A method of evaluating measurement uncertainty using means other than statistical analysis of observations, including manufacturer specifications, calibration certificates, published data, and scientific judgment.
Type B uncertainty evaluation uses any relevant information other than repeated measurement data to quantify an uncertainty component. Sources include manufacturer accuracy specifications, calibration certificate data for reference standards, handbook values for material properties (thermal expansion coefficients), resolution of digital displays, environmental measurement data, and engineering judgment based on experience.
Type B evaluations require assuming a probability distribution for each uncertainty source. Common distributions include rectangular (uniform) for situations where only upper and lower bounds are known (e.g., instrument specifications, resolution), normal (Gaussian) for calibration certificate uncertainties, triangular for quantities known to have a most likely value near the center of the bounds, and U-shaped for quantities equally likely to be at either extreme. The standard uncertainty is then derived from the distribution parameters.
For calibration uncertainty budgets, Type B components often dominate the combined uncertainty. They include the reference standard uncertainty (from its calibration certificate), resolution of the unit under test, thermal expansion effects, and environmental corrections. Properly evaluating Type B uncertainties requires understanding the source information and selecting appropriate probability distributions. While Type B evaluation involves judgment, it should be based on documented, defensible reasoning, not arbitrary guesses. The GUM and ISO 17025 treat Type A and Type B contributions equally in the combination process.
In aerospace calibration labs, Type B uncertainty is critical when calibrating torque wrenches using manufacturer specifications for the calibration standard's accuracy (±0.25% of reading). When the reference standard's certificate provides expanded uncertainty of 0.15% (k=2), this becomes Type B input data requiring division by coverage factor to obtain standard uncertainty. Medical device manufacturers face Type B evaluation when calibrating temperature chambers for sterilization validation. The chamber's display resolution (±0.1°C) and manufacturer's accuracy specification (±0.2°C) are Type B sources requiring rectangular distribution assumptions. A common audit finding occurs when labs incorrectly treat calibration certificate uncertainties as Type A data, failing to recognize these are already evaluated uncertainties requiring proper Type B treatment. Another frequent error involves ignoring environmental conditions like temperature coefficients from manufacturer specifications. For pressure calibrators in defense applications, ignoring the reference standard's temperature coefficient (0.001%/°C) as a Type B source can underestimate total uncertainty by 20-30%, leading to failed proficiency testing or invalid measurement decisions for critical flight components.
ISO/IEC 17025:2017 Section 7.6.2 explicitly requires evaluation of measurement uncertainty including Type B methods per GUM principles. AS9100D Section 7.1.5.2 mandates aerospace calibration labs demonstrate uncertainty evaluation competence. ISO/IEC Guide 98-3 (GUM) Sections 4.3.1 and 4.3.2 define Type B evaluation methodology and acceptable information sources including calibration certificates, manufacturer specifications, and handbook data. ISO 13485:2016 Section 7.6 requires medical device manufacturers to validate measurement processes with appropriate uncertainty analysis. ANSI/NCSL Z540.3-2006 Section 11.2.2 specifically addresses Type B evaluation requirements for calibration laboratories. ILAC-P14:07/2013 Section 7.6 provides policy guidance on uncertainty evaluation including Type B sources. Auditors specifically examine: (1) documentation of Type B information sources, (2) proper application of probability distributions (rectangular, triangular, normal), (3) correlation considerations between uncertainty components, (4) coverage factor applications, and (5) validation that uncertainty budgets include all significant Type B sources rather than solely relying on statistical analysis of repeated measurements.
CalibrationOS captures Type B uncertainty through the Uncertainty Budget module, where technicians input manufacturer specifications, environmental coefficients, and resolution limits for each measurement parameter. The system automatically applies appropriate probability distributions (rectangular for specifications, normal for calibration certificates) and calculates standard uncertainties. During certificate generation, the software combines Type A and Type B components using root-sum-squares methodology per GUM requirements. The Audit Trail feature documents all Type B source justifications and calculations for ISO/IEC 17025 compliance. CalibrationOS generates uncertainty budget reports showing individual Type B contributions, enabling labs to demonstrate comprehensive uncertainty evaluation during external audits. The system validates that expanded uncertainties from calibration certificates are properly converted to standard uncertainties before combination with other sources, preventing common Type B evaluation errors that result in audit findings.
Type B uncertainty is evaluated using information other than repeated measurements — such as manufacturer specifications, calibration certificates, published data, or scientific judgment. It requires assuming a probability distribution for each source.
Use rectangular (uniform) distribution when only upper and lower bounds are known, normal distribution for calibration certificate uncertainties, and triangular distribution when the most likely value is known to be near the center of the range.
This article is licensed CC BY-SA 4.0. Share, adapt, and reuse with attribution to calibrationos.com/glossary/type-b-uncertainty.
Industry benchmarks, best practices, and calibration tips — delivered to your inbox.
GUM-compliant expanded uncertainty with Welch-Satterthwaite degrees of freedom. No signup required.
Open the Calculator