Quick answer: Translating a laboratory proficiency-testing report correctly means preserving the assigned value, participant result, comparison group, assessment standard deviation, score formula, uncertainty, rounding rule and the provider’s actual interpretation. Do not turn a satisfactory proficiency-testing result into accreditation, a warning score into a laboratory failure, or an outside-range result into a claim that a patient or product was harmed. Translate the evidence before summarising the judgement.
How should you translate proficiency testing reports, interlaboratory comparisons, z scores, z-prime scores, zeta scores, assigned values, PT evaluation statements and laboratory performance tables? Begin with the original scheme instructions and its result table, identify which statistic was used, protect every sign and unit, and then render the interpretation in language that distinguishes a numerical score from the provider’s formal performance category. This is a practical guide to accurate scientific translation and interlaboratory proficiency test report QA.
A common translation error happens when a report says ‘questionable performance’ and the target version says ‘failed accreditation’. Another occurs when a result of z = −2.4 loses its minus sign or when ‘assigned value uncertainty’ is merged with ‘standard deviation for proficiency assessment’. These phrases are not interchangeable. ISO 13528:2022 sets out statistical approaches for proficiency testing by interlaboratory comparison, while ISO/IEC 17043:2023 covers requirements for competent, impartial proficiency-testing providers. Neither reference permits a translator to manufacture an outcome that the scheme has not stated.
This page has one canonical job: translate the participant-facing proficiency test performance report. General calibration certificates, measurement traceability and uncertainty have their own translation owner, and the calibration certificate guide remains the next route when the source is a certificate rather than an interlaboratory comparison. Mathematical explanations of scores help learners understand the report, but translating a report is not the same as redesigning its assessment method.
Why a readable report can still carry the wrong scientific conclusion
Suppose a fictional organiser sends a laboratory an evaluation for measuring lead in water. The assigned value is 10.0 micrograms per litre, the participant reports 11.4, the assessment standard deviation is 0.7, and the provider obtains z = +2.0. A translator replaces ‘satisfactory according to the scheme’s stated boundary’ with ‘measurement certified accurate’. The first phrase is a scheme-specific assessment of one reported result; the second suggests a broader certification. It loses limits, evidence context and the distinction between a comparison and a credential.
A proficiency test often combines laboratory measurements, context supplied by the organiser, a statistical performance criterion and a final narrative. Each layer has a different owner. The participating laboratory reports a result. The organiser specifies a comparison value, method and evaluation rule. The translator communicates these facts without recalculating them except as an independent checking aid. Accreditation bodies, customers and laboratory managers may draw further conclusions, but those conclusions do not originate in translation.
First principles before translating any PT document
1. The measurand decides what the number refers to
A measurement is not just an isolated number; it identifies a property in a material, often with method conditions. ‘Mass concentration of lead in drinking water’ and ‘total lead in dried sediment’ cannot share an unqualified label. Preserve analyte, matrix and any reporting basis so a reader does not compare results that answer different questions. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
2. Assigned value is a defined comparison reference
The PT provider may derive an assigned value from a reference measurement, a consensus of participants or another declared method. A translator should say assigned value rather than inventing ‘true value’. The source may explicitly describe limitations and uncertainty; keep them attached to that assigned value. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
3. Assessment standard deviation is not automatically laboratory uncertainty
The sigma used for proficiency assessment measures the performance criterion in the chosen score. It is not necessarily the participant’s uncertainty and does not always equal the spread of current participant results. If the report explains how sigma was selected, translate that explanation without reinterpreting its meaning. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
4. Signs matter as much as magnitudes
A positive z score indicates a result above the assigned value under the usual formula; a negative score indicates below. An absolute-value category can disregard direction for classification, but the signed number still conveys diagnostic information. Never omit a negative sign to make the report look tidier. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
5. z, z-prime, zeta and En are not alternative spellings
Different scores incorporate differing uncertainties and conventions. z is typically the difference divided by the proficiency assessment standard deviation. z-prime also accounts for uncertainty of the assigned value in its denominator. zeta and normalised-error approaches use their own specified uncertainty constructions. Preserve the exact score name. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
