Improving the performance of A-Not AR discrimination test using a sensory panel: Effects of the test protocols on sensory data quality

Abstract

The A-Not A with reminder (A-Not AR) test using a two-step rating is an efficient method for measuring the overall sensory differences between multiple products and a reference product. An efficient reference -familiarization procedure and a test protocol that helps the subjects to adopt stable decision criteria are essen-tial for using the A-Not AR test using a two-step rating based on a pooling approach and analyzing the pooled data from different sessions effectively. Two experimental factors for improving the A-Not AR test performance, the familiarization procedure (FP) and reference presentation probability (RPP), were examined using three test protocols following the same pre-test utilizing unspecified tetrad tests. In the Control protocol, the reference was tasted twice as the FP and the RPP was 0.25 with three test products included. In the two Modified protocols, either specified tetrad tests or 2-AFCR tests were used as the FP, both with feedback and re-tasting, and the RPP was 0.5. An independent samples design using three groups (each N = 40) having equal sensitivity was used to test the efficiency of the test protocols over two days. The three groups were formed based on the results of the pre-test. The results of discrimination d' and recognition index d'Rec analyses confirmed that the Modified pro-tocols led to superior test results; the specified tetrad FP was optimal for one-day testing, while the 2-AFCR FP was optimal for repeated testing. Further research will be needed to examine the effects of such comparative task-type FP and 0.5 RPP on the test stability over repetitions and the reliability with different product ranges and different types of pre-test.