A method-scope and evidence-gap matrix is a short written record that pairs each question you want a certificate of analysis to answer with the exact statement in the certificate that addresses it, the basis on which that statement was reported, and a note of every question the certificate leaves unanswered. It is the step after reading. Reading tells you what the document says; the matrix tells you which of your questions it settles, on what terms, and which stay open. This page shows how to build one from a certificate already matched to its container.
Field-by-field reading, method meaning, and lot matching belong to the shorter guide to reading a certificate of analysis. Every example below uses placeholders in angle brackets and the words not stated; all are teaching examples. Nothing here reports a value from a real certificate or describes any Glow material.
Why begin COA interpretation with an analytical question, not a purity percentage?
A certificate is usually read from the number outward: find the percentage, judge whether it looks high enough, stop. The matrix reverses that direction. You first write down, in ordinary words, the property you need to know about the material, and only then look for the report statement that speaks to it. Eurachem's guide on the fitness for purpose of analytical methods, 2025 third edition, covers method validation and verification, performance characteristics, reporting, and quality control, and that framing starts from the question an analysis is meant to answer rather than from a result. With the question written first, each report statement yields one of three findings: it answers the question on a stated basis, it answers a neighbouring question easily mistaken for yours, or it does not address the question at all. Each finding gets its own row.
A method-scope matrix adds a written, question-by-question comparison between what a reader needs answered and what a certificate actually states. Reading establishes what the document says. The matrix establishes whether each question is addressed, on what reported basis, within what stated scope, and where the gap sits when it is not. Its output is settled questions, partly settled questions, and precise requests instead of one overall impression.
Four plain questions cover most of what a reader wants from a certificate:
- Is the sampled substance the compound named on the label, and by what criterion?
- How much of the sampled material is that compound, measured against what?
- What else was found, and at what level?
- Which portion was sampled, and what does it stand for?
A single percentage speaks to at most one of these. The others stay open until a separate statement, on its own basis, is located or requested.
How do you build a method-scope matrix from report statements?
The matrix is built from statements, not impressions. A statement is a sentence, table cell, or labelled value the report actually contains, copied word for word into the row it addresses. Each row holds one question and one statement; two statements about the same question take two rows, because they may rest on different samples or methods. If no statement addresses a question, the row still exists and its result column reads not stated. That phrase is the most important entry in the worksheet: it records that you looked, that the report was silent, and that the silence is now a documented gap rather than a private assumption. The blank layout below has the eight columns the rest of this page fills, with two teaching rows showing how a nonnumeric entry and a placeholder value are written.
| Question | Identified sample | Named method and version | Result and reporting basis | Source location | Stated scope or limitations | Evidence gap | Follow-up requested |
|---|---|---|---|---|---|---|---|
| Is the sampled material the compound named on the container? | <report-sample-id>; lot <supplier-lot> | <method-name>; revision not stated | Consistent with <compound-name>; criterion not stated | Page 1, results table, row 1 | not stated | No identity criterion; no revision; sample-to-lot link is printed text only | Request the criterion and the revision |
| How much of the sampled material is the named compound? | <report-sample-id> | <chromatographic-method>; revision not stated | Purity <value-as-printed>; denominator not stated | Page 2, summary line | not stated | No denominator; no reporting range or limit | Request the percentage's basis and calculation |
Copy the report's wording, qualifiers included, into the result column and keep your reading of it in the gap column, so the two never blend. Give each row a source location a second reader can find quickly. The matrix carries no chromatogram, no certificate image, and no number the report did not print.
What is the difference between a catalog specification and a measured result?
Four kinds of document can carry a number that looks like a purity figure, and they play different roles. A sales specification is a seller's statement of what a listed item is intended to meet. An acceptance criterion is a limit a buyer or laboratory has agreed to apply. A reported finding is what the laboratory measured on the identified sample. A report conclusion is the laboratory's judgment on whether the finding meets a criterion it was given. The matrix keeps them apart because they are written by different parties, at different times, about different objects: a listing, a decision rule, a sample, and a comparison. Writing a specification into the result column silently converts an intention into a measurement, and every conclusion drawn from that row inherits the error.