6. Provider interpretation governs the output category
Many schemes use familiar thresholds such as absolute z at or below 2, between 2 and 3, and at or above 3, but details, inclusivity, terminology and decision procedures belong to the actual scheme. A translation should quote the scheme’s published criteria, not silently impose a general textbook rule. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
7. Standard uncertainty and expanded uncertainty are different
Standard uncertainty commonly uses a coverage factor of one; expanded uncertainty uses a stated factor or coverage approach. A reported numerical uncertainty without its type and coverage information can mislead. Keep u, U, k and the associated units distinct, particularly in columns with similar headings. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
8. A qualitative result is not a quantitative score
Some proficiency tests compare positive/negative, present/absent or categorical observations. Turning ‘incorrect identification of organism’ into an invented z score is nonsensical. Translate the actual nominal outcome and the method used to judge it. Preserve ‘not evaluated’ when evaluation was impossible. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
9. An accredited provider does not accredit each participant
A report can be issued by a provider whose activities fall within an accredited scope. That status does not itself certify the competence of every participating laboratory or validate every measurement method. Do not make the provider’s status sound like an automatic pass for a participant. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
10. Missing, censored and invalid results need their own states
‘Not reported’, ‘below detection limit’, ‘withdrawn’, ‘excluded from scoring’ and ‘not applicable’ are different data conditions. They must not all become ‘zero’ or ‘not detected’. The report’s statistical treatment depends on which condition occurred. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
11. Comparison groups and units determine comparability
A result compared within one analytical method subgroup might not have the same assigned value as another subgroup. Units such as mg/kg versus µg/L cannot be interchanged without density, matrix or reporting-basis information. Preserve subgroup, matrix and unit labels in every table. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
12. Translation cannot alter the scientific record
If calculations are suspicious, the translator may verify arithmetic, mark a query and ask the provider. The translator must not change the laboratory’s submitted result, adjust the score or reclassify performance without documented authority. A corrected scientific report requires a controlled revision from the responsible organisation. A reliable translation maps the term to its precise role in the evaluation, then checks that every occurrence in table headings, explanatory notes and conclusions uses the same distinction. A single accurate paragraph is not enough when another label elsewhere silently contradicts it.
A twelve-step translation and verification workflow
Step 1: Secure the right report revision
Record the PT scheme, round, issue date, provider, participant code, method and document version. Some reports are preliminary and later corrected. Translate the current approved version, while retaining the previous version’s identifier for a change audit. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 2: Classify the document’s purpose
Is this a provider’s final performance report, participant result sheet, technical appendix, statistical plan, appeal decision or accreditation correspondence? The terms ‘evaluation’ and ‘assessment’ can perform different jobs across these documents. Identify the reader before choosing phrasing. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 3: Build a protected-field register
Copy sample identifiers, analyte names, participant codes, units, assigned values, uncertainties, deviations, statistical score labels and provider conclusions. Keep the values in a separate review sheet, so they can be compared independent of the prose. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 4: Read the score definition
Find the exact equations and limits used by the scheme. Do not rely on memory of a familiar z-score formula. Some providers adjust comparison groups or specify treatment of uncertainty and bias; the translation must follow the stated design. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 5: Preserve notation and mathematical operators
Protect minus signs, absolute-value bars, inequalities, subscripts, multiplication signs and denominators. Translate adjacent descriptions but do not ‘simplify’ an equation at the risk of changing the criterion. Mathematical checking should occur against source values. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 6: Translate assessment categories narrowly
Use the provider’s terms for satisfactory, questionable, unsatisfactory, not evaluated or other stated outcomes. Do not make these into universal labels of pass, fail, certified, invalid laboratory or regulatory compliance unless those words are explicitly authorised. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 7: Check the reference value provenance
