Tiktok Linkedin Icon-instagram-1 X-twitter
FAQ
APPLY

Measurement Uncertainty in Laboratories: The Hidden Aspect of Reliable Results

What is Measurement Uncertainty?

A measurement result is never “single and absolute.” Every result reported by a laboratory actually represents a point with a probability distribution around the true value. In other words, a measurement result does not merely indicate a single point, but rather an interval defined together with its measurement uncertainty. Measurement uncertainty quantifies the width of this distribution—expressing, in numerical terms, how reliable the result truly is.

The ISO/IEC 17025 standard defines measurement uncertainty as “a fundamental indicator of a laboratory’s ability to produce accurate and comparable results.” The GUM (Guide to the Expression of Uncertainty in Measurement), on the other hand, establishes this concept on statistical foundations:

  • Type A evaluation of uncertainty: Statistical analysis derived from repeated measurement results.
  • Type B uncertainty: Non-statistical sources such as calibration certificates, manufacturer data, literature, and prior experience.

The Significance of Measurement Uncertainty

Uncertainty is not merely a technical calculation; it is a critical element that builds confidence in results. When a laboratory report is presented without an associated uncertainty value, it creates an illusion of certainty. In international trade, a reported value can determine acceptance or rejection decisions, and if uncertainty is not properly defined, products may be rejected. In medical laboratories, measurement uncertainty directly influences clinical decisions. A value located at the borderline of a reference interval may lead to misdiagnosis if measurement uncertainty is not taken into account

What are the impacts of your reported results?

Measurement uncertainty creates a chain reaction of effects:

  • In a calibration laboratory, an incorrectly calculated temperature uncertainty may disrupt the settings of ovens used in production.
  • In a food laboratory, when the uncertainty overlaps with the measurement of heavy metals, the decision for export approval or rejection may change.
  • In a medical laboratory, if uncertainty is ignored in a hormone test, a patient may be subjected to unnecessary treatment.

Therefore, measurement uncertainty is not merely a numerical add-on; it serves as the safety barrier of the result.

Why are nonconformities concerning measurement uncertainty so prevalent among laboratories?

In assessments, measurement uncertainty is almost always a “pain point.” The main reasons include:

  1. Use of generic templates – Copying a table from another source without adapting it to the laboratory’s own method.
  2. Omission of contributors – Overlooking critical parameters such as operator influence, environmental conditions, or repeat measurement data.
  3. Incorrect statistical approach – Miscalculating standard deviation or selecting the wrong type of distribution.
  4. Integration errors – Calculating uncertainty but failing to report it in the results in the “±” format.

The most common question asked by ISO/IEC 17025 assessors is:
👉 “You have calculated the uncertainty, but how have you integrated it into your results?”

How is it calculated?
Uncertainty evaluation is not merely a matter of solving an equation; it requires a systematic analysis.

Step-by-step approach:

    1. Define the measurement model
      • Formalize all parameters that influence the measurement result.
      • E.g., in a mass density measurement ρ=(m1–m0)/Vρ = 
  •  Identification of uncertainty contributors

  Type A evaluation: Standard deviation derived from repeated observations.

  Type B evaluation: Based on information from calibration certificates, instrument performance

  1. Convert the identified components into standard uncertainties
    • Normal distribution → u=σ/√n 
    • Rectangular distribution → u=a/√3
    • Triangular distribution → u=a/√6
  2. Combine

Method of the Root-Sum-of-Squares (RSS)
uc=√(u12+u22+…+un2)

  1. Calculate the expanded uncertainty
    • Select the confidence level (typically %95 → k=2).
    • U = k · uc
  2. Report
    • Measurement Result ± Uncertanity, e.g.:
      100,0 °C ± 0,5 °C (k=2, %95 confidence level).

Why are nonconformities concerning measurement uncertainty so prevalent among laboratories?

  • GUM (JCGM 100:2008) → Fundamental framework
  • EURACHEM/CITAC Guide: Measurement Uncertainty → Applied examples and practices
  • ILAC P14 → How to report uncertainty
  • EA-4/02 → A methodological framework for calibration and testing

Assessors typically ask the following question:
👉 “According to which reference did you calculate your uncertainty?”
these references strengthens the credibility of the laboratory..

How does it influence everyday life?

  • Measurement uncertainty is actually everywhere: 

    •  When your car shows 100 km/h, you may actually be driving within the range of 98–102 km/h

    (if U=2 km/h)

    • Your home blood pressure monitor may have an error margin of ±3 mmHg.

    A kitchen scale may display 1 kg of flour as anywhere between 995 g and 1005 g (if U = 5 g)


    While such variations may appear insignificant in everyday life, within the laboratory environment they can cause erroneous decisions, potentially leading to multi-million-dollar commercial losses or serious medical mismanagement.


    Author: Program Manager Dr. Serap Koktas Koca
    Author: Operations Manager Z. Naz Ayvaz