The Acceptance Criteria page defines, for each method–element combination, the acceptable limits, the measurement uncertainty and the rule by which a statement of conformity is reached (EN ISO/IEC 17025:2017 §7.6, §7.7, §7.8.6).
Found under: Quality Control → Acceptance Criteria.
| Button | Action |
|---|---|
| Add — Create a new record | |
| Columns — Show/hide table columns | |
| Filters — Open/close the filter panel. A red dot indicates active filters | |
| Export CSV — Download table data as a CSV file | |
| Orientation — Toggle between portrait and landscape printing | |
| Print — Generate a PDF of the current view | |
| Help — Open the help dialog | |
| Search — Quick search through table data |
Clicking the icon on each row shows:
| Field | Description |
|---|---|
| Document No. | Unique criteria code (for example AC-YYYY-0001). Leave blank for auto-numbering. |
| Lab Method | The analysis method to which the criteria apply |
| Element | The chemical element or parameter being evaluated |
| Lower Limit | The minimum acceptable measurement value |
| Upper Limit | The maximum acceptable measurement value |
| Uncertainty | The expanded measurement uncertainty (±U) |
| Calculation Formula | Mathematical formula (optional, hidden column); a preview of the result shows below the field |
| Decision Rule | The rule behind the statement of conformity — see below |
| Decision Rule Notes | Rounding, averaging of replicates, outlier handling (Eurachem MUC §3) |
| Specification / standard | §7.8.6.2(b) — where the limits come from; printed beside the conformity statement |
| Notes | Additional remarks or instructions |
| Rule | When |
|---|---|
| Simple Acceptance | The value is judged as it stands; needs an uncertainty pre-condition, see below |
| Guard Bands | The limit is shifted by a guard band so the risk of false acceptance is contained |
| ILAC-G8 | The rule from the ILAC-G8 guidance |
| Customer Specified | The rule agreed on the order or in the standing terms |
| Qualitative Match | §7.8.2 — for examinations with no numeric result |
Simple acceptance is not "no rule". A value landing exactly on the limit carries a 50 % false-decision risk unless we know how narrow the uncertainty is relative to the specification interval. So the criterion accepts two bounds, either or both:
| Field | Meaning |
|---|---|
| Min. capability index C95 | C95 = (TU − TL) / (2·U95) — how many times the uncertainty "fits" into the specification interval |
| Max. expanded uncertainty Umax | An absolute cap on U. Target per Eurachem STMU §5.1.2: (upper − lower limit) / 8, or the proficiency-testing scheme's σ |
Under the guard-bands rule, the band comes from two fields:
| Field | Meaning |
|---|---|
| Coverage factor k | The k of the stored expanded U (§7.6) |
| Max. false-acceptance probability PFAmax (%) | How much false-acceptance risk you accept |
The band becomes w = kw·u, with u = U/k and kw from the standard normal distribution for PFAmax (UKAS LAB 48 App C). Leave PFAmax empty and the band becomes w = U95 — the typical band, at roughly 2.5 % false acceptance.
A U value with no provenance is not a record. The criterion holds four fields for it:
| Field | Description |
|---|---|
| Uncertainty Basis | Type A — statistical (e.g. QC) · Type B — literature/certificates · From method validation · Not estimable |
| Uncertainty Scale | Linear (± in the unit) or Logarithmic log10 — see below |
| U Evaluated On | When it was evaluated |
| U includes sampling | §7.6.1 — whether sampling uncertainty is inside U |
| U Derivation Record | §7.6.3 — sources, data, combined u and k. The full worksheet can be attached to the method validation |
Above the derivation field there is a collapsible box listing the minimum contributors a complete uncertainty budget must name under §7.6.3. Open it the first time you write a derivation.
The Uncertainty field shows, where one exists, a Type A suggestion from the most recent locked QC series for the same method–element pair. It is an indication only: it never fills the field by itself.
For microbiological counts (ISO 19036) the uncertainty is not symmetric around the value. With Logarithmic log10, U is expressed in log10 units, the interval becomes asymmetric, and the guard band is computed on the log scale.
When the examination yields no number, the "specification" is not an interval but the expected verdict, and the limits above do not apply. Choosing the Qualitative Match rule reveals an Expected Result field with the four verdicts: Detected, Not detected, Presumptive positive, Invalid examination.
With no expected result the rule judges nothing. The criterion says so on the spot, rather than leaving the lab wondering why nothing is being evaluated.
Points at — this page does not stand without these:
error appears; there is simply no ✓ or ✗. The report goes out with a number and no judgement.
sitting exactly on the limit is called "in" with a 50 % chance of a wrong decision. That is
why the criterion accepts either a C95 capability index or a maximum-uncertainty cap — fill one in.
until a person decides.