Verifies that an assay's observed precision and bias meet the manufacturer's claims, using a one-sided chi-square for imprecision and a bias test for trueness.
When you bring an assay in-house you don't re-establish its performance from scratch — you verify the manufacturer's claims on a smaller sample. Two claims matter: is my imprecision no worse than claimed, and is my bias within the allowable limit (trueness).
Imprecision is verified with a one-sided upper chi-square test: given the claimed SD, the observed SD is acceptable unless it is significantly larger. Trueness is verified by comparing the observed mean to the claimed/target value (a paired-t or difference against an allowable bias). This mirrors the CLSI EP15 user-verification protocol.
A column of measurements on a known material, plus the claimed SD (for precision), the claimed/target mean (for trueness), and an allowable bias.
The precision verdict (observed vs claimed SD, chi-square) and the trueness verdict (observed bias vs the claim, with p-value).
"Not verified" on precision means your SD is significantly worse than claimed — investigate before use. Because verification uses few replicates, a pass is reassurance rather than proof; a borderline fail may reflect low power, so consider more replicates before rejecting the method.