Doctoral Dissertation Research: Using a Prediction and Option Generation Paradigm to Understand and Improve Decision Making

Michigan Technological University, Houghton MI

Investigators

Abstract

In this project the Principal Investigators will examine the effects of skill and decision type on strategy use, and then examine whether skilled decision-making strategies can be trained to become resilient under stress. In many complex domains, as opposed to selecting between explicitly presented alternatives, decision makers are first required to generate alternative options. However, skilled performers use different option generation strategies during perception and action, for instance, when predicting situational events compared to when selecting personal responses. Current research suggests that the accuracy and quality of prediction is positively related to the number of decision alternatives generated, whereas the quality of response selection is negatively related to the number of options generated. The goals of the proposed research are to: (i) reconcile theoretical differences in explanations of strategy use by examining the role of decision type (i.e., prediction, response) and skill level on the adoption of each option generation strategy; (ii) investigate the cognitive mechanisms used to support successful option generation by examining the role of context, (iii) determine the extent to which option generation performance can be improved through training, and (iv) evaluate training transfer to high-stress tasks. The research will employ quasi-experimental and experimental research methods. Using an option generation paradigm, skilled and less-skilled individuals will predict the outcome of, and generate responses to complex, dynamic law enforcement situations presented via video-based and live role-player simulations. An expert model will then be derived from the data and used to train option generation strategies of less-skilled participants during prediction and response. In terms of broader impacts, this research will help refine existing theoretical models of decision-making in complex environments by articulating the skill-based differences in strategies used when deciding upon a course of action versus how a situation will unfold. The ability to predict the future state of the world and to assess the relative merits and consequences of one?s own actions in that context is a requisite skill for many challenging and/or life-threatening domains, such as health care, military, and finance. Accordingly, the ability to improve these skills has the potential to increase the effectiveness, efficiency, and safety of operational environments, both for operators and for citizens that may otherwise be negatively affected by poor decision-making. This research will demonstrate how the analysis of skilled performance can be used to develop critical-thinking-based training methods to improve complex decision-making in stressful environments. The adoption of an evidence-based approach to training decision-making skills will serve as a best-practice example of how to create reproducible and scalable training programs that capture the strategies employed by superior performers.