Report whether an assigned value is reference-based, consensus-based or determined in another declared way. Maintain notes on uncertainty and traceability. The reader should see which number is an assigned value and which number was measured by the participant. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 8: Respect rounding and significant figures
A displayed z of 2.0 may be based on a more precise calculation. Do not reclassify a boundary score by calculating from rounded display columns alone. Preserve stated significant figures and refer to the provider’s classification whenever rounding affects the category. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 9: Separate participant-specific findings from all-laboratory statistics
Tables may show all-participant histograms, method subgroups, robust averages and a laboratory’s own score. Translate axes and captions so readers do not mistake a population distribution for the single participant’s result. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 10: Run a numerical identity audit
Use a second pass for values, signs, symbols, units, decimal separators, confidence statements and category wording. Perform arithmetic checks where possible, but never overwrite the scientific source because a translation checker produced a different result. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 11: Review limitations and required response
A provider may recommend investigating a discrepancy, documenting corrective action or considering possible causes. Translate the recommendation as a recommendation unless it is expressed as a mandatory condition. A participant’s internal corrective action may be governed by additional requirements. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Step 12: Approve, version and distribute responsibly
Obtain bilingual technical review from a laboratory professional, release a controlled version and document corrections. If an assigned value or score changes, propagate the update to tables, graphs, conclusions, notices and archived language copies. Verification question: which source field, criterion or authorised interpretation supports this exact target wording? If that evidence is missing, log a query. Clarity is useful only when it preserves the actual distinction between measurement, comparison and judgement.
Worked cases: 26 fictional reports and translation decisions
Every number and laboratory in these exercises is fictional. These are not actual PT scheme assessments. Where a threshold or interpretation is mentioned, it is a hypothetical scheme rule for teaching; the real provider’s rules always govern real reports.
Case 1: Assigned value versus true value
Source wording: Assigned value: 10.0 mg/L. Misleading translation: True concentration: exactly 10.0 mg/L. Correct interpretation: The assigned value is the scheme’s designated comparator. Its derivation and uncertainty can be material; ‘true’ overstates certainty.
Independent review: Check the provider’s method section before choosing the label. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 2: Participant result versus consensus
Source wording: Your laboratory reported 10.8 mg/L. Misleading translation: All laboratories reported 10.8 mg/L. Correct interpretation: The row belongs to one participating laboratory. Do not turn an individual submission into the group average.
Independent review: Trace the participant ID back to the report heading. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 3: Positive z-score direction
Source wording: z = +1.5. Misleading translation: z = −1.5. Correct interpretation: The magnitude is unchanged but the sign reverses the direction of departure from the assigned value.
Independent review: Check whether the reported result lies above or below the comparator. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 4: Negative score is not a poor score by itself
Source wording: z = −0.6. Misleading translation: Performance is negative. Correct interpretation: A minus sign gives direction, not an adverse performance category. Interpretation depends on the magnitude and scheme criteria.
Independent review: Read the provider’s evaluation wording before giving a summary. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 5: z-score denominator
Source wording: z = (x − assigned value) / σpt. Misleading translation: z = difference / participant uncertainty. Correct interpretation: Assessment standard deviation and individual measurement uncertainty are not interchangeable.
Independent review: Confirm the denominator from the published scheme description. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 6: z-prime typography
Source wording: z′ = −1.3. Misleading translation: z = −1.3. Correct interpretation: The prime mark identifies a different statistic. Keep it visibly attached to z in the target title, table and conclusion.
Independent review: Compare the glossary and formula to the score column. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 7: zeta uses uncertainty differently
Source wording: ζ = 0.9. Misleading translation: z = 0.9. Correct interpretation: A zeta score cannot be relabelled z merely because both are dimensionless scores.
Independent review: Verify how participant and reference uncertainties were combined. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 8: En label versus score
Source wording: Eₙ = 0.4. Misleading translation: z-score = 0.4. Correct interpretation: The normalised error uses a distinct formulation and often expanded uncertainties. Keep symbol and evaluation rule together.
