QA/QC, Duplicates, Blanks, and Cross-Checks
QA/QC, Duplicates, Blanks, and Cross-Checks
This domain tests the quality system behind each result.
Core rules
- Follow a written QA/QC plan. Typical field QC includes about 10% duplicates and 5% field blanks, plus lab spikes and annual cross-checks with a reference lab.
- Duplicates must be exposed in the same location and time; side-by-side placement is not enough if one is near a draft.
- Field blanks travel to the site unopened and are analyzed to detect contamination or shipping exposure. Blank results should be below the lab’s reporting limit or the protocol’s blank threshold.
- Review duplicate precision, such as relative percent difference (RPD), against the protocol criterion. Investigate failures before releasing results.
- CRMs require calibration verification, background checks, and documented response to tamper or alarm events.
Worked example
For 100 charcoal detectors, deploy at least 10 duplicates and 5 field blanks. If a duplicate pair reads 4.2 and 6.8 pCi/L, RPD = |4.2−6.8| ÷ ((4.2+6.8)/2) × 100 ≈ 47%, which exceeds a 20% precision criterion. Do not simply average the pair-trigger a QA investigation.
Traps
Skipping blanks; placing duplicates in different rooms; ignoring failed QC; using a CRM without current calibration; treating a lab’s detection limit as zero.