ISO 17034: Evidence Readiness for Reference Material Producers

ISO 17034 addresses the competence and consistent operation of reference material producers, including producers of certified reference materials. This guide asks a narrower, practical question: can a reviewer retrieve the evidence supporting a particular material claim, distinguish what was known at different dates and identify exactly what is still missing?

A useful readiness exercise starts with a disputed certificate input or a customer unit, rather than a tour of production folders. Work backward to the assigned property, evaluation, measurement inputs, preparation revision and physical units. Then work forward again to the decision, document version and recipients affected. A file being present is not the same as its technical conclusion being accepted.

For your own exercise, choose one material and one decision date. Ask the technical lead to define the claims being considered and the evidence needed for that decision. Ask a second person to retrieve the packet without relying on the preparer's memory. The useful output is not a general completeness percentage: it is a list of missing objects, affected claims, responsible reviewers and unresolved decisions.

The ISO 17034 catalogue identifies ISO 17034:2016 as edition 1, published and current following confirmation in 2022. Its lifecycle indicates revision is pending; that is not a newly published replacement edition. Keep the edition used for your assessment explicit.

The ISO 33405 scope excludes statistical methods for assessing homogeneity and stability of qualitative-property reference materials and excludes multivariate quantities such as spectral data. Do not stretch its catalogue description into a statistical procedure for those cases. The ISO 33401 catalogue describes mandatory, recommended and other information categories and minimum label requirements, but does not supply their full contents here.

Use the ISO/IEC 17025 catalogue for its laboratory scope and status, not as evidence that every measurement in this exercise requires a universally accredited system. If accreditation is claimed, examine the actual claim and scope separately. Historical ISO and ILAC references quoted inside VIM or NIST documents are not substitutes for current primary requirements.

In VIM 5.13, a reference material is sufficiently homogeneous and stable for specified properties and established as fit for its intended use in measurement or examination of nominal properties. The definition is not limited to numerical concentration standards. VIM 5.14 adds authoritative documentation, valid procedures and one or more specified property values with associated uncertainties and traceabilities for a CRM.

VIM 5.14 Note 3 explicitly includes both quantitative and nominal meanings of uncertainty and traceability. Under VIM 5.13 Note 1, uncertainty associated with a nominal property is not measurement uncertainty. Consequently, do not define every CRM as quantitative or SI-only. Use the certified reference material definition for the canonical RM/CRM explanation; this guide concentrates on retrieving producer evidence.

Terminology also needs its source attached. In NIST SP 260-136:2021, RM denotes a particular NIST-provided class, not the whole generic reference-material category. NIST also describes certified and non-certified values appearing in CRM documentation, with those categories distinguished. Do not silently apply NIST's naming conventions to another producer.

For evaluation purposes, ask which property, value status, use and reference are actually claimed. Handle accreditation evidence separately rather than adding accreditation to the VIM definition. Likewise, possession of a CRM is not a demonstration that every customer measurement result is valid: VIM explicitly warns that metrological traceability alone does not establish adequate uncertainty for a purpose or absence of mistakes.

The index below is the exercise packet. Its identifiers refer to the fictional record contents described in this guide, not downloadable documents. Start with snapshot S0, frozen at 2026-07-14T10:00Z. PLAN-041 revision 2 stipulates that initial storage and transport evaluations, an accepted calibration-chain assessment, value assignment and uncertainty review are required before this example's certificate can be authorized. Future monitoring is listed separately.

The exercise uses three decision states. S0 exposes unresolved initial evidence. S1, dated 2026-07-21T14:00Z, records later closure and certificate issue. S2, dated 2026-10-22T16:00Z, introduces a supplier-document change after distribution. Retrieve the requested state, not whichever record happens to be newest. A later approval cannot serve as evidence that approval existed at S0.

For each retrieved object, record its identifier, revision, decision status, author or reviewer and relevant timestamp. Also preserve an explicit missing-object entry. A blank attachment field is ambiguous; an entry stating that ST-T-EV-041 is required but absent makes the decision impact inspectable.