A stated specification does not equal a measured finding. A specification is a target or limit written in advance by a seller or buyer; a finding is a value or conclusion a laboratory reports for one identified sample after running a stated method. They can agree, but agreement must be shown by a report statement, not assumed from a listing. The two occupy different columns of the matrix and are never merged.
| Role | Written by | States | Cannot establish alone | Matrix placement |
|---|---|---|---|---|
| Sales specification | Seller, in a listing | A target or minimum the item is meant to meet | Anything about a particular sample or lot | Note beside the question |
| Acceptance criterion | Buyer, laboratory, or agreement | The rule a finding is compared against | Whether any sample was measured | Stated-scope column, quoted |
| Reported finding | Laboratory, for the identified sample | A value or categorical result with its basis | Compliance with a rule it was not asked to apply | Result column, verbatim |
| Report conclusion | Laboratory, comparing finding to criterion | Pass, fail, conforms, or similar | Anything the criterion did not cover | Result column, with its criterion |
A pass or fail word is readable only next to the criterion that produced it, so the matrix records the pair: conforms, against the criterion on page <n>, or conforms, criterion not stated. Do not supply a threshold the report did not state, and do not write that a finding meets a named standard unless the report names it. A catalog listing's specification and a report in the available lot documentation land in different columns of the same row.
How do you check whether a method's stated purpose matches your question?
A method statement says what an analysis was set up to find out, and that purpose must match the question in your row before its result can be entered as an answer. Eurachem's fitness-for-purpose guide, third edition, 2025, covers method validation and verification, sampling and sample handling, performance characteristics, reporting, and quality control. Validation, in that vocabulary, is showing a method suits a stated purpose; verification is confirming that a method established elsewhere performs as expected where it is now used. For a reader, that scope becomes three questions for every method line: what purpose the method serves, what material or sample matrix it was shown to work on, and on what basis its results are reported. A technique named without any of the three tells you which instrument was used, not which question was answered. In analytical writing, sample matrix means the material surrounding the substance of interest; the bare word matrix here means the worksheet.
| Question class | Reader wants to know | Look for in the method statement | Record if absent |
|---|---|---|---|
| Identification | Whether the sampled substance is the compound named on the label | The property compared, the reference compared against, and the criterion for calling them consistent | Identity conclusion present; criterion not stated |
| Quantity | How much of the compound is present, per container or unit of material | Unit, denominator, calibration or reference basis, and reported range | Value present; unit or denominator not stated |
| Impurity characterization | What else is present and at what level | Which substance classes the method can and cannot register, and its reporting limit | Purity figure present; impurity scope not stated |
The table does not say which technique answers which class; that pairing depends on the method's demonstrated scope, which only the report or its referenced method document can state. No single named assay settles every structural question about a compound, and rows on sequence order or configuration stay open until a statement addresses them. This page supplies no procedure for running or validating any method.
Why do quantity, area proportion, and identity results need separate columns?
Three kinds of result appear on research-material certificates, and they are not interchangeable. A quantity states how much of a substance is present, with a unit and a stated or implied denominator, such as an amount per container or a fraction of sample mass. A chromatographic area proportion states what share of the total detected signal in one run belongs to one peak; it is a percentage of signal, not of mass, and covers only substances the detector responds to under that run's conditions. An identity statement is categorical: a measured property was found consistent with an expectation, and no amount is attached. Preserve the unit, the percentage label, the declared denominator, and the stated scope exactly as printed, and keep each type in its own row. Introductory method meaning is covered in the shorter reading guide.
