Do neither by default. First establish whether the laboratories tested comparable specimens using the same defined method and reporting basis. Then have qualified reviewers examine the difference under the agreed dispute procedure. Equipment age and a simple average do not resolve incompatible tests or uncertain sample identity.
Match specimen identity, preparation and named method before comparing numbers
Obtain both complete reports and the specimen-submission records. Compare sample identifiers, production origin, tested property, method version, preparation and any deviations reported by the laboratories. “Same grade” is not necessarily “same specimen” or “same production lot.”
Check whether one report concerns a separate coupon while the other concerns a finished insert. Ask the laboratories to explain the relevant distinction rather than drawing your own metallurgical conclusion. The carbide testing guide provides background on test interpretation; this case needs a precise comparison of the two actual measurement records.
Review calibration, uncertainty and any reporting conversions with the labs
Compare units, scales, conversions, reported uncertainty and any stated limitations. If a result was converted for a summary, obtain the original reported value and basis for conversion. Do not average values that do not represent the same defined quantity under comparable conditions.
Ask each laboratory to review the discrepancy against the same neutral comparison sheet. Include the ordered acceptance criterion and avoid leading questions that demand a preferred answer. A laboratory may identify a reporting error, method difference or need for additional investigation; retain its explanation as issued.
Have the responsible technical authority assess method suitability and the significance of the difference. Calibration information and competence evidence can matter, but neither a familiar laboratory name nor a modern instrument automatically decides which result should govern the order. Use the agreed evidence and dispute framework.
Use the agreed dispute procedure and qualified retest instead of averaging incompatible results
If a retest is needed, agree the laboratory, specimen selection, custody, method and decision rule before testing. Clarify whether the new work is investigating the discrepancy or serving as the contractual referee result. Those purposes should not be confused after the outcome is known.
For the carbide blade material review, hold the affected material decision until the authorized reviewer can support a disposition. Keep the original reports, laboratory responses and any retest together. Do not request that an inconvenient report be withdrawn unless its issuer has identified a genuine error and documents the correction.
The final explanation should state what caused or resolved the disagreement, or acknowledge what remains uncertain. If the purchase terms do not specify how conflicting results are settled, refer that gap to the contract owner and qualified technical adviser. An unsupported average is not a substitute for an agreed basis of acceptance.
NIST makes uncertainty part of measurement traceability. NIST Metrological Traceability: FAQ and Policy ↗
