On Amelioration Of Human Cognitive Biases In Binary Decision Making

Abstract

We study the behavior of cognitively biased humans in decision making under the binary hypothesis testing framework. Rationality of humans is modeled via prospect theory, which utilizes a value function and a weight function to characterize humans’ distorted perception of costs and probabilities. Following psychology studies which suggest that humans make decisions by choosing an alternative that yields the minimum expected costs based on received evidence, we show that a cognitively biased human employs an likelihood ratio test (LRT) in the prospect theoretic hypothesis testing framework. We further propose three approaches to ameliorate the side effects of cognitive biases and help humans make higher quality decisions. Simulations are presented to corroborate the theoretical results.