Applied decision making with fast-and-frugal heuristics.

Abstract

In applied settings, such as aviation, medicine, and finance, individuals make decisions under various degrees of uncertainty, that is, when not all risks are known or can be calculated. In such situations, decisions can be made using fast-and-frugal heuristics. These are simple strategies that ignore part of the available information. In this article, we propose that the conceptual lens of fast-and-frugal heuristics is useful not only for describing but also for improving applied decision making. By exploiting features of the environment and capabilities of the decision makers, heuristics can be simple without trading off accuracy. Because decision aids based on heuristics build on how individuals make decisions, they can be adopted intuitively and used effectively. Beyond enabling accurate decisions, heuristics possess characteristics that facilitate their adaptation to varied settings. These characteristics include accessibility, speed, transparency, and cost effectiveness. Altogether, the article offers an overview of the literature on fast-and-frugal heuristics and their usefulness in diverse applied settings.

A systematic review of algorithm aversion in augmented decision making

“…This is a serious research agenda in its own right, but the ability to bridge the paradoxical relationship of intuition and rationality is needed so that algorithmic decision aids can pick up people where they stand and make improvements to the decision process that people already follow. That is, in place of requiring people to learn a new process from scratch, one can develop prescriptive aids for intuitive and effective use (Hafenbrädl et al, , p. 217). This means that overcoming algorithm aversion requires carefully examining the subconscious processes that lead up to an intuitive decision to identify the criterion used by human decision makers for gathering and evaluating information under environmental restrictions (Mullins & Rogers, ; Thayer, ).…”

“…This can be conspicuously seen in research that measures decision performance by comparing descriptive results with normative optimality, which has in many ways been considered the gold standard for decision analysis (e.g., Kahn & Baron, ; Lim & O'Connor, ; Sieck & Arkes, ; cf. Brown & Vari, ; Hafenbrädl et al, ). Undoubtedly, the heuristics‐and‐biases program's view of decision making and rationality has added to the algorithmic decision making literature by identifying individuals' cognitive and motivational deficiencies that could benefit from complementary augmentation.…”

“…Although the explicit focus on task structure has seemingly dwindled in recent years, this literature is reminiscent of the fast‐and‐frugal view that human decision makers' decision strategies are contingent on the statistical structures available in the environment (e.g., What calculable risks and alternatives are known?). Put simply, human decision makers often operate in a world of uncertainty (where alternatives, consequences, and probabilities are unknown and optimization is unfeasible) whereas algorithms operate in a world of risk (where probabilities are known or calculable and optimization should be the objective; Hafenbrädl et al, ). The best decision strategy under risk is often not the best decision under uncertainty.…”

“…Existing research points out that organizational and social structures favor the expert intuiter over a cold algorithmic decision maker and incentivize accordingly (e.g., Alexander et al, 2018;Brown, 2015;Eastwood, Snook, & Luther, 2012;Highhouse, 2008b;Klimoski & Jones, 2008;Kuncel, 2008;Önkal et al, 2009). Brown (2015) and Hafenbrädl, Waeger, Marewski, and Gigerenzer (2016) argue that augmented decision making requires extra motivation because it involves combining multiple judgements rather than the acceptance of a single calculation. This means that the successful implementation of algorithmic decision making requires motivating, or incentivizing, human decision makers to utilize algorithmic aids in order to balance the costs of effort with the benefits of decision performance (Christin, 2017).…”

“…This can be conspicuously seen in research that measures decision performance by comparing descriptive results with normative optimality, which has in many ways been considered the gold standard for decision analysis (e.g., Kahn & Baron, 1995;Lim & O'Connor, 1996;Sieck & Arkes, 2005;cf. Brown & Vari, 1992;Hafenbrädl et al, 2016).…”