N is the number of test samples measured by an individual service laboratory in a given

test category. The sample size N shall be at least five. For performance testing purposes, Br

shall be within -0.25 to + 0.50 when Aai is at or above the MTL to be considered acceptable.

5.2.3.3 The relative precision of the measurement process, selected for the purposes of this

document, is the relative dispersion of the values of Bri from its mean Br. The relative

precision (SB) is defined as