What Creates Measurement Risk?
Measurement risk is driven by the relationship between three things:
- Measurement uncertainty
- The specification limit or acceptance limit
- The decision rule used (Pass/Fail – or conditional Pass/Fail)
If uncertainty is small relative to tolerance, risk is low. If uncertainty approaches the tolerance limit, risk rises sharply.
Additionally, as the measurement approaches the acceptance limits, the risk rises even as the measurement uncertainty stays the same.
This is why a calibration result without a clearly stated uncertainty is incomplete — and potentially misleading.
Two Direct Types of Measurement Risk
Measurement risk is commonly expressed using two statistical probabilities:
Probability of False Accept (PFA)
The risk of accepting a device or product that is actually out of tolerance.
Probability of False Reject (PFR)
The risk of rejecting a device or product that is actually within tolerance.
Both matter, but they affect organizations differently.
Real-World Statistical Examples
Example 1: False Accept (PFA)
A pressure gauge has a tolerance of ±1.0 psi.
The calibration result is 0.8 psi from nominal, with a measurement uncertainty of ±0.6 psi (k=2).
Although the reported value is inside tolerance, part of the uncertainty interval extends beyond the tolerance limit.
Statistical analysis with standard distribution shows there is a 20–25% probability of false accept — meaning one out of every four gauges is actually out of tolerance, even though it “passed” calibration.
Impact:
- Risk of process drift
- Potential nonconforming product
- Product recalls
- Undetected quality issues
Example 2: False Reject (PFR)
A torque wrench has a tolerance of ±4 lbf-in.
The measured value is 3.6 lbf-in from nominal, with a measurement uncertainty of ±1.5 lbf-in (k=2).
Here, the uncertainty overlaps both sides of the tolerance band. Applying a conservative decision rule may result in rejection.
Statistical analysis with standard distribution shows there is a 30% probability of false reject — meaning nearly one-third of rejected tools may actually meet specification.
Impact:
- Unnecessary recalibration or repair
- Increased downtime
- Higher operating costs
Why Quality Managers Should Care
If your calibration provider does not actively manage measurement risk:
- Passing instruments may still be unreliable, creating customer failures/recalls
- Good instruments may be removed from service unnecessarily, costing extra money
- Audit findings may lack defensible technical justification
- Quality decisions become guesswork instead of data-driven
A calibration certificate alone does not control risk. Uncertainty analysis and decision rules do.
How Measurement Risk Is Properly Determined
A competent calibration laboratory evaluates risk by:
- Developing a complete uncertainty budget
- Comparing uncertainty to actual instrument tolerances
- Applying defined statistical decision rules (such as guard banding)
- Considering real-world use conditions, not just bench testing
- Clearly documenting assumptions and limitations
This approach aligns with ISO/IEC 17025 and modern risk-based quality systems.
What to Look for in a Calibration Supplier
When evaluating a calibration provider, ask:
- Do they report measurement uncertainty, not just pass/fail?
- Can they explain PFA and PFR in practical terms?
- Do they uncertainty as it relates to your instrument’s specification (TUR)?
If the answers are unclear, your measurement risk probably is too.
Bottom Line
Measurement risk is not theoretical — it is statistical, measurable, and manageable. Quality managers who understand false accept and false reject risk make better calibration decisions, reduce risk, reduce cost, reduce waste, and protect product integrity.
A calibration provider’s true value is not in the certificate — it is in how well they control measurement risk.