From Evidence to Threshold: Why Occurrence Data and Certification Limits Must Remain Separate

“What concentrations have been observed?” and “What concentration will this certification permit?” both produce numbers. They are not the same scientific or institutional question.

Occurrence research describes a sampled world. It may reveal typical values, tails, regional patterns, method effects, vulnerable matrices, or gaps in surveillance. Its distributions depend on which products were sampled, when and where they were collected, how non-detects were treated, which analyte species were measured, and whether the reported basis matches the product as sold or consumed.

A certification limit is a rule of conformity. It determines whether a defined product, on a defined basis, for a defined analyte, satisfies a published requirement. That rule has legal, commercial, and governance consequences. It therefore needs an authority basis that remains valid even when the available occurrence literature is sparse, biased, or rapidly changing.

A percentile may be an informative description of an occurrence pool. It does not, by itself, establish what should be permitted. If the pool overrepresents one geography, laboratory method, product form, or quality tier, the percentile inherits that structure. If contaminated products dominate the observed market, a market-derived limit can normalize poor performance. If only high-performing products were tested, the same method can set a limit that the broader category was never given a fair opportunity to satisfy.

A descriptive distribution cannot supply its own normative authority. It can inform context, surveillance, and later review; it cannot silently become the rule it is being used to judge.

This is not an argument against statistical evidence. It is an argument for assigning statistical evidence to the correct decision. Occurrence findings are indispensable for understanding exposure, prioritizing research, identifying anomalous matrices, and testing whether a standard is producing meaningful improvement. Those uses do not require the occurrence pool to select the initial limit.

Heavy Metal Tested & Certified begins with the strictest applicable credible government maximum for the exact product-and-analyte cell. The purpose of that rule is not to assume that every government value is optimal. It is to make the initial authority source public, external, and reproducible.

1. 1
   Define the exact cell

   Fix product matrix, form, intended consumer, analyte species, unit, and native reporting basis.

2. 2
   Identify applicable sovereign rules

   Collect credible maximum levels that genuinely govern that cell rather than a superficially similar one.

3. 3
   Select the strictest applicable maximum

   Preserve formulation, species, age, and jurisdictional distinctions instead of averaging them away.

4. 4
   Disclose the unresolved edge

   Use a documented sovereign read-across or control, or leave the cell in development rather than manufacture false precision.

A reader can challenge any step: whether the rule is credible, whether it applies to the matrix, whether the basis is comparable, or whether a stricter instrument was missed. The challenge is possible because the method does not hide the authority source inside a proprietary score.

A concentration without its analyte species and reporting basis is incomplete. Total arsenic is not interchangeable with inorganic arsenic. Total mercury is not automatically equivalent to methylmercury. Chromium measurements do not answer the same question when hexavalent chromium is the material species of concern. The standard must state which form it regulates and how an applicable total-metal control interacts with a species-specific rule.

Basis is equally consequential. Dry, wet, prepared, drained, reconstituted, and as-sold values can differ substantially without any change in the underlying material. Converting between them requires explicit assumptions, and those assumptions can introduce more uncertainty than the apparent precision of the resulting number suggests.

HMTc therefore keeps the initial standard in the native basis of the controlling instrument wherever possible. A conversion may be shown for context, but it does not erase the native record from which the rule derives.

Government anchoring answers the initial-authority problem. It does not require a standard to remain static. A certification system may tighten a published value when its own eligible operating evidence demonstrates that the category can sustain a more demanding requirement.

That later decision uses a different dataset: governed, eligible certified-lot results collected under the program’s defined laboratory, basis, identity, and quality controls. It does not draw the new threshold from the public occurrence pool. The evidence is narrower but more comparable to the population the program regulates.

The separation of datasets also protects the public index. Certified-lot findings may contain brand-level and commercial information that does not belong on a literature surface. The ratchet may publish its rule, aggregate basis, and resulting edition without turning the independent index into a private program-data mirror.

Separating occurrence evidence from certification limits creates clear accountability. Literature curators are responsible for faithfully describing the source record. Standards architects are responsible for applying the published initial-value and ratchet rules. Certification operators are responsible for lot eligibility, mark use, and current status. Institutional oversight is responsible for conflicts, review, overrides, corrections, and version history.

The model also makes claims easier to interpret. The index can report a high or low observed concentration without implying that a product would pass certification. The program can publish a strict limit without implying that every product below it is safe in every context. Each statement remains within the authority of the system that made it.

Occurrence evidence and certification limits should inform one another without becoming one another. The boundary preserves the scientific meaning of the evidence and the institutional legitimacy of the standard.

An initial government anchor makes the authority source external and reproducible. A separate certified-lot ratchet permits later tightening under controlled, category-relevant evidence. Public occurrence research remains free to describe the world as it is found—including evidence that challenges the standard—without being recruited to justify a decision it did not make.