Independent review: Check units and coverage-factor discussion where supplied. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 9: A score near a category boundary
Source wording: Displayed z = 2.00; provider says satisfactory. Misleading translation: Questionable because rounded score is exactly two. Correct interpretation: Do not invent an inclusive boundary or overturn the provider’s computed classification from rounded output.
Independent review: Compare the unrounded value or documented rule if available. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 10: Uncertainty of assigned value
Source wording: u(xpt) = 0.20 mg/kg. Misleading translation: Participant uncertainty = 0.20 mg/kg. Correct interpretation: The uncertainty belongs to the assigned value, not necessarily the participant result.
Independent review: Map subscript and glossary definitions before translating. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 11: Expanded versus standard uncertainty
Source wording: U = 0.40 mg/kg, k = 2. Misleading translation: u = 0.40 mg/kg. Correct interpretation: Lowercase u and uppercase U are not cosmetic. Preserve coverage-factor information.
Independent review: Look for the uncertainty statement associated with the table. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 12: Unit prefix
Source wording: Lead 12 µg/L. Misleading translation: Lead 12 mg/L. Correct interpretation: Micrograms and milligrams differ by a factor of one thousand. A wrong prefix can create a dangerous apparent deviation.
Independent review: Run a dedicated Greek-letter and SI-prefix check. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 13: Concentration basis
Source wording: Moisture-corrected result: 21 mg/kg dry matter. Misleading translation: 21 mg/kg wet matter. Correct interpretation: Dry and wet mass bases define different quantities; the text must keep the basis in labels and notes.
Independent review: Find the preparation and calculation footnote. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 14: Analyte identity
Source wording: Total chromium. Misleading translation: Hexavalent chromium. Correct interpretation: Different analytes are being described; a familiar generic word must not collapse the chemical distinction.
Independent review: Match chemical names and element forms to the laboratory method. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 15: Sample matrix
Source wording: Sample: seawater. Misleading translation: Sample: drinking water. Correct interpretation: These sample matrices can have very different analytical and comparison implications.
Independent review: Check sample description in both performance table and instructions. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 16: Method-group score
Source wording: Compared within GC-MS method group. Misleading translation: Compared with every laboratory. Correct interpretation: The reference subgroup is a core condition of the performance score, not optional background.
Independent review: Identify the group’s label and whether other method groups have different values. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 17: Not evaluated
Source wording: Result not evaluated due to insufficient comparison data. Misleading translation: Unsatisfactory result. Correct interpretation: No score is not the same as a failed score. Preserve the reason evaluation was withheld.
Independent review: Do not populate an absent score with zero. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 18: Less-than result
Source wording: Reported < 0.05 mg/L. Misleading translation: Reported 0.05 mg/L. Correct interpretation: The less-than operator means the result is below a threshold, not a measured equality.
Independent review: Check the reporting limit and treatment of censored values. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 19: Zero and missing are different
Source wording: Result field blank—no submission. Misleading translation: Reported zero. Correct interpretation: A missing response is not a measurement of zero concentration. Empty fields should remain explicitly unreported.
Independent review: Confirm source data status with the organiser. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 20: Qualitative organism identification
Source wording: Organism not identified correctly. Misleading translation: Negative culture result. Correct interpretation: The assessment concerns identification of a material, not necessarily the absence of growth.
Independent review: Use the provider’s categorical scheme and specimen details. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 21: Proficiency versus accreditation
Source wording: Performance satisfactory in this PT round. Misleading translation: Laboratory accredited for the method. Correct interpretation: Passing one scheme round does not itself confer accreditation or method scope.
Independent review: Read the provider’s accreditation and disclaimer text. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 22: Reference material status
Source wording: PT test item was prepared for the scheme. Misleading translation: Certified reference material. Correct interpretation: A PT item is not automatically a certified reference material; technical status must be sourced.
