Two methods agree beautifully on a reference material. Must they agree on ordinary samples? Conversely, if methods have a predictable difference on ordinary samples, should that same difference appear on the reference material?
These questions introduce commutability: whether a reference material follows the relationship between measurement procedures that is observed for representative samples. This guide uses fictional measurements to show why a well-characterised material can still be unsuitable for a particular comparison. The clinical laboratory literature supplies the definition and assessment context; the examples are educational and make no claim about a named test or product.
THE GOLD STANDARD OF CHEMISTRY
Choose the question you need to answer
Define the comparison · Chapters 1–2. Build the distinction before calculating.
Work the material cases · Chapters 3–4. Follow the original worked cases.
Keep the claim in scope · Chapters 5. Identify what the evidence cannot establish.
Test a new relationship · Chapters 6. Change the conditions and check your reasoning.
Read and connect · Chapters 7–8. Resolve questions and choose the next reading.
Open the chapter contents · Try the transfer checks · The Gold Standard of Chemistry
Chapter contents
Define the comparison
Work the material cases
Keep the claim in scope
Test a new relationship
Read and connect
A reference material can carry carefully established information about a quantity. That information addresses an important question: what value is assigned, with what uncertainty and traceability? A further question concerns how that material behaves when different procedures measure it.
Commutability asks about that second relationship. NIST’s general-design publication explains the comparison between results for reference materials and representative clinical samples across measurement procedures. The property is conditional on the procedures and sample types being considered.
A processed material can differ from ordinary samples in composition or physical state. Such differences may affect procedures differently. A precise assigned concentration does not by itself establish that the material reproduces the relevant sample behaviour.
Keep three statements separate: the material has an assigned value; the laboratory can measure it repeatedly; it behaves appropriately for the intended calibration or comparison. One does not automatically establish the others.
Begin a review by naming the measurand—the quantity intended to be measured—the sample type and the exact procedures. “This is a reliable standard” is too broad unless its intended use and evidence are also clear.
Imagine procedures A and B measuring representative native samples. In this synthetic example, their results follow B = 1.1A + 2 over the chosen range. A sample giving A = 10 gives B = 13; A = 20 corresponds to B = 24; A = 30 corresponds to B = 35.
This relationship contains a proportional difference and an offset. We are describing observed method comparison, not declaring either procedure correct. A reference material giving A = 20 and B = 24 follows the same relationship.
A second material gives A = 20 and B = 30. Its deviation from the native-sample relationship is 30 − (1.1 × 20 + 2) = 6. That discrepancy raises a different question from the ordinary A–B difference already present in samples.
The reference comparison must therefore use the relevant sample relationship, not an automatic expectation that A equals B. A material whose two results are identical could be inconsistent with the native-sample relationship in this example.
These few exact points illustrate the idea only. Real assessment needs representative samples, characterised procedure performance, uncertainty and a prespecified criterion. A line drawn through three invented points is not a validation study.
| Material | A result | B result | B predicted from native relationship | Deviation |
|---|---|---|---|---|
| Native sample 1 | 10 | 13 | 13 | 0 |
| Native sample 2 | 20 | 24 | 24 | 0 |
| Native sample 3 | 30 | 35 | 35 | 0 |
| Candidate Q | 20 | 24 | 24 | 0 |
| Candidate R | 20 | 30 | 24 | +6 |
Suppose an analyst chooses material R because its assigned value is 20 and applies a crude multiplicative adjustment to B so that R’s result of 30 becomes 20. The factor is 20/30 = 2/3.
Apply that factor to the native sample with A = 20 and B = 24. Its adjusted B result becomes 16. The material now meets its target while this native sample moves farther from A’s value of 20. A success on the calibration material has not demonstrated success on ordinary samples.
This deliberately oversimplified adjustment is not a recommended calibration method. It exposes the transfer problem: a correction derived from a material-specific response can be carried into samples where that response does not apply.
Nor does this example prove that A represents the true value of every sample. Establishing trueness and traceability requires additional evidence. The narrower conclusion is that forcing R to its assigned value does not establish equivalent native-sample results.
When a report shows a perfect control result, ask what it was designed to verify. Repeatability, assigned-value agreement and transfer into routine samples are related checks with different evidential jobs.
Real native samples do not fall on a perfect line. Repeat measurements vary; procedures may respond differently to interferents; uncertainty can change across the concentration range. An assessment must account for those features before interpreting a reference material’s difference.
For a teaching decision only, suppose an allowed deviation from the native relationship is ±2 units. A candidate with estimated deviation +1 and uncertainty interval from 0 to +2 fits within that selected band. A candidate with estimated deviation +6 and interval +5 to +7 lies outside it. A candidate estimated at +2 with interval +1 to +3 crosses the boundary and leaves the decision unresolved under this simple interval rule.
