What does k=2 mean on a calibration certificate?

The letter k is the coverage factor. k=2 means the stated uncertainty covers roughly a 95% confidence level for a normal distribution — so there is about a 95% probability the true value lies within the quoted range. It is the most common convention on accredited calibration certificates.

Why must a calibration certificate state measurement uncertainty?

Because a result without uncertainty cannot be used to make a reliable pass/fail decision. Regulators and ISO/IEC 17025 expect uncertainty to be reported so you can judge whether a reading is genuinely within limits once its uncertainty is taken into account. A certificate with no uncertainty is of limited compliance value.

How does uncertainty affect a pass/fail decision?

If a reading plus or minus its uncertainty could cross your acceptance limit, the result is not a clear pass. Good practice is to apply a guard band — effectively tightening the limit by the uncertainty — so you only accept results that are inside the limit with margin. Ignoring uncertainty risks accepting out-of-tolerance instruments.

Measurement Uncertainty Explained (What k=2 Means)

Q: What is measurement uncertainty?

Measurement uncertainty is the range within which the true value of a measurement is expected to lie. No measurement is perfect, so a calibrated result is only meaningful when stated with its uncertainty — for example 25.00°C ± 0.05°C. It quantifies how much you can trust the number.

Measurement uncertainty is the honest answer to "how much can I trust this number?" Every measurement has it, which is why a proper calibration certificate never just says "25.00°C" — it says "25.00°C ± 0.05°C (k=2)". Understanding that little suffix is the difference between a defensible compliance decision and a lucky guess.

In shortUncertainty is the range the true value is expected to lie within. The coverage factor k=2 means about 95% confidence. Certificates must state uncertainty because you cannot make a sound pass/fail decision without it — and near a limit, you should apply a guard band so you only accept results that are inside the limit with margin.

Why every measurement has uncertainty

No instrument, reference or method is perfect. The sensor, the reference standard, environmental conditions and the process all contribute small errors. Uncertainty gathers all of those contributions into a single number that tells you how far the reading might reasonably be from the truth. Far from being a weakness, stating uncertainty is a mark of rigour.

The coverage factor: what k=2 means

Uncertainty is quoted at a confidence level, set by the coverage factor k. For a normal distribution, k=2 corresponds to roughly 95% confidence — there is about a 95% chance the true value lies within the stated range. (k=1 is ~68%, k=3 is ~99.7%.) Accredited certificates almost always use k=2, so "± 0.05°C (k=2)" means you can be about 95% sure the true value is within 0.05°C of the reading.

Measurement uncertainty on a calibration certificate

Why "±0.5°C accuracy" isn't the full story

A logger's headline accuracy spec is a manufacturer's claim about the device. Calibrated uncertainty is the tested, traceable statement of how much to trust your instrument as used. When you make compliance decisions, it is the calibrated uncertainty — not the datasheet accuracy — that matters.

Using uncertainty in decisions: the guard band

Here is where it becomes practical. Suppose your limit is 8.0°C and your logger reads 7.98°C with an uncertainty of ±0.1°C. The true value could be anywhere from 7.88 to 8.08°C — which means it might already be over the limit. A clear pass requires the reading plus its uncertainty to stay inside the limit. Applying a guard band (tightening the acceptance limit by the uncertainty) ensures you only accept results that are genuinely in tolerance.

Test uncertainty ratio (TUR)A useful rule of thumb: the reference used to calibrate should be several times more accurate than the item under test — a TUR of 4:1 is a common target. If your reference is barely better than the instrument you are checking, the calibration tells you little.

Frequently asked questions

What is measurement uncertainty?

The range within which the true value is expected to lie — reported alongside a result, e.g. 25.00°C ± 0.05°C.

What does k=2 mean?

The coverage factor for ~95% confidence on a normal distribution — the standard convention on accredited certificates.

Why must a certificate state uncertainty?

Because you cannot make a reliable pass/fail decision without it; ISO 17025 and regulators expect it reported.

How does it affect pass/fail?

If reading ± uncertainty could cross the limit, it is not a clear pass — apply a guard band to accept only in-tolerance results with margin.

Key takeaways

Every measurement has uncertainty; a result without it is incomplete.
k=2 means about 95% confidence — the accredited-certificate standard.
Use calibrated uncertainty, not datasheet accuracy, for decisions.
Apply a guard band near limits so you never accept out-of-tolerance instruments.

Related guides

Get new guides by email

Practical monitoring, mapping and calibration guides, straight to your inbox. No spam — unsubscribe anytime.