PREP-041 records execution on 2026-06-02 under PI-CU-07 revision 4, effective 2026-05-28. Revision 5 became effective on 2026-06-08 and changed transfer-record fields. The S0 export header says revision 5 because its report template used the current instruction title. That header is an inconsistency to investigate, not evidence that the operator followed a future revision.

These are recorded weighing inputs, not a certified property value. The net stock addition follows from the two displayed readings; the matrix addition follows from its own tare and gross. This exercise does not supply purity corrections, a preparation-value calculation or a measurement-uncertainty model. Do not derive a certified concentration simply because the arithmetic of the recorded masses is available.

Resolve exception EX-041-A by comparing the execution copy, instruction effective dates, report configuration and the revision-change assessment. In the fictional closure dated 2026-07-15T11:20Z, reviewer Mara Vale confirms execution under revision 4 and corrects the export mapping. The original export remains identifiable as superseded. No fictional operator signature is recreated and the preparation is not relabelled as having followed revision 5.

Keep the preparation-equipment evidence distinct from the characterization calibration chain. CAL-B14-2604 and CAL-B52-2605 concern the weighing inputs. They do not repair missing annex B for the separate analytical calibrant CAL-SOL-88. Closing the instruction-header exception therefore does not close the assignment exception.

Before a received material becomes a preparation input, make a lot-and-container-specific acceptance decision for the stated property and use. Supplier approval status alone is not evidence that this material is fit. For this readiness check, keep permission to buy from a supplier separate from the technical decision to use a particular received material under a particular preparation revision.

NIST SP 1012 (September 2004) is guidance for high-purity organic/neat chemical materials only. Within that scope, §2.3 identifies confirmation of chemical identity as the first step of purity determination, makes acceptable purity contingent on intended use and explains that impurities can make material unsuitable when they significantly influence the properties of interest. Use those distinctions here without importing the publication's historical RM/CRM definitions or prescribing its analytical procedures.

Define the Acceptance Question

1. Specify the property and proposed use. Have a qualified technical owner identify the required chemical identity and form, the purity quantity and reporting basis needed for preparation, and any impurity whose effect matters to that use. Record the acceptance rationale before comparing supplier results; a generic preference for the highest purity is not the decision basis for this exercise.
2. Bind the evidence to the received material. Record the supplier item and lot, received container identifier, receipt record and the linkage between the material examined and the material received. Where repacking occurred, preserve the parent-to-received-container relationship. Retain each supplier document's identifier, revision or issue date and receipt date, together with the actual version reviewed—not only a link that opens the latest document.
3. Review identity and purity separately. Identify what evidence supports the specified identity of this received material. Then identify what supports its suitability for the proposed use, including use-critical impurities and the reported purity basis. Preserve qualifications such as dry basis or as-received basis rather than silently treating them as equivalent. Distinguish evidence present, evidence technically accepted and evidence still missing.

Work the Two-Container Hold Exercise

Fictional manufacturer practice: two containers from an approved supplier are alternative starting materials for the same proposed preparation—not ingredients to combine. USE-Q17 revision 1 proposes a solution reference material for calibrating a specified procedure, with mass fraction of fictional organic compound Q as the intended property. A positional isomer, Q-I, also contributes to that procedure's response in this scenario, so the plan treats its assessment as necessary to starting-material acceptance. Proposed instruction PI-Q17 revision 3 expects an accepted starting-material purity input on an as-received mass-fraction basis. Receipt record RCV-Q17 revision 1 identifies both containers and their supplier lots.

The exercise assigns disposition to a technical reviewer with documented competence in the relevant organic identity, impurity and purity-basis assessments. That reviewer records both HOLD decisions in MAR-Q17 revision 1. Procurement owns the supplier requests; the technical reviewer owns evidence adequacy and the subsequent decision. Each open entry names the missing evidence, affected container, responsible owner and next review action. An acknowledgement from the supplier does not close the technical question.

