REPLICATED TRIANGLE TESTS - EFFECT OF FEEDBACK AND PRODUCT COMPARISONON PERFORMANCE

The feed-back procedure consists of informing assessors of the quality of their response (correct or incorrect) after each triangle test. Th is procedure showed a contrasted effect according to assessors' detect ion abilities. A decrease in performance was observed for assessors wh o had low detection abilities whereas a slight increase was observed f or assessors who had higher detection abilities. When the feed-back pr ocedure is followed by a comparison by tasting of the two products, th e increase in performance from the beginning to the end of the session is larger than that of both feed-back and control (no information) gr oups. We suggest that information inference from previous tests, which allows assessors to perform better at subsequent tests, is facilitate d by the presence of a double information (feed-back and product compa rison). This effect might be due to an attentional process allowing as sessors to focus their attention on the difference between the two pro ducts, An attempt was made to model learning effect by variation of at tentional weights.