Course:FNH200/Lessons/Lesson 03/Page 03.4

3.4 Discriminative or Difference Tests

Figure 3.5. Example of a triangle test set-up. In this test panelists receive three coded samples. They are told that two samples are the same and one is different and are asked to identify the “different” sample. Analysis of the results is based on the probability that the different sample will be identified by chance one-third of the time.

Discrimination or “Difference” tests are used to determine whether a difference exists between samples. The difference can be defined or it may not. These tests would be used to evaluate if a new processing treatment, yeast type or aging treatment has changed the character of the food product. Panelists’ personal likes and dislikes are not a concern.

Difference tests are the backbone of sensory analyses. They allow the experimenter to document the presence of perceived differences among samples for quality control, product development and/or research and development purposes.

Examples of commonly used different tests include: triangle tests, pair-difference tests. Figure 3.5. shows an example of a triangle test set-up.

Descriptive Analysis Tests

Descriptive analysis requires detection, description and quantization of the sensory aspects of a product. It is used only with trained panelists. These trained panelists must be trained for several weeks in order to be able to be "calibrated" and accurately detect, describe and rate the intensity of each attribute (whether it is appearance, aroma, flavour, mouth feel, etc.). Examples of descriptive analysis tests include: Flavour profile method, quantitative descriptive analysis (QDA^TM), and free-choice profiling.

Hedonic/Preference Tests

Preference tests are also known as consumer tests. The objective of these tests is to evaluate a personal (subjective) response to a product. Consumers can give their preference between products, degree of liking of a product, or their overall acceptance of a product. Preference tests require a large number of panelists (100s to 1000s) in order to represent the target population for the product being tested. Examples of preference tests include: Paired-preference, ranking and hedonic scales.