Show solutionHide solution
Question 9
Separate calibration bias from repeatability
Assume the reference value is reliable.
Diagram shown in the original SVG worksheet.
Calculate mean mass error.
Express the mean error in ppm.
Why does the small replicate spread not remove the need for calibration?
Question 9 - Solution
Separate calibration bias from repeatability
Diagram shown in the original SVG worksheet.
(a)
The mean error is +0.0020 u.
(b)
One million times 0.0020/400 = +5.0 ppm.
(c)
The replicates agree closely with each other but are biased relative to the reference. Precision is not the same as accuracy.