A Note on Semi-Probabilistic Grading

This note describes the results of using a semi-probabilistic grading system to determine students' grades in a graduate course in quantitative methods. That is, the students are required to make policy decisions to handle traditional sorts of problems, such as inventory and queuing problems. The points a student receives are some known function of his firm's profits for a particular period, the period having been simulated for each student's firm in accordance with a prespecified probability distribution, given the student's recommendations. In addition, he receives a grade on the assignment based on its a priori quality rather than its ex post success. This note describes the results of using this system, results which in turn lead to some interesting speculations about decision-making and decision makers in general. In particular, it is suggested that even persons well aware of sophisticated procedures for making a rational choice among alternatives will not necessarily take advantage of these procedures, and further, that luck plays a more significant role in success than one might be inclined to concede.