For RC-Q-01, evidence of identity does not answer the unassessed Q-I question: missing assessment is not evidence of absence. For RC-Q-02, the accepted impurity assessment does not establish the missing identity linkage or resolve the purity basis. Neither container is ranked by its reported purity or accepted because the supplier is approved. The exercise ends with HOLD and requests for evidence—not a basis conversion, derived concentration or assumed preparation value.

Bind the Decision to the Preparation Input

Use the following disposition meanings for this practice record. A qualified person makes the decision; the record arrangement merely preserves it.

• ACCEPT: the reviewer concludes that the identified lot/container is suitable for the specified use and preparation revision, using the listed evidence versions and an explicit rationale. No unresolved use-critical issue is hidden behind the acceptance status.
• HOLD: identity linkage, a use-critical impurity assessment, purity basis or another necessary input remains unresolved. State the evidence request and owner; do not describe an unavailable assessment as a failed analytical result.
• REJECT: the reviewer records an established reason for excluding the material from the specified use. Preserve that reason and its evidence rather than using rejection as a synonym for an unanswered request.

If a later qualified review supports acceptance, issue an approved-input record linking the acceptance decision to the exact lot/container, supplier evidence versions, intended property and use, accepted purity basis, restrictions and permitted preparation revision. Include the qualified owner's authorization and the disposition of previously open questions. The preparation record should reference that approved-input revision, not merely the supplier name or latest attachment. At this exercise snapshot, neither container has an approved-input record.

Before carrying an acceptance forward after a change in use, supplier evidence or preparation revision, have the technical owner assess whether its rationale still applies. Preserve the earlier decision and the evidence it actually used. This closes the starting-material gate into preparation; it does not authorize a finished-material property assignment or certificate.

FILL-041 maps sequential unit identifiers to three fill periods on 2026-06-02: 0001–0288 at 08:30–10:02Z, 0289–0576 at 10:15–11:47Z and 0577–0864 at 12:30–14:08Z. Each period contains 288 units. Prefix every identifier below with CU-W26-041. The selected study ranges are stipulated solely to make this retrieval exercise enumerable.

The complete partition is 12 rejected + 28 consumed + 40 reserves + 4 returned + 176 external + 604 unreserved local = 864. The unreserved ranges contain 64, 120, 66, 166 and 188 units respectively, totalling 604. No range overlaps another disposition. Current local on-hand is 648: 604 unreserved units, 40 reserves and four quarantined returns.

At S0, before any distribution, the surviving intact population was 824 units: 864 minus 12 rejects and 28 consumed study units. Forty of those 824 were reserved, leaving 784 potential distribution units, but none was distributable while the initial evidence remained unresolved. At S1, the fictional authorization made those 784 eligible for distribution; reservations remained physical stock.

Historical dispatches total 180. Four subsequently returned, leaving 176 in confirmed external custody. Never add the historical 180 to the current local 648: that would count the four returned units twice. At S2 the unreserved 604 are physically present but held for technical review, so current distributable stock is zero. Quantity, location, reservation and permission to distribute are different fields.

A single studies-complete flag would conceal the central problem at S0. The homogeneity evaluation is accepted, characterization results have been reviewed, and one required initial stability evaluation is still missing. None of these statements is interchangeable. The following statuses are stipulated by PLAN-041, not presented as universal ISO release rules.

The missing transport evaluation is not a failed study result. Its technical conclusion is unavailable in S0. The appropriate exercise finding is therefore evidence absent for a required initial decision, rather than transport failed or stability complete. Preserve the distinction when assigning the corrective task: recover and evaluate the evidence, then decide whether it supports the proposed conditions.

Similarly, future monitoring does not replace missing initial support. Conversely, a future result is not overdue merely because it has not yet occurred. In this example, ST-V-041 revision 1 later supports the stipulated validity statement through 2027-06-30, while MON-041 retains its separate future schedule. These fictional dates are not recommended validity periods or monitoring intervals.

