Premium
The two‐factor triangle test
Author(s) -
HARRIES J. M.,
SMITH G. L.
Publication year - 1982
Publication title -
international journal of food science and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.831
H-Index - 96
eISSN - 1365-2621
pISSN - 0950-5423
DOI - 10.1111/j.1365-2621.1982.tb00172.x
Subject(s) - mathematics , statistics , binomial distribution , set (abstract data type) , binomial (polynomial) , interpretation (philosophy) , selection (genetic algorithm) , test (biology) , quality (philosophy) , econometrics , computer science , artificial intelligence , paleontology , biology , programming language , philosophy , epistemology
Summary The triangular test, commonly used in sensory tests of food quality, is generally interpreted on the assumption that the probability of a correct selection is the same for all assessors, and independent for separate trials leading to the use of the binomial distribution in setting up tests of significance. Only when there is no real discrimination is it true that using one taster n times is the same as using n tasters once, the basis for the column heading ‘number of tasters or tastings’ in published tables of significance. This paper attempts to remove this assumption by the use of the beta‐binomial distribution, and suggests an interpretation in which the chance of success is different between and within assessors so that a two‐factor analysis becomes possible. A set of data relating to the flavour of fish has been reinterpreted in these terms.