Practical example of the application of the simple triangular test ("forced choice" option)*

A manufacturer who changes the formulation of his product wishes to know whether the new product will be similar to the old from the sensory point of view. He has at his disposal 12 assessors.

Two batches are prepared, one representing the old formulation (A) and one the new formulation (B).

Presuming that each assessor makes only one assessment, it will be necessary to prepare 18 test samples of A and 18 test samples of B in two series of six sets, as follows:

two sets ABB
two sets AAB
two sets ABA
two sets BAA
two sets BBA
two sets BAB

These sets are distributed at random among the assessors.

The test supervisor chooses a significance level of 5%, i.e., he accepts a risk of 5% of concluding that the test will reveal a difference when there is none.

The number of correct answers (8) is referred to the table.

With 12 replies, it is concluded that the two products are different at the 5% significance level.

* From ISO 4120-1983(E)