Overview
ISO 7347:1987 - "Ferroalloys - Experimental methods for checking the bias of sampling and sample preparation" specifies statistical experimental procedures to check whether a sampling or sample‑preparation method introduces bias when determining quality characteristics of ferroalloys. The standard defines how to compare a method under test (Method B) with a reference method (Method A) using repeated experiments and statistical significance tests at the 5% level (two‑sided t‑test). It is intended to be used together with ISO 3713, ISO 7087 and ISO 7373.
Key topics and technical requirements
- Scope: Applies to ferroalloys covered by relevant ISO sampling standards; establishes procedures to detect systematic error (bias) introduced by sampling or sample preparation.
- Experimental design:
- Repeat experiments at least 10 times (for example on 10 consignments, parts of consignments, gross samples or sub‑samples).
- Take two gross samples from the same consignment: one by Method A (reference) and one by Method B (method to be checked).
- When checking sample preparation, prepare separate test samples A and B from the same gross sample or sub‑sample.
- Use pairing (matched increments from the same point) where technically possible - this increases statistical sensitivity.
- Data recording: Use standard data logs (examples provided for paired and unpaired increments).
- Statistical analysis:
- Paired data: Apply the paired t‑test (Annex A) to the differences between matched measurements.
- Unpaired data: First apply the F‑test (Annex B) to check equality of variances; if passed, proceed to the unpaired t‑test (Annex C).
- Decision rule: compare observed t or F with critical values (tables provided); significance set at 5% (two‑sided for t).
- Outcome interpretation:
- If difference is statistically insignificant, Method B may be adopted.
- If statistically significant but practically negligible, Method B may still be used by agreement.
- If significant and practically important, reject Method B or repeat experiments.
Applications and users
- Who uses it: metallurgical laboratories, quality assurance teams, ferroalloy producers, material inspection bodies, buyers and standardization/technical committees.
- Practical uses:
- Validate new sampling procedures or alternative sample‑preparation devices.
- Demonstrate equivalence between routine methods and reference methods for contract compliance.
- Support quality control, dispute resolution, and supplier qualification in ferroalloy trade.
Related standards
- ISO 3713 - Ferroalloys: sampling and sample preparation general rules
- ISO 7087 - Experimental methods for evaluation of quality variation and precision of sampling
- ISO 7373 - Experimental methods for checking the precision of sample division
Annexes A–D provide calculation methods for paired/unpaired t‑tests, the F‑test, and a worked example to guide implementation. Keywords: ISO 7347:1987, ferroalloys sampling, sampling bias, sample preparation, paired t‑test, F‑test, statistical validation.