Independent review: Translate documentation, not assumed material credentials. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 23: A warning requests investigation
Source wording: Questionable performance; laboratory should investigate. Misleading translation: Laboratory must suspend operations. Correct interpretation: A suggested investigation is not a mandatory regulatory shutdown.
Independent review: Preserve modality and who has authority to require action. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 24: Revision of assigned value
Source wording: Corrected report replaces assigned value 6.9 with 7.1. Misleading translation: Both values are currently valid. Correct interpretation: The revised report may supersede the earlier comparator and affect calculations.
Independent review: Check report version and effect on the final category. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 25: Chart axis not a percent
Source wording: Frequency of participating laboratories. Misleading translation: Probability of success, percent. Correct interpretation: A histogram count cannot be renamed a likelihood or performance rate.
Independent review: Read axes, plotted population and scale before localising. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Case 26: Scheme’s overall conclusion
Source wording: Results indicate satisfactory performance for the reported analytes. Misleading translation: Laboratory is competent for all testing. Correct interpretation: The conclusion is constrained to the analytes, test items, time and specified evaluation criteria.
Independent review: Carry the scope qualifiers into the final translated executive summary. This case demonstrates a general principle: translation must preserve the mathematical object and the authority that interprets it. Build a reversible trail from translated label to source term, reported value, scheme rule and the original conclusion, rather than substituting a convenient nearby phrase.
Frequent failure modes and meaningful repairs
Repair 1: Using pass/fail without the provider’s permission
Keep the scheme’s precise category. When it defines satisfactory, questionable or unsatisfactory ranges, explain the category without claiming wider conformity status. The practical remedy is to identify the data owner, confirm the source criterion, document the local translation decision and then test the result in the table and surrounding prose. A second reviewer should be able to explain the category without guessing at a formula.
Repair 2: Treating sigma for assessment as a standard uncertainty
Maintain separate glossary entries for σpt, participant u and assigned-value uncertainty. A QA scan should flag any target heading that collapses them. The practical remedy is to identify the data owner, confirm the source criterion, document the local translation decision and then test the result in the table and surrounding prose. A second reviewer should be able to explain the category without guessing at a formula.
Repair 3: Recalculating from rounded data
Calculate only as a check and note when the source may use unrounded values. Never overwrite the organiser’s formal score or judgement without an authorised corrigendum. The practical remedy is to identify the data owner, confirm the source criterion, document the local translation decision and then test the result in the table and surrounding prose. A second reviewer should be able to explain the category without guessing at a formula.
Repair 4: Dropping a negative sign
Use a character-level review for all scores. Test copy/paste, PDF export and font fallback; a small glyph can determine whether the result lies above or below the comparator. The practical remedy is to identify the data owner, confirm the source criterion, document the local translation decision and then test the result in the table and surrounding prose. A second reviewer should be able to explain the category without guessing at a formula.
Repair 5: Changing the measurement matrix
Link each analyte to sample, reporting basis and unit in a protected glossary. A chemical name translated well but attached to the wrong material still misinforms. The practical remedy is to identify the data owner, confirm the source criterion, document the local translation decision and then test the result in the table and surrounding prose. A second reviewer should be able to explain the category without guessing at a formula.
Repair 6: Assuming an international standard supplies every scheme decision
ISO guidance describes methods and provider competence, but the actual proficiency scheme may define detailed classifications and groupings. Translate the scheme document, not a generic memory of it. The practical remedy is to identify the data owner, confirm the source criterion, document the local translation decision and then test the result in the table and surrounding prose. A second reviewer should be able to explain the category without guessing at a formula.
Repair 7: Confusing provider and participant responsibilities
Label the organiser’s statistical decisions separately from a participant’s investigation, root-cause analysis or corrective action. Translate recommendations and obligations with their correct agent. The practical remedy is to identify the data owner, confirm the source criterion, document the local translation decision and then test the result in the table and surrounding prose. A second reviewer should be able to explain the category without guessing at a formula.