For context, NIST's 2021 publication treats homogeneity and stability as relative concepts and discusses storage and shipment stability in its own program. Its practice is useful context, not a sample-count rule for this exercise or a substitute for the producer's technical justification.

VIM 2.26 describes measurement uncertainty in terms of the dispersion of quantity values attributed to a measurand, based on the information used. It can include contributions associated with systematic effects and assigned measurement-standard values. A useful review therefore asks what information and evaluations support the uncertainty statement, rather than whether a document named uncertainty budget exists.

VIM 2.41 makes metrological traceability a property of a measurement result connected to a reference through a documented, unbroken calibration chain, each calibration contributing uncertainty. VIM 2.42 describes the chain as the standards and calibrations used to relate the result to that reference. Supplier-lot genealogy answers a different question.

At S0, CAL-SOL-88 is present but its referenced annex B is missing. In this scenario the annex contains the calibration relationship and uncertainty information used by TRACE-041. The exact gap is therefore not no calibration certificate: it is missing supporting annex for a particular result relationship. A balance calibration file or supplier shipment record cannot fill that gap.

ROLE-041 assigns metrologist Ivo Chen to evaluate the calibration chain, value assignment and uncertainty, with competence evidence indexed in ROLE-041. Quality reviewer Mara Vale checks packet identity, unresolved exceptions and document consistency. Authorized signatory Nia Rowe makes the certificate authorization decision. These are fictional role allocations, not a universal requirement for three separate people.

Annex B is received and assessed on 2026-07-17. Ivo accepts TRACE-041 revision 1, ASG-041 revision 1 and UB-041 revision 1 on 2026-07-20 after reviewing the now-complete inputs. Record storage preserves that reasoning and its versions; it does not make the assignment. As VIM cautions, a traceable result can still have uncertainty unsuitable for the intended purpose or contain mistakes.

FIELD-041 is a field-to-source review, not a copied certificate checklist. For this exercise, the reviewer begins with each proposed statement and follows it backward to an accepted technical source. Where the source is provisional, the certificate input remains provisional. A polished document layout does not resolve the dependency.

For a real, bounded comparison, the NIST SRM 1643f certificate, issued 28 April 2025, concerns trace elements in water and is intended primarily for evaluating methods for trace elements in fresh water. It reports mass fractions in µg/kg and mass concentrations in µg/L; the latter are derived using measured density. These are different quantity expressions, not interchangeable column headings.

That certificate's stated validity runs to 1 August 2027 subject to its conditions; improper storage or use, damage, contamination or modification nullify its certified values. It also directs users to verify that they have current documentation. Those dates and instructions belong to that product. They do not establish a general producer protocol, dating interval or permission to copy its preparation or uncertainty approach.

The practical lesson for FIELD-041 is to review meaning as well as transcription. Does the property expression match the assignment? Does the uncertainty statement describe the same value? Is the traceability statement supported by the evaluated relationship? Are conditions and validity tied to the applicable evidence rather than borrowed from another material?

At S1, Nia authorizes CERT-041 revision 1 at 2026-07-21T14:00Z after the named exceptions and technical reviews have closed. SNAP-041 S1 preserves that certificate, FIELD-041 revision 1, accepted evaluations, the 824-unit on-hand population and the 784-unit distribution authorization. It does not contain later shipment events.

On 2026-10-21, supplier document SC-417 revision 3 replaces revision 2. The fictional change clarifies the stock's uncertainty statement without changing its stated nominal value. CHANGE-041 asks whether the clarification affects UB-041, ASG-041 or certificate wording. The answer is not stipulated. At S2, a proposed CERT-041 revision 2 exists only as a draft, and the remaining unreserved stock is technically held pending review.

NIST's 2021 publication describes revisions that can include editorial changes, validity extensions, changes in value status or removal of values. It also describes its own practice of not notifying customers of extended expiration dates. Neither statement is a universal notification rule. For this exercise, record the communication decision and its reasons rather than importing NIST's policy.

