Compatibility effects in the prescriptive application of psychological heuristics.

Abstract

This paper examined how people use prescriptive decision aids based on fast-and-frugal principles. It showed that the effectiveness of a decision tool depends not only on model recommendations, but on the fit between the person, the task and the decision strategy being recommended. The work highlights an important principle for analytics and AI-enabled decision support – better decisions do not come from better models alone. They come from tools that are interpretable, cognitively usable and well matched to the context in which decisions are made. I discuss the implications of these findings for the development of prescriptive tools informed by pragmatic behavioural considerations.

Shashwat M. Pande

Data, Models, Decisions.

My research interests include behavioural decision-making, experiment design, data science and interactions between humans, models and decision-making environments involving risk, uncertainty and/or ambiguity.