A design for interlaboratory quality-control programs.

Abstract

A design for interlaboratory quality-control programs is described. Speimens are despatched in pairs to participating laboratories. The results returned by laboratories are compared with each other and with reference results, but the frequent brief reports which are prepared are not regarded as the main laboratory assessments, although these reports can alert laboratories to gross imprecision and inaccuracy. When sufficient results have been returned, a linear regression analysis is carried out between results from each laboratory and the reference results. The statistics obtained from the regression data provide a concise source of information about the form of inaccuracy (imprecision and systematic error) present.