How we turn laboratory reports into evidence
Sources, analyte selection, exclusions, calculations, funding and versioning behind ChemVerify’s Research & Evidence reports.
Published by ChemVerify · Data snapshot: · 2026-11-13.v1
Rules used in this edition
Source and unit of observation
The source is ChemVerify’s public /tests/data.json, captured on 13 November 2026. One entry is one sample; a sample may contain several laboratory reports, each with several results. We preserve the published grouping and do not count each report as an independent purchase.
Comparable measurements
We reuse the test database’s analyte-selection rules. A result must uniquely match the labelled analyte, or be the single unambiguous unnamed result. Complexes and their constituents are not summed. Only exact numerical values in % for purity and mg for quantity qualify. Bounded results, incompatible units, ambiguous duplicates and invalid numbers are excluded.
Quantity versus label
The structured label quantity must be finite and positive. Difference (%) = (measured mg / label mg − 1) × 100. Missing labels are not inferred from usual product sizes. A paired row requires both an eligible purity result and a comparable quantity. Rounding is applied only for display.
Coverage and uncertainty
Coverage counts at least one published result per sample and test category. Comparisons use reported point estimates; uncertainty remains in the source JSON and original reports. We derive no confidence intervals, population estimates or significance claims from this selected set.
Funding and responsibility
ChemVerify publishes these analyses of its commissioned tests. The platform also earns affiliate commissions and fees for listings and commissioned testing. Those relationships matter when interpreting sample selection. These are ChemVerify publications, not independent peer-reviewed papers. Operating and funding disclosures are linked below.
Endotoxin results and document fields
The endotoxin analysis retains every structured endotoxin result, its original comparison sign and unit. Bounds are not substituted with exact values; no mean or safety threshold is derived. The traceability audit counts non-empty text and finite numeric fields in the frozen export. It does not authenticate documents or infer missing information from PDFs.
Analytes and qualitative non-detections
The GHK-Cu report uses the same labelled-analyte selection as the paired analysis; constituent values remain separate. The heavy-metal report includes only structured heavy_metal results. It counts samples, report records and analyte entries separately, preserves qualitative verdicts and does not substitute null values with zero or infer numerical reporting limits.
Versions and corrections
Edition 2026-11-13.v1 is tied to the downloadable JSON and its SHA-256 checksum. New observations require a new dated edition. Corrections must identify the change and retain the prior source version. Report problems through the contact page, with the protocol number, field concerned and supporting source.
Download and cite this edition
This edition uses a frozen copy of the public dataset. Later results may change the live database without changing this report. JSON preserves the sample, report and measurement fields, including uncertainty where reported.
Source data: CC BY 4.0. Attribute ChemVerify and retain the original laboratory references.
Suggested citation
ChemVerify (2026). How we turn laboratory reports into evidence. Research & Evidence, 2026-11-13.v1. https://www.chemverify.com/research/methodology
Source checksum (SHA-256)
451bee625402c1f15b7cf29ea6cbdf46d7bd915631aec9aef0753bca6c470ad8Edition history
13 November 2026 — first edition, 2026-11-13.v1.
