What your En number actually says — and what it does not

Proficiency testing results for calibration laboratories usually arrive as an En number: the difference between your result and the reference value, divided by the root-sum-square of the two expanded uncertainties. |En| ≤ 1 is scored satisfactory; beyond 1, your result and the reference disagree by more than your combined claimed uncertainties can explain.

Two things make En sharper than the z-scores common in testing PT schemes. First, it judges your uncertainty claim, not just your value: a laboratory that measures well but understates its uncertainty will fail En exactly as it should. Second, it is symmetric evidence — an En of 0.05 with an enormous claimed uncertainty is a weak performance dressed as a safe one, which is why assessors read the uncertainty column alongside the score.

The habit that separates strong laboratories: treat |En| between about 0.7 and 1.0 as a finding even though the scheme calls it a pass. Something — an unstable standard, an optimistic budget line, a procedural drift — consumed most of your uncertainty margin, and ISO 13528's statistics cannot tell you which. Investigating a 0.9 today is dramatically cheaper than investigating a 1.3 after your accreditation body has seen it.