Measurement Uncertainty Calculator

Build a complete uncertainty budget with the GUM method (JCGM 100, ISO/IEC Guide 98-3). Add Type A and Type B components and get the combined standard uncertainty and the expanded uncertainty for ISO/IEC 17025 reports. Free and in your browser, with no spreadsheet templates to buy.

STEP 1 — ADD AN UNCERTAINTY COMPONENT

ci = 1 in most cases. Use a different value if the input quantity affects the result through a different factor.

TYPE A — enter the repeated readings

u = s/√n (experimental standard deviation of the mean)

STEP 2 — COMPONENTS ADDED

STEP 3 — COVERAGE FACTOR

What is the GUM method?

The GUM (Guide to the Expression of Uncertainty in Measurement, published as JCGM 100:2008 and as ISO/IEC Guide 98-3) is the international document that sets out how to evaluate and express measurement uncertainty. It is the reference for laboratories accredited to ISO/IEC 17025, which must evaluate the uncertainty of their results.

The method has four steps: list the sources of uncertainty, evaluate the standard uncertainty u(xi) of each one, combine them into the combined standard uncertainty uc = √Σ(ci × u(xi))², and multiply by a coverage factor k to obtain the expanded uncertainty U = k × uc. The table that lists every component with its contribution is the uncertainty budget.

Type A and Type B evaluation

Type A evaluation uses statistics: you take several repeated measurements and calculate their standard deviation. The standard uncertainty is u = s/√n, the experimental standard deviation of the mean.

Type B evaluation uses any other information: calibration certificates, manufacturer specifications, published data or the laboratory's experience. The standard uncertainty is the half-width of the interval divided by the divisor of the assumed distribution.

Distributions most often used in Type B evaluation

Rectangular (divisor √3): when you only know that the value lies within an interval, with no further information. Example: the resolution of a digital instrument.

Normal (divisor 1): when the certificate already gives the standard uncertainty. If it gives an expanded uncertainty with k = 2, divide it by 2 first.

Triangular (divisor √6): when values near the center are more likely than those at the limits. Example: the tolerance of a volumetric flask or the temperature setting of an oven.

Frequently asked questions

How many components should I include?

Every component that makes a non-negligible contribution to the final result. At a minimum: instrument repeatability (Type A), calibration of the instrument or reference material (Type B) and instrument resolution (Type B). More complex procedures add others, such as the purity of the standard, solution preparation, temperature and operator effects.

Which coverage factor (k) should I use for ISO/IEC 17025?

Most laboratories and accreditation bodies, following the ILAC P14 policy, expect k = 2, which gives a level of confidence of approximately 95%. Use it unless your method or your client specifies a different value.

Is the sensitivity coefficient (ci) always 1?

Not always. The sensitivity coefficient ci describes how strongly an input quantity affects the final result. If the result is the measured quantity itself, ci = 1. If there is a dilution factor of 10, ci = 10 for the components that come before the dilution. If the quantity enters a nonlinear equation (pH, for example), ci is the partial derivative of that equation with respect to the input quantity.

Related tools

  • LOD and LOQ Calculator: limit of detection and limit of quantitation by three methods (blank, calibration curve and signal-to-noise), for method validation under ISO/IEC 17025.

This calculator is also available in Spanish.