Consumer preferences of coffees in relation to sensory parameters as studied by analysis of covariance

Eight different coffees have been evaluated for liking by consumers of different sexes and agegroups and from Jive different European countries. Analysis of variance showed that Sample, Country, Sex, Sample-Country and Country-Agegroup contributed most to the liking scores. In a subsequent analysis, the 7 dr's from the nominal variable Sample were replaced by the samples' coordinates on seven continuous variables (each with 1 df) denoting a 7-dimensional sensory space. this space was obtained from five European coffee panels and is described in an accompanying paper. The results from this analysis showed that the first four dimensions as well as the interaction of the first dimension with Country contributed to the liking scores. These results were visualised in a biplot. (C) 1998 Elsevier Science Ltd. All rights reserved.