Skip to content

Decision rule for statements of conformity

Every result of an analysis is subject to a measurement uncertainty, because the measured value only represents an estimate of the true (unknown) value due to limitations such as display accuracy or signal noise, or envi­ron­mental influences.

If a measured value is used to make a statement regarding compliance with a specified tolerance limit (statement of conformity), this decision can be made either with or without consideration of the measurement uncertainty.

ISO/IEC 17025 requires accredited testing laboratories to make statements of conformity based on a defined decision rule. This must be communicated to the customer and is applied unless the standard or technical specification on which the statement of conformity is based specifies a different decision rule.

We report an extended measurement uncertainty Uk=2 for a confidence interval of 95% for all measurement results. In some cases, this information can be provided on request. nless otherwise specified or agreed, we make statements of conformity on the basis of a simple decision rule. This means that the measurement uncertainty is not taken into account for the decision. Measurement results are rounded to the number of decimal places specified in the standard or technical specification before conformity is assessed.

As illustrated in the adjacent diagram, only the measured value, or the mean value of a multiple measurement, must lie within the acceptance interval (A‑B) to pass the conformity assessment. Measured values outside the acceptance interval fail the assessment (D‑E). If a measured value lies exactly on the tolerance limit (C), it passes the assessment if the tolerance limit represents an acceptable minimum or maximum value, or if the acceptance interval includes the tolerance limit (≥ or ≤). Otherwise, the measured value fails the conformity assessment. Without a clear indication of the role of the tolerance limit, measured values in case C pass the conformity assessment.

The probabilities for false positive declarations of conformity (error probability) arising from these decision rules are shown in a table for the 5 different cases A‑E. On test certificates with a statement of conformity, we note the applied decision rule in each case.

Case Decision

Error probability

A Pass < 2.5%
B Pass 2.5% to < 50%
C According to specification* 50%
D Fail 2.5% to < 50%
E Fail < 2.5%
* f the acceptance interval includes the tolerance limit (acceptable minimum or maximum value, or ≥ or ≤), the measured value passes the conformity assessment. Otherwise, it fails. If the rule of the tolerance limit is undefined, it passes.