Repair 8: Inventing reference traceability
Only report a measurement’s stated traceability, calibration and reference-value method. Do not add ‘certified’ or ‘SI traceable’ as prestige words. The practical remedy is to identify the data owner, confirm the source criterion, document the local translation decision and then test the result in the table and surrounding prose. A second reviewer should be able to explain the category without guessing at a formula.
Repair 9: Flattening qualitative findings into a number
Keep categorical states and missing-result descriptions. The right answer for a qualitative identification test may be a named organism rather than a z score. The practical remedy is to identify the data owner, confirm the source criterion, document the local translation decision and then test the result in the table and surrounding prose. A second reviewer should be able to explain the category without guessing at a formula.
Repair 10: Ignoring updates to numerical tables
A revised assigned value can propagate to plots, score calculations and conclusions. Controlled document review must compare every location where the affected result appears. The practical remedy is to identify the data owner, confirm the source criterion, document the local translation decision and then test the result in the table and surrounding prose. A second reviewer should be able to explain the category without guessing at a formula.
Practice and transfer: eight laboratory translation exercises
Exercise 1: Compute and label a simple z score
Use fictional x = 12.5, assigned value = 12.0 and σpt = 0.5 in the stated formula to obtain z = +1.0. Translate the table headings into another language and explain that the score is dimensionless while the measurement remains in its original unit. Success criterion: the reviewer can point to the assigned value, participant result, criterion, scope and final status without assuming information absent from the source. Repeat with a different analyte or reporting system to test whether the procedure transfers.
Exercise 2: Do not conflate minus and unsatisfactory
Make two fictional results with z = −1.2 and z = +1.2. Compare the interpretation under a symmetric criterion. Explain why signs describe direction while category depends on the actual scheme’s rule. Success criterion: the reviewer can point to the assigned value, participant result, criterion, scope and final status without assuming information absent from the source. Repeat with a different analyte or reporting system to test whether the procedure transfers.
Exercise 3: Explain z versus z-prime
Prepare a document containing separate formulas and a paragraph describing uncertainty in the assigned value. Translate both labels and use a back-translation check to confirm the distinction survives. Success criterion: the reviewer can point to the assigned value, participant result, criterion, scope and final status without assuming information absent from the source. Repeat with a different analyte or reporting system to test whether the procedure transfers.
Exercise 4: A qualifier check at the boundary
Give a reviewer a source statement with a displayed rounded score and formal provider judgement. Ask whether the translation should reverse the judgement. Require an answer grounded in the provider’s actual criteria, not informal recalculation. Success criterion: the reviewer can point to the assigned value, participant result, criterion, scope and final status without assuming information absent from the source. Repeat with a different analyte or reporting system to test whether the procedure transfers.
Exercise 5: A multi-analyte table
Translate a table with nitrate, lead and arsenic results, each with different units. Make a field ledger and check whether analytes, uncertainty and performance categories stay on the right rows. Success criterion: the reviewer can point to the assigned value, participant result, criterion, scope and final status without assuming information absent from the source. Repeat with a different analyte or reporting system to test whether the procedure transfers.
Exercise 6: Report status exercise
Create three rows: no submission, below reporting limit and not evaluated. Make three distinct translations. Ask a reader whether any row means an actual numerical zero was measured. Success criterion: the reviewer can point to the assigned value, participant result, criterion, scope and final status without assuming information absent from the source. Repeat with a different analyte or reporting system to test whether the procedure transfers.
Exercise 7: Method group interpretation
Show one assigned value for a chromatographic subgroup and another for all methods. Translate the group headings so the participant is compared against the appropriate set. Success criterion: the reviewer can point to the assigned value, participant result, criterion, scope and final status without assuming information absent from the source. Repeat with a different analyte or reporting system to test whether the procedure transfers.