Retain the historical certificate actually supplied to each recipient. If a later revision is authorized, associate it with a new event and an impact assessment; do not overwrite earlier dispatch records. A revised certificate does not retroactively change what a laboratory received or which document it recorded using. Lab A remains a historical recipient of 80 units even though its confirmed current holding is 76.

Give a reviewer the index and the following questions before discussing the answer key. Ask for identifiers, versions, quantities and decision impacts—not simply present or absent. The exercise is successful when another person can reproduce the answer from the same snapshot and can see where technical judgment remains necessary.

1. At S0, which instruction was actually followed, and why does the export disagree?
2. Identify the exact missing required initial stability object. Is it a failed result, a future monitoring result or an absent evaluation?
3. Explain why the available balance calibration records cannot close P06.
4. Reconcile all 864 units at S2 and calculate current local on-hand, external holdings and distributable stock.
5. Identify every recipient of revision 1, distinguishing historical dispatched quantity from confirmed current holdings.
6. Can the supplier-document change be called editorial for the finished material merely because its nominal stock value is unchanged?

Answer Key: What the Evidence Supports

First, PI-CU-07 revision 4 was executed. Revision 5 postdates preparation; the export used a current-title mapping. EX-041-A requires documentary resolution, later recorded on 15 July. Second, ST-T-EV-041 is the missing initial evaluation. Raw transport observations exist, but the required conclusion does not. MON-041's January 2027 event is future monitoring, not the missing initial evidence.

Third, P06 lacks CAL-SOL-88 annex B, supporting the characterization calibration relationship and uncertainty information. CAL-B14-2604 and CAL-B52-2605 concern preparation weighing and answer a different question. S0 therefore cannot support the example's final assignment or certificate authorization, even though homogeneity and characterization review records are present.

Fourth, 12 destroyed + 28 consumed + 648 local + 176 external equals 864. The 648 local units comprise 604 unreserved stock, 40 intact reserves and four quarantined returns. Current distributable stock is zero at S2 because of the technical hold. Fifth, revision 1 accompanied 180 dispatched units across three recipients. The current external total is 176, and four returned units retain their historical recipient association.

Finally, unchanged nominal value does not resolve the uncertainty impact. CHANGE-041 remains open until the technical reviewer assesses the revised source statement against the finished-material evaluations. The exercise does not predetermine a revised value, expanded uncertainty, withdrawal or customer notice. It requires the reviewer to state what remains undecided.

Use this exercise to evaluate a proposed record arrangement, whether it involves document repositories, laboratory tools, inventory records or an integrated system. Keep the evaluation product-neutral. Require demonstrated retrieval of the decision state and evidence relationships; do not treat a feature name, dashboard colour or count of attachments as proof of technical readiness.

• Retrieve S0 without later approvals appearing as if they already existed.
• Find a document by lot, property, equipment or certificate field, retaining its exact revision.
• Represent required-but-missing evidence separately from drafts, rejected conclusions and future scheduled work.
• Keep reservations within physical on-hand while distinguishing them from consumed study units.
• Reconcile returns without erasing historical dispatches or duplicating current stock.
• Preserve calculation inputs, evaluation outputs and qualified decisions separately from storage metadata.
• Produce a readable packet and exception list that another reviewer can inspect outside the original screen sequence.

For each demonstrated task, record the test data, expected result, observed result and unresolved limitation. Have technical reviewers assess scientific content and record reviewers assess identity, history and reconciliation. If an export loses the revision or cannot distinguish a missing annex from an accepted evaluation, document that limitation rather than assuming it will be solved later.

The final readiness statement should be specific: which snapshot was reviewed, which claims are supported, which evidence is missing and which decisions remain with qualified people. It should not promise compliance, accreditation or a successful inspection. The value of the exercise is a reproducible explanation of the evidence—not a claim that storing it makes the material certified.