An undefined purity percentage is recorded as the word purity, the value exactly as printed, and the phrase basis not stated. It is not recorded as a mass fraction, an area proportion, or a content figure, because the report did not say which it is. The matrix's evidence-gap column notes the missing definition and its follow-up column holds a request. Defaulting to mass fraction replaces the laboratory's statement with a guess.
| Result type | Basis to preserve | Cannot become |
|---|---|---|
| Quantity, a number with a unit | The unit and what it is measured against | A share of signal, or an identity conclusion |
| Chromatographic area proportion, a percentage labelled area or relative area | Total detected signal in that run, under its conditions | A mass fraction, an amount per container, or a statement about unregistered substances |
| Identity statement, such as consistent with or confirmed | The property compared and the criterion applied | Any amount or proportion, or a statement about other substances |
The bare word purity is the most common way these types get merged: it may label an area proportion, a mass-based content figure, or a composite defined elsewhere in the laboratory's documents. A value copied without its unit has lost its type; a percentage copied without its label has lost its denominator. Both go back to the report.
When should reporting limits and uncertainty be recorded on a COA matrix?
A reported value comes with boundaries, and the boundaries are part of the result. The reporting range is the span of values the method was set up to report. A detection limit is the level below which the method does not claim to see the substance; a quantification limit is the level below which it does not claim to measure it. Uncertainty is the laboratory's stated interval around a value. Caveats are qualifying sentences near a result, such as a note that a figure was calculated rather than measured. The matrix has a field for each, filled only from the report. An absent value stays absent: a reader does not estimate a limit from experience, borrow one from another document, or compute a confidence figure from the digits printed. The blank worksheet below records what the report supplied for one result and, just as usefully, what it did not.
BOUNDARIES FOR ONE RESULT ROW (teaching worksheet; every value is a placeholder) Value as printed ........ <value-as-printed> unit: <unit-as-printed> Reporting range ......... [ ] stated: <range-as-printed> [ ] not stated Detection limit ......... [ ] stated: <limit-as-printed> [ ] not stated Quantification limit .... [ ] stated: <limit-as-printed> [ ] not stated Uncertainty ............. [ ] stated: <interval-as-printed> [ ] not stated Caveat text nearby ...... [ ] yes: <caveat-as-printed> [ ] none found Source location ......... page <n>, <section-or-table> A box left at not stated is the finding. Do not fill it from memory.
Two entries need care. The phrase not detected describes the method's reach, not a zero amount: the substance was not seen above whatever limit the method applies, and if that limit is not stated the row reads not detected, limit not stated. Turning it into zero adds a claim the laboratory never made. The second is uncertainty. The NIST policy on metrological traceability states that traceability is a property of a measurement result and requires a documented chain of calibrations with stated uncertainty. It is not a consumer acceptance rule, and the same policy is explicit that traceability by itself does not establish fitness for a particular purpose. A stated uncertainty is recorded as part of the result, its absence as a gap, and neither becomes a verdict. Nor does a title supply boundaries: NIST's reference material definitions describe a certified reference material certificate as one that specifies property values with their uncertainty and metrological traceability, and a document titled certificate of analysis for ordinary catalog material is not automatically that kind of document. This page applies that terminology to no Glow material.
What does the sampled material on a certificate actually represent?
Every result on a certificate was produced from a test sample, the physical portion that went through the instrument. That portion was taken from submitted material, which is what reached the laboratory. The submitted material may have come from a production lot, a quantity the producer identifies with one code, and the lot may be one component of a finished combination that contains several. These are four different objects, and a result belongs to the first. The Eurachem and CITAC guide on measurement uncertainty arising from sampling, second edition, 2019, describes sampling and physical sample preparation as parts of the measurement process, each contributing uncertainty to the result. So a result cannot be moved up the chain from test sample to lot, or across it from one component to a combination, unless the report or a referenced document states the sampling basis that justifies the move.
| Level | What it is | Identified on paper by | Says about the next level up |
|---|---|---|---|
| Test sample | The portion measured | A sample identifier on the report | Nothing, unless the report states how the portion was taken |
| Submitted material | What reached the laboratory | A submission record, received-on date, or description | Nothing, unless a documented submission-to-lot link exists |
| Production lot | A quantity identified by one producer code | The lot code on the container and any report field repeating it | Nothing about a combination containing other components |
| Finished combination | An article assembled from several components | A component list on the label or listing | Top of the chain; needs its own report or a documented basis per component |
Neither the table nor this page says how a laboratory should sample or asserts that any sampled item represents an entire lot; both need a documented sampling basis. If a report prints a sample code and a lot code, the identified-sample column holds both, with a note on whether the report states how they are linked or merely prints them side by side. A report on one component fills only the rows about that component; rows about the combination or the other components stay open.
