Overview
ISO 5725-5:1998 is Part 5 of the ISO 5725 series on the accuracy (trueness and precision) of measurement methods and results. It specifies alternative methods to the basic interlaboratory design for determining the precision (repeatability and reproducibility) of a standard measurement method. The part complements ISO 5725-1 and ISO 5725-2 by providing experimental designs and robust statistical approaches for situations in which the basic uniform‑level design is not optimal.
Key topics and requirements
- Alternative experimental designs
- Split‑level design: Each laboratory receives two similar (but not identical) samples at each level to reduce bias caused when operators let one measurement influence another. Includes layout, statistical model, data analysis, outlier scrutiny and reporting.
- Design for heterogeneous materials: Addresses situations where identical samples cannot be prepared. Provides layout, statistical modelling and methods to separate within‑sample and between‑sample variability so reproducibility estimates are not inflated.
- Robust data analysis
- Describes robust methods (including Algorithm A and Algorithm S) to estimate repeatability and reproducibility without relying on outlier exclusion tests - reducing analyst judgement effects.
- Provides formulae and worked examples for applying robust analysis at different experiment levels.
- Statistical detail
- Guidance on planning experiments (sample sizes, number of laboratories), statistical models, scrutiny for consistency and outliers, and reporting of results.
- Normative references to ISO 3534 (statistics vocabulary) and other ISO 5725 parts.
- Annexes
- Symbols/abbreviations (normative) and informative derivations, formulae and bibliography.
Practical applications and users
Who benefits:
- Metrology and testing laboratories conducting interlaboratory studies to quantify method precision.
- Analytical chemistry and quality-control teams facing operator bias risks or working with heterogeneous materials (e.g., soils, foods, composite products).
- Standards developers and method authors who need validated precision estimates for standard measurement methods.
- Statisticians and quality managers implementing robust analysis when outliers or irregular data threaten standard precision estimates.
Typical uses:
- Reducing operator-induced bias in precision studies (use split‑level design).
- Separating sample heterogeneity from method variability to obtain realistic reproducibility estimates.
- Applying robust statistical procedures to produce repeatability/reproducibility estimates insensitive to outliers.
Related standards
- ISO 5725-1: General principles and definitions (trueness, precision)
- ISO 5725-2: Basic method for repeatability and reproducibility
- ISO 3534-1 / ISO 3534-3: Statistical vocabulary and experimental design
Keywords: ISO 5725-5, accuracy, trueness and precision, split-level design, heterogeneous material, repeatability, reproducibility, robust methods, interlaboratory experiment.