ABSTRACT
E-mail prioritization involves placing all of the ‘useful’ or ‘good’ unread e-mails at the top of the inbox, and all of the bad ones at the bottom. We use two cognitive decision models—a rational model, which considers all of the available information, and a fast and frugal model that uses one reason decision making—to prioritize e-mails. Experimental results, using real data obtained by unobtrusively logging e-mail user behavior, show that the fast and frugal model is just as effective as the rational model. The results also show that a Bayesian approach to learning is superior to the standard frequentist approach, because it balances the competing demands of exploration and exploitation in finding good e-mails. We use the results to draw some applied conclusions about the development of an e-mail prioritization system, and note some theoretical implications of the results for the cognitive modeling of human decision making in general.
