Discrete confidence levels revealed by sequential decisions

Whether humans are optimal decision makers is still a debated issue in the realm of perceptual decisions. Taking advantage of the direct link between an optimal decision-making and the confidence in that decision, we offer a new dual-decisions method of inferring such confidence without asking for its explicit valuation. Our method circumvents the well-known miscalibration issue with explicit confidence reports as well as the specification of the cost-function required by ’opt-out’ or post-decision wagering methods. We show that observers’ inferred confidence in their first decision and its use in a subsequent decision (conditioned upon the correctness of the first) fall short of both an optimal and an under-sampling behavior and are significantly better fitted by a model positing that observers use no more than four confidence levels, at odds with the continuous confidence function of stimulus level prescribed by a normative behavior. Significance statement A normative decision behavior requires that one’s confidence be a continuous function of the difficulty of decision to be made (the posterior probability that the decision was correct given the evidence). People however concede that they could not discriminate between two indefinitely close confidence states. Using a new experimental paradigm that does not require participants’ explicit confidence evaluation nor knowledge of their utility function, we show that people can discriminate only two and at most four confidence states, the implication of which is that they are not optimal decision-makers. This finding counters a plethora of studies suggesting the contrary and also rejects an alternative stand according to which people’s decisions are based on a sample of an available posterior function.