GB/T 39558-2020 Active

This standard specifies a method for determining whether there are sensory differences between two types of products. This method is suitable for determining differences in a single or multiple sensory attributes.
In sensory analysis, the "A"-"Non-A" test can be used for:
a) Difference testing: particularly suitable for samples with differences in appearance (it is difficult to prepare identical appearance of repeated samples) or aftertaste (differences in aftertaste between samples make direct comparison difficult);
b) Identification testing: particularly suitable for determining whether evaluators or evaluation groups can identify new stimuli related to known stimuli (e.g., identifying the sweetness quality of new sweeteners);
c) Sensation testing: used to determine evaluators' ability to distinguish stimuli.
The "A"-"Non-A" test is not suitable for determining whether two types of products are similar enough to be interchangeable (e.g., for similarity testing), because the essence of the "A"-"Non-A" test is that all evaluators repeatedly evaluate the same product, and such repeated evaluation violates the basic assumption that similarity testing is statistically valid.
Application examples are provided in Appendix A.
Note: Bi and Ennis pointed out that regardless of how the repeated evaluation is implemented in the test, the difference between the "A" and "Non-A" samples, d′, remains constant, but the variance of d′ depends on the specific implementation of the repeated evaluation. This standard does not discuss the Thurstonian analysis and test validity of the "A"-"Non-A" method. Interested readers can refer to Reference [1] for a detailed discussion of this topic.