Proceedings of the Annual Meeting of the Cognitive Science Society

The extent to which people utilize advice from others differs depending on whether the source of the advice is an algorithm or a human. However, no unifying evidence can be used for advice design. Moreover, the use of advice given at intervals (e.g., 70–90%) has not been fully studied. This study proposed a three-step model of the cognitive process of the use of advice with intervals and conducted a simulation and four behavioral experiments. These experiments showed that differences in advice sources affected the cognitive process in which judges decide whether to update their initial judgment based on the advice; this cognitive process was influenced by the relative weight between their initial judgment and the advice interval. These results suggested that for judges to adjust their judgments, designing advice itself is insufficient and advice must be designed according to the relationship between the advice and judge’s judgments.