Exercise 8: Technical review simulation
Trade roles: one student translates the narrative, another checks the numerical tables, and a third plays the PT organiser. Produce a revision log and state which disagreements require scientific source correction rather than linguistic editing. Success criterion: the reviewer can point to the assigned value, participant result, criterion, scope and final status without assuming information absent from the source. Repeat with a different analyte or reporting system to test whether the procedure transfers.
Frequently asked questions
What does a z score mean in proficiency testing?
Under a common definition it compares a participant result with an assigned value in units of the proficiency assessment standard deviation. Interpretation depends on the scheme’s criteria; a negative sign alone does not indicate failure. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
Is z′ the same as z?
No. z-prime typically incorporates uncertainty of the assigned value in its denominator. A translator must preserve the prime mark, the provider’s formula and any related decision rule. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
Should translators change decimal commas to decimal points?
Locale formatting may be appropriate, but the numeric value, order of magnitude and unit must remain identical. Apply one consistent number format and independent numerical checking. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
Is a satisfactory PT result the same as accreditation?
No. It is an evaluation within a specific scheme and round. Accreditation concerns a broader assessment against applicable requirements and scope. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
Can a translator correct an arithmetic error?
A translator may flag a suspected mismatch for the scheme organiser. A corrected source report should be issued by the responsible authority before the translation is changed as an official record. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
How should less-than results be handled?
Retain the inequality and reporting-limit meaning. A value such as <0.05 is not the same as 0.05 or zero. Follow the scheme's documented treatment of censored results. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
Do ISO standards prescribe one interpretation for every provider?
Standards guide statistical methods and provider competence, but a scheme’s own stated plan, decision limits and notes must be consulted for the actual report. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
Can AI help review technical translations?
It can flag discrepancies in labels, units or numerical strings. It cannot independently establish the correctness of a scheme-specific judgement or assume access to unpublished technical context. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
How do I translate the term ‘assigned value’?
Use a stable target-language term that means the value designated for proficiency comparison, not an absolute true value. Explain the meaning in a note if the audience is unfamiliar with PT. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
What should an auditor receive?
Provide the source report version, translated version, terminology register, numerical comparison record, query resolutions and bilingual technical approval. That evidence shows how the final language preserves the underlying evaluation. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
Is proficiency testing always numeric?
No. Qualitative and categorical comparisons also exist. The report’s actual observations and evaluation method should determine how the results are translated. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
What is the best first check?
Check the score name, sign, units, assigned value and final provider wording before reading for elegance. These elements usually control whether a translation changes the scientific conclusion. In practice, keep mathematical terms, measurements and interpretation in separate labelled columns before combining them into an English-language explanation. This is much safer than paraphrasing a statistics table in one unstructured sentence.
Final release checklist and reading routes
- The scheme, round, participant and report revision are identified correctly.
- Every measurand, matrix, unit, assigned value, participant result and uncertainty is copied exactly.
- z, z′, zeta, En, sign and score formula remain distinct.
- Score thresholds and category words follow the actual provider’s published rules.
- Qualitative, missing, excluded, censored and not-evaluated states remain distinguishable.
- The provider’s accreditation and the participant’s performance are not conflated.
- Second-person technical review checks tables, graphs, legends, footnotes, conclusions and updates.
- The final translation includes an escalation route for suspected source errors.
Primary reference documents: ISO 13528:2022 — Statistical methods for proficiency testing by interlaboratory comparison and ISO/IEC 17043:2023 — Requirements for proficiency testing providers. The standards are authoritative specifications; this open guide is a teaching explanation, not a substitute for the complete texts or the provider’s scheme instructions. ISO 13528:2022 also lists Amendment 1:2026; consult the current applicable edition and amendments.
Continue to the translation series index, calibration certificates and traceability, Vocabulary Learning Hub and How English Works. For the mathematics underlying a score, study the statistical method; for translation, return to the source’s precise meaning and its intended reader.
The larger lesson: A proficiency-testing report is both a measurement record and a communication about evaluation. Its numbers constrain the words, and its stated assessment rules constrain the conclusion. The most reliable translation lets an independent reader reconstruct both.