| Selection and detection problems represent some of the most challenging decision making tasks, especially in the fields of health and medicine. In a population of pregnant women, who is a candidate for a cesarean delivery? Does this mammogram indicate the presence of cancer? Should antibiotics be prescribed for this illness? We must judge cues, make decisions and evaluate feedback, in an uncertain, high stakes environment. This particular type of problem is relevant and timely, especially in the arena of public policy, yet we have limited understanding about how people make decisions in these circumstances.;The purpose of this investigation is multifold. In the following pages, I examine the literature, identifying principles and modeling rules based on the large volume of prior research. This review covers tool based analysis of judgment and decision making, explores the relationships between feedback and accuracy, feedback and confidence and feedback and implicit learning, while untangling the results of years of model building on cue judgment and threshold learning.;I identify the key elements of decision maker behavior, based on an analysis of recent experimental data, applying, wherever possible, those elements to a current problem in public policy and model decision making: the significant rise in cesarean deliveries. I present a simple model that combines error correction and hill-climbing principles, which provides a good match to empirical data, as well as demonstrating that it effectively seeks an optimal threshold in an uncertain environment.;I conclude by offering some general conclusions about what we know about this sizable intellectual landscape, and what additional insight has been drawn from this modeling investigation. In particular, I note that this research has shed some light on why the cesarean rate has risen so dramatically, and why it is a direct and very reasonable result of the decision making environment. And I will offer some prescriptions, based on model based exploration, and supported by the literature and empirical testing, about where we are at, and what we still need to uncover, in order to make better decisions... |
