12. Precision and Bias
12.1 Precision - The precision of this test method as determined by a statistical examination of interlaboratory test results is as follows:

12.1.1 Repeatability - The difference between successive results obtained by the same operator with the same apparatus under constant operating conditions on identical test materials would, in the long run, in the normal and the correct operation of the test method, exceed the following values in only one case in twenty.


where:
X = the salt concentration in mass % (as NaCl).

12.1.2 Reproducibility - The difference between two single and independent results, obtained by different operators working in different laboratories on identical test materials, would, in the long run, and in the normal and the correct operation of the test method, exceed the following values in only one case in twenty.


where:
X = the salt concentration in mass % (as NaCl).

NOTE 7 - The precision data from the interlaboratory program were obtained on three different crude oils ranging in density between 825 and 950 kg/m3, spiked with varying amounts of (artificial) sea water and formation water where the salt content (as NaCl) varied from 0.0005 to 0.15 mass %. Statistical evaluation could not detect any dependency of precision on crude oil density or water type, except for reproducibility. The reproducibility variance (power transform) proved to be inversely related to crude oil density and varied from 0.49 for the heavy crude to 1.35 for the light crude. The above quoted reproducibility represents the overall variance of 0.99.

12.2 Bias - The procedure in Test Method D 6470 has no bias since salt content is defined only in terms of this test method and certified reference materials are unavailable. However, since the samples from the interlaboratory study where neat desalted crudes, spiked with known quantities of salt (as sea water and formation water), bias might be defined as percent recovery of halide added. Over the range 0.0005 to 0.0400 mass % salt added, the recovery proved to be constant and averaged 97 %. Over the range 0.0400 to 0.1500, the recovery proved to be a function of concentration and gradually decreased from 97 % at 0.04 mass % to 88 % at the 0.15 mass % level.

13. Keywords
13.1 crude oil; extraction; potentiometric; salt