The use of multiple correspondence analysis for the evaluation of inter-laboratory comparisons

Participation in inter-laboratory comparisons (ILC) is one of the recommended means of external quality control according to ISO/IEC 17025:2005. Providers of ILC or proficiency test (PT) schemes collect, besides the measurement results on the test samples, information on the sample treatment and measurement procedure. The objective of this paper is to evaluate in a non-traditional way, using numerical and non-numerical data provided by the participants in IMEP-20 (trace elements in tuna fish), the additional information concerning the applied analytical methods and the accreditation/nomination status. Arsenic was taken as an example. The basic statistical procedure for the evaluation of questionnaire information was the multiple correspondence analysis (MCA). Univariate clustering techniques were applied for the categorization of the numerical data (measurement values). The methodology of the evaluation of supplementary non-numeric information used in this paper might serve (a) to providers of ILC (PT) schemes to modify/improve their questionnaires and (b) to give laboratories better guidance in the methods used for the determination of various analytes in different matrices. This paper is meant serve as a guide for the possible interpretation of the questionnaires accompanying ILC schemas. Presented at the 3rd International Conference on Metrology, November 2006, Tel Aviv, Israel