IJCA Vol 5 Issue 1 - Flipbook - Page 44
The International Journal of Conformity Assessment
6. Performance Evaluation and Statistical
Considerations in PT and ILC
Performance evaluation is a defining element
of proficiency testing and a primary means by
which PT and ILC participation provides objective
evidence of technical competence. For calibration
laboratories, understanding how performance is
evaluated — and whether the evaluation method is
technically appropriate — is essential for interpreting
results and supporting accreditation decisions.
ISO/IEC 17043:2023 reinforces the importance of
transparent and technically justified performance
evaluation by aligning PT provider requirements
with ISO 13528:2022, which specifies statistical
methods for proficiency testing by interlaboratory
comparison. Although ISO 13528 is primarily
directed at PT providers, its principles are
directly relevant to participant laboratories, as
PT outcomes are evaluated and accepted on the
basis of these methods.
6.1 Purpose of Performance Evaluation in
Proficiency Testing
The purpose of performance evaluation is
to assess participant results relative to an
assigned value using predefined criteria, enabling
meaningful comparison among laboratories.
Through this process, PT schemes convert
reported measurement results into standardized
indicators of performance that can be interpreted
by laboratories and accreditation bodies.
For calibration laboratories, performance
evaluation supports validation of measurement
accuracy and claimed uncertainty. Unsatisfactory
results may indicate issues related to
measurement procedures, reference standards,
environmental conditions, or uncertainty
estimation and therefore require investigation and
documented follow-up. Even satisfactory results
contribute evidence supporting the reliability and
consistency of measurement processes.
6.2 Common Performance Metrics Used in
PT and ILC
ISO 13528 describes several statistical metrics
commonly used to evaluate participant
performance, including z‑scores, z′‑scores, and
En values. Each metric is based on different
assumptions regarding uncertainty and reference
values, and their suitability depends on the context
of the comparison.
The z-score expresses the difference between
a participant’s result and the assigned value,
44
2026 | Volume 5, Issue 1
normalized by a standard deviation for proficiency
assessment. It is widely used in testing contexts
where the uncertainty of the assigned value is
considered negligible. The z′‑score extends this
approach by incorporating the uncertainty of the
assigned value, providing a more conservative
assessment when reference uncertainty cannot be
ignored.
In calibration contexts, the En value is widely
regarded as the most appropriate performance
metric because it explicitly incorporates both
the participant’s reported expanded uncertainty
and the uncertainty of the assigned value. By
evaluating metrological compatibility within
combined uncertainties, the En value provides
insight into whether a laboratory’s claimed
uncertainty is supported by its performance.
6.3 Selection of Appropriate Metrics for
Calibration Laboratories
ISO 13528 emphasizes that statistical evaluation
methods must be appropriate for the objectives
of the comparison and the characteristics of
the data. This principle is particularly important
for calibration laboratories, where uncertainty
is a fundamental component of demonstrating
competence.
Use of z-based metrics alone may be
misleading when uncertainty plays a central
role in performance assessment. In such
cases, evaluation methods that do not account
for participant uncertainty may fail to reflect
whether reported results are metrologically
compatible with the assigned value. For this
reason, calibration laboratories should understand
which evaluation metric is applied and assess
its suitability for their scope and claimed
measurement capabilities.
Even when participating in PT schemes operated
by accredited providers, laboratories retain
responsibility for reviewing evaluation methods
and documenting their assessment of technical
appropriateness. Accreditation of the PT provider
does not remove this responsibility.
6.4 Assigned Values, Uncertainty, and Sample
Size Considerations
The determination of the assigned value is a
critical component of performance evaluation.
ISO 13528 describes multiple approaches for
establishing assigned values, including the use
of reference laboratories, certified reference
materials, or robust statistical methods applied
to participant results. The suitability of each
