In this study, the Temporal Dominance of Sensations (TDS) method and its derived multi-intake TDS method were used to measure the change in taste and flavor when solid foods differing in taste between the outside and inside are eaten continuously. The former approach was used to describe the major sensory perceptions, and the latter was used for evaluating the sensation of continuing to eat food by continuing the TDS evaluation for each bite multiple times. The samples used were rice crackers with a spicy soy sauce-flavored seasoning liquid attached to the surface and fried peanuts with salt attached to the surface. When the sample was placed in the mouth, the surface taste and flavor followed by the internal taste and flavor were perceived, and the sensory stimulation became the dominant sensation near the end of swallowing. Furthermore, it was possible to capture the accumulation of sensory stimuli the sample continued to be eaten. This method allowed us to capture the changes in taste and flavor from the time the sample was placed in the mouth to the time it was swallowed, and to capture all the sensations of continued eating.

Triangle test is conducted for similarity testing to verify if there are (or are not) any detectable changes between the products. As per ISO4120, before performing a triangle test, it is necessary to decide four parameters, i.e., alpha risk, power, sample size and proportion of discriminator (p_d) for a rational sensory evaluation. However, p_d is inferred based on experiences or business strategies, thus, it is difficult to precisely decide p_d. In this study, we tried to introduce the tau criteria for consumers’ acceptable psychological differences using the same-different test with sureness judgement and degree of difference testing. Moreover, correlation between consumer and in-house panels’ sensitivities using Thurstonian d-prime (d′) value was evaluated. Fifteen pairs of various soup samples were used for the consumer tests (N＝80–100) and discrimination test by the in-house panels (N＝20–74). The tau values of consumer and Thurstonian d′ of both consumer and in-house panel were calculated. The mean consumer tau value was 0.93. The regression analysis comparing the d′ values between consumer and in-house panel elucidated that the in-house panel was 1.27-fold more sensitive than the consumers. Therefore, a fewer number of in-house panelists are sufficient to estimate the results of the consumers’ acceptability for the difference. Once the tau criteria of consumer and the sensitivity of the in-house panel are known, it would be possible to introduce more rapid and consumer-relevant designs for similarity testing.