Experiment 1: Difference tests. The first experiment
was split into six separate tests, conducted on three different
days with different subjects participating and different
closures compared each day: tests 1 and 2 compared
screwcap and natural cork, tests 3 and 4 compared
screwcap and synthetic cork, and tests 5 and 6 compared
natural cork and synthetic cork. The triangle test design
was selected because the probability of choosing the odd
sample by chance is 0.3, whereas the probability for other
three-sample reference difference tests is 0.5. The triangle
test minimized the probability that any differences were
due to chance or guessing and it achieved greater test
power with fewer participants. Even though this test has
been identified as more difficult for the taster (Lawless and
Heymann 1999), an additional palette cleansing rinse was