Organization and evaluation of interlaboratory comparison studies among southern african water analysis laboratories

Interlaboratory comparison studies provide a useful "external" supplement to the various "internal" quality-control procedures which must be employed in the water-analysis laboratory in order to maintain a high degree of reliability in the production of analytical results. Such studies have been made regularly for many years by various overseas organizations, and were introduced in South Africa in 1976. Since that time 16 studies have been made, involving more than 40 laboratories. In this paper the various factors involved in the successful organization of interlaboratory comparison studies are discussed, and details are given of the sample-preparation, analysis-instruction and result-reporting procedures used in the southern African studies. Techniques used for the statistical evaluation of the analytical results submitted are described and discussed, for example, methods for the rejection of outliers, measures of accuracy and precision, determination of total error, tests of significance, Greenberg's assessment technique, Madden's ranking technique and Youden's graphical technique. A brief review is given of the studies made to date, along with specific findings and recommendations arising from them. The need for a recognized updated set of standard methods for use by water and wastewater analysis laboratories in southern Africa is highlighted.