Some statistical procedures for analytical method accuracy tests and estimation.

In 1974 the National Institute for Occupational Safety and Health (NIOSH) and the Occupational Safety and Health Administration joined to complete exposure standards promulgated by federal regulations. In that effort NIOSH scientists developed an accuracy criterion (AC) and a statistical protocol for evaluating fulfillment of that AC by an analytical method. This article extends that foundation and proposes a new approach to accuracy analyses. It concentrates on the case of known bias, but attempts to bridge the procedures from that case to one in which the bias is estimated. The article emphasizes a general and flexible approach to the design and analysis of more informative and effective method accuracy studies. These empower the user/investigator to design and analyze studies to be most useful and informative for specific requirements.