The discriminant test allows to demonstrate whether two products are different or not according to a general difference or to a specific attribute.

Objectives

• Verification of tiny changes in the overall product sensory quality (overall difference) or in specific characteristics (specific difference) following to modifications to the production chain like: ingredients, process variables, packaging, storage, etc.
• Determination of the presence of off-flavour and/or of contaminants (e.g., soaps, compounds deriving from the package, etc.).
• Selection, training, and monitoring of judges.
• Selection of products to submit to further analyses (e.g., consumer tests).

The FEM Sensory Laboratory has gained over the years a great expertise in the application of numerous tests, like: A-nonA, triangle test, duo-trio, two out of five, Harris-Kalmus, pairwise comparison (single or multiple), ranking.

Triangle method (UNI EN ISO 4120)

In every triangle test, 3 samples are presented of which 2 are the same and one is different.

E.g., considering sample A and sample B, the judge is presented with 3 samples ordered according one out of the 6 possible combinations (A-A-B, A-B-B, A-B-A, B-B-A, B-A-A, B-A-B). The judge should pick the different samples out (single).

Typical Triangle Test

Every test consists in the sensory evaluation of the perceivable difference between 2 products. Two samples are thus compared: sample A, which is generally the reference product vs. sample B, which is the developed product (e.g., different in the concentration of a same ingredient or in the place of origin or in the supplier). Usually, an analysis session consists in the performance of 2-4 triangle tests with breaks between each other. This allows to compare the reference product with 2-4 variants. The rationale behind the number of comparisons to perform during a session depends on the product of interest, as a function of which different degrees of sensory fatigue are produced.

The complete discriminant test consists in two testing sessions (analysis and replicate) with 2-4 triangle tests performed by a panel of 20-30 members trained in the method applied. As a function of the responses collected, the existence of perceivable sensory differences between samples can be statistically proved.

Panel reliability is also evaluated by presenting the judges during a control session with some reference products (bought or prepared in laboratory) having a matrix similar to that of the product of interest

Product samples

It is customer’s responsibility to identify and supply the products to investigate. Additionally, the customer is in charge of guaranteeing controlled conditions during each product sample preparation and storage. Both product quantity and evaluation conditions will be agreed with the customer.

Example: The QUALITY CONTROL OF TRENTINGRANA CHEESE project.