Module 3 of 6 · about 68 min · Badge: Assigned Value
Assigned value and σpt
Five ways to set the assigned value, Algorithm A, and choosing σpt.
Start this module free
Sign in with your email to watch the narrated lessons, read the notes, use the tools and take the checks. It is free, and your progress and badges are saved.
What you will learn
5 lessons and a case study
- 1Sign in
Five ways to set the assigned value
Compare the five ways ISO 13528 gives for setting xpt, from formulation to participant consensus, by when each fits, its uncertainty, its traceability and its weak points.
Animated lesson · notes · check · about 11 minFree account - 2Sign in
Traceability, and when a consensus misleads
Judge what xpt is traceable to, recognise the four ways a participant consensus can mislead, and check a consensus against an independent reference value.
Animated lesson · notes · check · about 10 minFree account - 3Sign in
Robust statistics with Algorithm A
See why the plain mean and SD fail with outliers, calculate the median, MADe and nIQR, and run Algorithm A by hand and in the interactive tool.
Animated lesson · notes · check · about 13 minFree account - 4Sign in
Choosing σpt
Compare the five ways ISO 13528 gives for setting σpt, see what each does to the scores, and understand why a σpt taken from the round itself hides real performance.
Animated lesson · notes · check · about 11 minFree account - 5Sign in
Uncertainty of the assigned value and z′
Estimate u(xpt) for a participant consensus and for a certified value, apply the 0.3 σpt negligibility rule, and decide when to score with z′ instead of z.
Animated lesson · notes · check · about 10 minFree account - CSign in
Case: setting xpt and σpt for a small trace-metals round
Choose xpt and σpt for a 14-laboratory blood lead round with one outlier and a CRM for one level, estimate u(xpt) and decide between z and z′.
Case study · about 9 minFree account