Statistical aspects of proficiency testing in analytical laboratories. Part 2. Testing for sufficient homogeneity

Abstract

The effective homogeneity of the test material distributed to participants in a proficiency test is essential to its successful operation, but demonstration of that homogeneity can amount to a major proportion of the costs of the scheme. Here it is shown that there are often no satisfactory alternatives to the method of test for sufficient homogeneity recommended in the Harmonised Protocol. However, it is argued that the criterion for demonstrating sufficient homogeneity is too stringent. In addition, the required precision for the analytical method used in the homogeneity study is clarified.