ISO 25337:2010

Overview

ISO 25337:2010 - Plastics - Production quality control - Statistical method for using single measurements provides a validated statistical framework (the single-measurement model, SD model) for using single test measurements in production quality control. The standard lets producers reduce replicate measurements or repeated operations specified in test methods - where justified by statistical analysis - to save time and cost while preserving control over product quality. It applies only where test results follow a normal (Gaussian) distribution and is intended for use at a producer’s manufacturing site.

Key topics and requirements

• Single‑measurement model (SD model): Describes how single measurements can replace replicates for production control when supported by statistical evidence.
• Production limits: Establishes lower (L_PL) and upper (U_PL) production limits for a stable production period and shows how to account for test method variability when defining these limits.
• Combined variability: Emphasizes the need to consider both production-process variability and test-method variability (combined standard deviation s_P&T).
• Precision requirements: Recommends that test method precision be at most 30% (preferably less than 10%) of the combined variability to reliably detect process variation.
• Within‑laboratory reproducibility (s_RLab): Requires estimates of within‑laboratory reproducibility; this reproducibility component is added to production limits to define a zone where results may fall due to testing/process variation.
• Statistical parameters: Uses familiar concepts such as repeatability/reproducibility (ISO 5725), k‑factors from the standard normal distribution (e.g., ±3s → ≈99.73% interval) for setting warning/acceptance limits.
• Scope limits and documentation: Only applicable to normally distributed numerical results; when applied alongside existing test method standards the combination must be clearly documented in product/process specifications and contracts.
• Need for statistical support: The involvement of a statistician is strongly recommended and sometimes indispensable for RRI design, analysis and limit setting.

Practical applications and users

• Who uses it: Project managers, heads of laboratories, production managers, quality engineers in the plastics industry - and other industries using ISO test methods - plus consulting statisticians.
• Typical benefits:
  • Reduce laboratory workload and response time by justifying single measurements.
  • Lower production and testing costs by reducing replicates or repeated operations (e.g., drying cycles).
  • Harmonize production limits with test-method precision to avoid rejecting conforming product or accepting nonconforming product.
  • Provide a defensible statistical basis for internal QC procedures, supplier agreements and production specifications.
• Caveats: Not intended for publication of marketing data or replacing full inter-laboratory precision studies; reductions in replicates should follow validated RRI work.

Related standards

• ISO 5725 (Accuracy - trueness and precision): normative reference for repeatability and reproducibility investigations used in ISO 25337:2010.