A problem that frequently occurs in metrology is one of assessing compatibility of data obtained by a user laboratory with the specified values and uncertainty estimates from the certificate of analysis. The users data are summarized by a measurand mean value and its coverage interval, which is typically based on a repeatability standard deviation, but could include errors estimated using a Type B approach. If the users interval and the certificate interval do not overlap or more generally when the conformance hypothesis is rejected, the user may seek guidance on how to confirm this fact or how to rectify the nonconformity. The suggested two-stage statistical approach allows for lab to obtain its coverage interval of the same width as the certificate, or to get a guaranteed power compatibility test for the given bias. Practical computationally simple formulas for each stage sample size are provided.
Citation: Accreditation and Quality Assurance
Pub Type: Journals
Certificate reference value, Conformance testing, Coverage interval, Critical bias, Necessary sample size, Noncentral t-distribution, Stein procedure.