Quantitative descriptive analysis developed in 1974 aims to minimise bias by careful choice of conditions and to provide a description of test products which is meaningful and reproducible. By comparison, the profile method is considered to use too few tasters, does not apply statistics and lacks reproducibility from laboratory to laboratory. In choosing test conditions for the descriptive analysis method developed by the authors the following are considered: minimisation of variability, number of tasters, repeating of tests, training of panelists and minimising bias. In selecting meaningful descriptors these can be selected from existing flavour terminology especially the international terminology ('flavour wheel') developed by EBC/ASBC/MBAA. The terms can be selected on known technological and physiological principles and should be tested on the panel using production samples which show clearly the characteristic to be demonstrated. The best terms selected allow the greatest degree of discrimination between closely similar samples. Finally the selection of a suitable scale is described. The scales can vary depending on the purpose of the test. Where overall opinion is important, free scaling, for example a line scale, is preferred whilst if reproducibility is predominant a heavily structured and anchored scale is favoured.
Keywords : analysis beer flavour quality sensory statistical terminology