IUPAC/CITAC guide: interlaboratory comparison of categorical characteristics of a substance, material, or object (IUPAC Technical Report)

This Guide is intended for harmonization of interlaboratory comparisons of categorical − nominal (qualitative, i.e., non-quantitative) and ordinal (semi-quantitative) − characteristics of a substance, material, or object. It provides guidance for application of relevant methods of mathematical statistics for design of such interlaboratory comparisons and analysis of the obtained data, when the methods developed for continuous quantitative values (e.g., ANOVA − analysis of variance) cannot be used without violation of their basic assumptions. The proposed approach employs recently-developed two-way nominal analysis of variation CATANOVA and two-way ordinal analysis of variation ORDANOVA. The Guide also addresses correlation between the categorical characteristics, as well as correlation between these characteristics and the chemical composition of the material or object. A multisensory quality index of a product, combining information on its categorical characteristics, is detailed. It allows for comparison of the quality of the same material produced by different producers. The examples provided in the Guide are from the fields of macroscopic examination of weld imperfections, comparison of odor intensity of drinking water, and comparison of sensory (ordinal) characteristics of a sausage. A corresponding calculation tool with an Excel spreadsheet including macros, and programs written in the R environment, are available in the specified references.