Cognitive Decision Strategies Adopted by Trained Judges and Consumers in Reminder Tasks

Abstract

Food manufacturers strive to produce products that are perceived as unchanging in order to maintain sales and keep up with their competitors. Food manufacturers rely on the results obtained from testing trained judges and consumers regarding the potential saleability of their products. This thesis employs some of the tasks available for estimating how well trained judges and consumers perceive differences between similar sample types. However, one needs to determine the decision strategy used by the judge in order to estimate sensitivity. Most of the literature investigating decision strategies for taste stimuli has focused on the same-different task. Five experiments were conducted to determine the decision strategies adopted on the A-Not AR and 2-AFCR tasks for discriminating between confusable stimuli. The A-Not A, 2-AFC, and same-different tasks were used to obtain baseline dʹ values which were compared with those obtained from the reminder tasks. The reminder tasks have more than one strategy that a judge may use (τ and β for A-Not AR, and β/τ and COD for 2-AFCR), each resulting in different dʹ values. The strategy used is assumed to be the one that obtained a dʹ value closest to that of the baseline task. The A-Not A and 2-AFC tasks were used for Experiments 1, 2, and 4, and the same-different task was used for Experiments 3 and 5. Experiment 1 used trained judges to determine the decision strategies adopted for three aqueous solutions which differed in the number of compounds dissolved in filtered water. Experiment 2 used trained judges to determine the decision strategies adopted for Ice Tea, Yoghurt, and Mayonnaise. Experiment 3 used trained judges to determine the decision strategies adopted for Ice Tea (with two levels of sucrose added). Experiment 4 used consumers to determine the decision strategies adopted for Ice Tea and Mayonnaise (similar to Experiment 2). Experiment 5 used consumers to determine the decision strategies adopted for Ice Tea (with two levels of sucrose added, similar to Experiment 3). Trained judges consistently used the A-Not AR β-strategy (6 out of 9 cases) and 2-AFCR β/τ-strategy (8 out of 9 cases). Similarly, consumers consistently used the A-Not AR τ-strategy (6 out of 8 cases) but in Experiment 4 they used the 2-AFCR β/τ-strategy and in Experiment 5 they used the 2-AFCR COD-strategy. There were problems fitting ROC curves to the data obtained from the same-different task for trained judges. The resulting goodness of fit statistics indicated that half of the judges adopted the τ-strategy and the other half the β-strategy making it unclear which strategy was really adopted. Consumer consistently used the τ-strategy for the same-different task. The A-Not AR strategy used was contingent upon the solution type but was consistent across products. The A-Not AR strategy was also contingent upon the type of judge. The A-Not AR task had the best task sensitivity for trained judges (7 out of 9 cases). The task with the best task sensitivity for consumers was variable but the same-different task had better task sensitivity twice. By providing a comparison sample on each trial, the A-Not AR task maybe a reliable alternative to the A-Not A and 2-AFC tasks (which require stimuli familiarisation).