Those values are invented. The ±2 rule is not a clinical recommendation or an official universal procedure. The purpose is to distinguish clear compatibility, clear incompatibility and insufficient precision under a stated rule.
NIST’s 2023 recommendations discuss quantitative criteria tied to the intended analytical performance and treatment of indeterminate outcomes. Use that professional framework for the assessment it describes, rather than importing our classroom band.
Decide what size of material-specific difference would matter before reviewing candidate results. Otherwise the criterion can quietly change to favour a convenient material.
A conclusion applies to the assessed material, measurand, procedures, range and conditions. Compatibility for A and B does not automatically establish compatibility for a third procedure C. A changed reagent lot or modified sample preparation can also call for further evidence.
Commutability does not mean that every laboratory result is accurate, that all procedures are interchangeable for every purpose, or that specimen collection and handling are sound. Those questions remain part of the wider measurement system.
Likewise, a noncommutable outcome does not make a material useless for every task. A material unsuitable for assessing trueness across procedures might still have a different, explicitly supported purpose. Intended use determines the interpretation.
Do not silently replace an unresolved result with a pass. An uncertainty interval crossing a selected boundary can indicate that the evidence lacks discrimination. More appropriate measurements or a better design may help; a stronger adjective does not.
The useful conclusion is specific: what comparison was made, which criterion was used, what evidence supported the result and where the conclusion stops. That specificity allows another laboratory or learner to judge whether the result transfers to their situation.
Task one: native samples follow B = 0.8A + 5. A candidate gives A = 25 and B = 25. Is its exact A–B agreement sufficient evidence of commutability?
The expected B result is 0.8 × 25 + 5 = 25. Here the identical results also happen to lie on the native relationship. That makes the point compatible with the illustrative relationship, but the point alone still lacks uncertainty, range and broader assessment evidence.
Task two: a candidate gives A = 40 and B = 40 under the same relationship. The predicted B is 37. Its deviation is +3. Identical A and B values now depart from the expected sample relationship. Equality is not the right general test.
Task three: the selected allowable deviation is ±2, and a candidate’s uncertainty interval runs from −1 to +3. Under the simple whole-interval rule in Chapter 4, what can you conclude? The interval crosses the upper boundary. The outcome is unresolved by that rule; it is not established as compatible merely because the point estimate is small.
Explain each answer in terms of the intended comparison. A good calculation preserves the distinction between a material response and a conclusion about ordinary samples.
Is commutability the same as traceability?
No. Traceability concerns the relationship of a measurement result to a reference through a documented calibration chain, with uncertainty. Commutability concerns whether a material behaves appropriately in the relevant comparison of procedures and representative samples.
Does a certified value guarantee commutability?
An assigned value alone does not. Inspect the certificate and supporting assessment for the intended measurand, procedures and use. Do not assume every certified material carries every possible suitability claim.
Can repeated measurements solve the problem?
They can improve information about repeatability and reduce some uncertainty. They do not automatically remove a systematic material-specific response or make an unrepresentative sample set representative.
Why use clinical chemistry sources in a Chemistry lesson?
They define and develop the particular comparison being taught. The general habit—test whether a reference behaves like the material of interest—is useful more widely, while specific clinical criteria remain within their documented setting.
Should a laboratory alter a calibration after reading this example?
The examples teach interpretation. Actual changes require the applicable validated procedure, competence, documentation and assessment evidence.
A reviewable record lists the material and batch, assigned-value information, native-sample selection, procedures and versions, measurement range, replicate structure, analysis method, uncertainty, criterion and outcome. Include unresolved findings rather than hiding them in a final label.
The existing blank-correction guide addresses a background contribution to a measured result. The standard-addition guide addresses calibration when a spike changes volume. This article asks a further question: whether a reference material supports the intended transfer between procedures and routine samples. These are complementary parts of an analytical argument.
For a student activity, use paper tables and compare each candidate with the native-sample relationship. Then change the relationship or introduce uncertainty. Ask which earlier conclusion survives and which needs revision.
The learning gain is a precise habit: identify the evidence that connects a successful reference check to the claim being made about real samples. A reassuring number becomes useful when its relationship to the intended measurement is visible.
Chemical measurement and adjacent methods
Blank corrections and honest titration results
Analytical Chemistry: methods and measurement evidence
Standard addition when spike volume changes
Quality infrastructure and learning
Writing a laboratory proficiency-testing investigation
Education, metrology and quality infrastructure
Education: learning that transfers into life
How X Works: connected subjects
Authoritative sources
Sources checked on 21 October 2026. Numerical cases and practice tasks are original teaching examples, not observations from a published experiment or official statistics.
Camara et al. (2018): assessing commutability, general experimental design — Definition, sample selection and scope of assessment.
Miller et al. (2023): setting a commutability criterion — Intended-performance criteria, uncertainty and indeterminate conclusions. The classroom bands here are not the recommended clinical calculation.