Overview
ISO 5725-5:2025 - part of the ISO 5725 series on accuracy (trueness and precision) of measurement methods and results - specifies alternative robust methods for determining the precision of a standard measurement method. The second edition (2025) focuses on analytical procedures that avoid conventional outlier tests by using robust statistical estimators. It provides algorithms and computational methods to derive repeatability and reproducibility estimates from interlaboratory precision experiments without excluding data items based on outlier decisions.
Key topics
- Robust data analysis approaches: detailed presentation of Algorithm A and Algorithm S for robust calculation of dispersion measures.
- Computationally intensive estimators: introduction and use of the Q method and the Hampel estimator (Q/Hampel method) for improved breakdown properties and resistant estimates of means and standard deviations.
- Design coverage: formulae and procedures for common experimental layouts, including uniform‑level designs, split‑level designs, and experiments on heterogeneous materials; treatment of staggered/nested designs with two factors.
- Outputs provided: robust estimates of repeatability standard deviation, reproducibility standard deviation, intermediate standard deviation, and robust mean values without reliance on outlier exclusion.
- Support material: normative references (ISO 3534-1, ISO 5725-1), derivations, worked examples and implementation guidance in annexes.
Practical applications
- Laboratories and interlaboratory study organizers can use ISO 5725-5:2025 to:
- Produce robust precision estimates that are less sensitive to single extreme results or borderline outliers.
- Analyse data from heterogeneous samples where identical sample preparation is impractical.
- Reduce subjectivity in the treatment of outlying data during proficiency testing, method validation, or collaborative studies.
- Statisticians, metrology experts, accreditation bodies, and measurement method developers will find the standard valuable for:
- Implementing reproducible procedures for calculating repeatability and reproducibility.
- Choosing appropriate robust algorithms (Algorithm A/S, Q/Hampel) tailored to the experiment design.
- Documenting method precision in compliance reports, technical dossiers, and standard operating procedures.
Related standards
- ISO 5725-1: General principles and definitions (required reading)
- ISO 5725-2: Basic methods for estimating precision (context/comparison)
- ISO 5725-3: Designs for heterogeneous materials (related experimental designs)
- ISO 3534-1: Statistical vocabulary and symbols
ISO 5725-5:2025 is essential when you need reliable, objective precision estimates from interlaboratory data using robust statistical methods-especially in situations where outlier exclusion would introduce subjectivity or distort results.