How do you turn evidence gaps into precise requests to a laboratory?
A finished matrix has entries in its gap column, and each is either left as a documented limit or turned into a request. A good request names one statement the reader could not find, says where the reader looked, and asks for a document or a definition rather than for reassurance. It does not ask whether the material is good, whether the laboratory is reliable, or whether a number is normal; those are judgments, and a request for a judgment returns an opinion that cannot be entered in a row. Three patterns cover most gaps this page produces: define a reported percentage, identify which submitted sample a result covers, and supply the method revision or scope statement a report references but does not include. Each is written below as a teaching example with placeholders.
| Pattern | Gap | Request wording | Reply that closes the row | Reply that leaves it open |
|---|---|---|---|---|
| Define the reported percentage | Purity or similar printed without a basis | On report <report-id>, page <n>, the value <value-as-printed> is labelled purity. Please state what quantity it is a proportion of and how it was calculated. | A written definition of basis and calculation, ideally a method or report-format document | A restated number or a general assurance |
| Identify the covered sample | A sample identifier missing or not matching the container | Please state which submitted sample the result on page <n> of report <report-id> describes, and how that sample is linked to lot <supplier-lot>. | A sample identifier plus a sourced statement of the sample-to-lot link | A repeated lot code with no statement of the link |
| Supply the referenced method or scope | A method cited by code or revision but not included | Report <report-id> cites method <method-name>, revision <revision-as-printed>. Please supply that revision's scope statement or the document defining it. | The scope statement, or a precise reference to it, at the revision cited | A newer or unspecified revision, or a summary omitting the scope |
Requests stay documentary: they ask for statements that exist or should exist in the laboratory's records and can be answered by sending a page. Wider questions, such as how a supplier handles documentation over time or what to weigh before a purchasing decision, belong to the broader vendor evaluation guide. A reply is entered exactly as received, with date and source; a documented non-answer is evidence too.
How do you write a conclusion with the same scope as the certificate?
The last line of the work is a sentence no wider than the rows above it. The template is fixed: this source reports X for identified sample Y under stated method Z, and the following questions remain unresolved. X is the result as printed with its basis. Y is the sample identifier as printed, with the lot link at whatever strength the report gave it. Z is the method as named, at the revision stated or with revision not stated. The unresolved list is the gap column, copied, so the sentence cannot say more than the certificate said. Three levels stay distinct and are reached independently: readable and present, meaning the document exists and can be read; authenticated and matched, meaning its identifiers have been checked against the container and, where possible, the issuer; and analytically adequate, meaning its statements address the questions asked on stated bases within stated scope.
When a desired attribute is not reported, the right conclusion is that the certificate does not address it. The matrix row records the question, the words not stated in the result column, and a request in the follow-up column, and the conclusion sentence lists the attribute under unresolved questions. It is not recorded as absent from the material, as satisfactory, or as zero, because the report made no statement about it.
CONCLUSION TEMPLATE (teaching example; every bracketed item is a placeholder)
This source, <report-id> dated <date-as-printed>, reports
<result-as-printed>, basis <basis-as-printed or not stated>,
for identified sample <report-sample-id>,
linked to lot <supplier-lot> by <printed text only | stated link | not stated>,
under stated method <method-name>, revision <revision-as-printed or not stated>,
with stated scope <scope-as-printed or not stated>.
