Confidence reports in decision-making with multiple alternatives violate the Bayesian confidence hypothesis

Abstract

Decision confidence reflects our ability to evaluate the quality of decisions and guides subsequent behavior. Experiments on confidence reports have almost exclusively focused on two-alternative decision-making. In this realm, the leading theory is that confidence reflects the probability that a decision is correct (the posterior probability of the chosen option). There is, however, another possibility, namely that people are less confident if the best two options are closer to each other in posterior probability, regardless of how probable they are in absolute terms. This possibility has not previously been considered because in twoalternative decisions, it reduces to the leading theory. Here, we test this alternative theory in a three-alternative visual categorization task. We found that confidence reports are best explained by the difference between the posterior probabilities of the best and the next-best options, rather than by the posterior probability of the chosen (best) option alone, or by the overall uncertainty (entropy) of the posterior distribution. Our results upend the leading notion of decision confidence and instead suggest that confidence reflects the observer's subjective probability that they made the best possible decision.

Bayesian confidence in optimal decisions

“…That such models can account for a wide range of patterns observed in previous research, and in this study, indirectly supports the view that confidence is Bayesian in the ongoing debate about this claim (Adler & Ma, 2018;Li & Ma, 2020;Meyniel et al, 2015;Navajas et al, 2017;Peters et al, 2017;Sanders et al, 2016).…”

Predictive processing as a systematic basis for identifying the neural correlates of consciousness

“…In this discussion we have focused on the promising explanatory and unificatory aspects of PP. However, PP, although increasingly pervasive, is certainly not universally accepted and there are conflicting findings and perspectives, driving ongoing debate (e.g., Alilović et al, 2019;Cao, 2020;Li & Ma, 2020;Litwin & Miłkowski, 2020;Rahnev & Denison, 2018). For example, one recurring challenge to PP is that worry that an imperative to minimize sensory prediction errors will lead to agents that do nothing at all -the so-called 'dark room problem' (Friston et al, 2012;Sun & Firestone, 2020a, 2020b.…”

Human visual motion perception shows hallmarks of Bayesian structural inference

“…Beyond the reported motion structure, we asked if our participants’ subjective confidence was dependent on the stimulus ambiguity, as measured by the Bayesian model. The relation between experimentally reported subjective confidence and theoretically derived Bayesian confidence has been studied for a variety of tasks (Drugowitsch, Moreno-Bote, and Pouget, 2014; Galvin et al, 2003; Hangya, Sanders, and Kepecs, 2016; Kepecs and Mainen, 2012; H.-H. Li and Ma, 2020; Mamassian, 2016; Pouget, Drugowitsch, and Kepecs, 2016; Sanders, Hangya, and Kepecs, 2016), and the literature is often equivocal about the exact nature of the relation. We analyzed the participants’ reported confidence as a function of the Bayesian predicted confidence, p ( S | X ), where the prior P ( S ) is given by the fitted biases b S for each participant, and the structure S is evaluated at the human choice (Pouget, Drugowitsch, and Kepecs, 2016).…”