In a recent paper, Brockhoff and Schlich (1998, Handling replications in di
scrimination tests. Food Quality and Preference, 9(5), 303-312) proposed a
statistically sound way of handling replications in difference testing. In
the present paper, this new test is applied to the data obtained in six exp
eriments on non alcoholic beverages, where triangle tests were intensively
replicated (between eight and 60 times) with groups of subjects composed of
12-61 students. The paper aims to estimate in these practical situations t
he extent to which a group of panelists is heterogeneous towards the abilit
y of detecting a sensory difference among two products. As the results indi
cate that group heterogeneity was lower than Brockhoff and Schlich first gu
essed, the value of using replicates in difference testing is emphasized. I
t seems that, average over experiments, at least 10 replicates were necessa
ry to properly estimate the level of heterogeneity and that a single subjec
t having done 10 replicates would bring to our test the same amount of info
rmation as five subjects having done the test once only. (C) 1999 Elsevier
Science Ltd. All rights reserved.