Unresolved: - <question copied from an open row>
Level reached: [ ] readable and present [ ] authenticated and matched [ ] analytically adequateTwo limits on the publisher belong in the conclusion too. Glow is a commercial publisher of research-material information; a guide from a seller can explain how to read evidence but cannot supply independent evidence about the seller's own listings, which comes only from the reports and the issuing laboratory. This page ranks no vendors or laboratories and awards no badge; no reading of a certificate yields a global verified pure label, because every statement it makes is bounded by a sample, a method, a basis, and a scope.
Frequently asked questions
Can one analytical result answer every question about a material?
No. A single result answers the one question its method was set up to address, on the basis it reports, for the sample it identifies. An identity conclusion says nothing about amount, and a chromatographic proportion says nothing about substances the detector does not register. In a method-scope matrix each question gets its own row, a result is entered only where its method addresses the question, and the other rows read not stated until a separate statement is found or supplied.
What is the difference between the specification and the reported finding?
A specification is written in advance by a seller or buyer and describes what a listed item is meant to meet. A reported finding is written afterwards by a laboratory and describes what was measured on one identified sample using a stated method. One describes a listing; the other describes a sample. They can agree, but only a report statement that compares them shows it, read with the criterion it used, and a method-scope matrix keeps the two in separate columns.
Should a missing result be recorded as zero?
No. A missing result is recorded as not stated. A result printed as not detected is recorded as not detected with the reporting limit if the report gives one, or with limit not stated if it does not. Zero is a measured value with a basis and a boundary, and a report that made no measurement, or one that stopped at a limit, has not supplied it. Writing zero adds a statement the laboratory never made.
What should I ask when a percentage has no stated basis?
Ask what quantity the percentage is a proportion of and how it was calculated, naming the report, the page, and the value as printed. A usable reply states the denominator, such as total detected signal in one run or the mass of the sample, and the calculation. A reply that only repeats the number does not close the row; until a definition arrives, a method-scope matrix records purity, the value as printed, and basis not stated.
Does a certificate for a component cover the finished combination?
Not by itself. A component certificate describes the sample of that component the laboratory analyzed. A finished combination is a different object, made from that component and others, so questions about it need either a report on the combination or a documented basis linking each component result to it. A component report fills only the method-scope matrix rows about that component; rows about the combination or the other components stay at not stated and generate a request.
Does a more readable certificate establish stronger analytical evidence?
No. Readability is the first of three separate levels. A clear, complete, easy-to-follow document has reached readable and present. Whether its identifiers match the container and the issuer is a second, independent check. Whether its statements address the reader's questions on stated bases within stated scope is a third. A well laid out report can be silent on the question that matters, and a hard to read one can answer it exactly.
Sources and scope notes
The four external sources were read as publisher landing or policy pages on September 20, 2026, Pacific time. The Eurachem records are guide landing pages; this page uses only their stated scope, claims no review of either full text, and reproduces no procedure, threshold, or performance value. Nothing here applies NIST certification terminology to any Glow material, and a lot code, report identifier, or web address is not metrological traceability.
Every row, worksheet field, request wording, and conclusion sentence is a teaching example built from placeholders and the phrase not stated. None reproduces a real certificate, laboratory, lot, value, limit, or price, and none describes any Glow material. This page is research documentation only: not a testing procedure, sampling design, method validation, purchasing recommendation, or ranking. Glow is a commercial publisher of research-material information and supplies no independent evidence about its own listings.
- Eurachem, The Fitness for Purpose of Analytical Methods, third edition (2025): guide landing page Read September 20, 2026; stated scope only.
- Eurachem / CITAC, Measurement uncertainty arising from sampling, second edition (2019): guide landing page Read September 20, 2026; sampling as part of the measurement process.
- NIST, Policy on Metrological Traceability (NIST P 5800.00, effective May 31, 2019) Read September 20, 2026; traceability definition and its limit on fitness for purpose.
- NIST, Standard Reference Materials: Definitions Read September 20, 2026; reference material definitions only.