Research

This paper experimentally investigates the role of confidence sensitivity and meta-confidence in career-related decision-making. While the effects of overconfidence on economic outcomes have been widely studied, the implications of confidence sensitivity - the ability to distinguish between accurate and inaccurate performance - have been largely overlooked. Yet, in various contexts, a lack of sensitivity can be just as detrimental as overconfidence, as it undermines individuals’ ability to assess their abilities accurately. Additionally, individuals may hold beliefs about their own confidence sensitivity, and they might use such “meta-confidence” beliefs to guide their behavior. In our experiment, participants performed a task and then chose between a competitive or noncompetitive compensation scheme. In addition, some participants were given the option to seek performance feedback before committing to a choice. Our results show that confidence sensitivity significantly improves the quality of compensation scheme decisions, while low meta-confidence is associated with increased feedback-seeking behavior prior to the compensation decision. We provide causal evidence for this latter finding: a treatment condition designed to reduce meta-confidence led to an increased willingness to seek feedback. Overall, this paper underscores the richness of individuals’ beliefs about their abilities and demonstrates how an in-depth analysis of these beliefs, using tools from cognitive psychology, can offer new insights into economic decision-making.

The ability to form judgments about one’s own cognitive performance—known as metacognition—has been widely studied across domains. More recently, a few studies have shown that individuals can also evaluate the quality of their own metacognition, an ability referred to as meta-metacognition. Yet despite its promising implications, meta-metacognition remains a poorly defined construct. Here, we contribute to this emerging theme with a novel experimental paradigm that offers a precise definition and incentivization of meta-metacognitive judgments, enables direct comparisons between metacognitive and meta-metacognitive levels, and does so using observable data alone—without relying on unobserved model parameters. In a first experiment, participants judged which of two perceptual decisions was more likely to be correct (a metacognitive decision), and then rated their confidence in this judgment (a meta-metacognitive evaluation). In a second experiment, we incorporated trial-by-trial probabilistic confidence ratings and a counterfactual urn task to disentangle the information available at the metacognitive and meta-metacognitive levels. Crucially, we show that meta-metacognitive judgments predict the accuracy of metacognitive decisions beyond what is captured by participants’ probabilistic confidence ratings alone. Moreover, we identify systematic inconsistencies between confidence ratings and metacognitive decisions—inconsistencies that improve decision accuracy—revealing that metacognitive evaluations draw on multiple, dissociable sources of information. Together, these findings provide direct evidence for meta-metacognitive insight, while challenging the assumption that the metacognitive level is functionally unified.

Many studies report that following feedback, individuals do not update their beliefs enough (a conservatism bias), and react more to good news than to bad news (an asymmetry bias), consistent with the idea of motivated beliefs. In the literature on conservatism and asymmetric updating, however, only one prior study focuses on judgments on absolute performance (Grossman & Owens, 2012), which finds that belief updating is well described by the Bayesian benchmark in that case. Here, we set out to test the replicability of these results and their robustness across several experimental manipulations, varying the uncertainty of participants’ priors, the tasks to perform, the format of beliefs and the elicitation rules used to incentivize these beliefs. We also introduce new measures of ego-relevance of these beliefs, and of the credibility of the feedback received by participants. Overall, we confirm across various experimental conditions that individuals exhibit no conservatism and asymmetry bias when they update their beliefs about their absolute performance. As in Grossman & Owens (2012), most observations are well-described by a Bayesian benchmark in our data. These results suggest a limit to the manifestation of motivated beliefs, and call for more research on the conditions under which biases in belief updating occur.

Perceptual confidence has been an important topic recently. However, one key limitation in current approaches is that most studies have focused on confidence judgments made for single decisions. In three experiments, we investigate how these local confidence judgments relate and contribute to global confidence judgments, by which observers summarize their performance over a series of perceptual decisions. We report two main results. First, we find that participants exhibit more overconfidence in their local than in their global judgments of performance, an observation mirroring the aggregation effect in knowledge-based decisions. We further show that this effect is specific to confidence judgments and does not reflect a calculation bias. Second, we document a novel effect by which participants' global confidence is larger for sets which are more heterogeneous in terms of difficulty, even when actual performance is controlled for. Surprisingly, we find that this effect of variability also occurs at the level of local confidence judgments, in a manner that fully explains the effect at the global level. Overall, our results indicate that global confidence is based on local confidence, although these two processes can be partially dissociated. We discuss possible theoretical accounts to relate and empirical investigations of how observers develop and use a global sense of perceptual confidence.

Although conflicts in bargaining have attracted a lot of attention in the literature, situations in which bargainers have to share the product of their performance have been less commonly investigated empirically. Here, we show that overplacement leads to conflict in these situations: individuals overestimate their contribution to the joint production and consequently make unreasonable claims. We further decompose overplacement into three types of cognitive biases: overestimation of one’s own production (i.e. overconfidence bias), underestimation of others’ production (i.e. superiority bias) and biases in information processing. We show that they all contribute to overplacement. To quantify these biases, we develop a novel experimental setting using a psychophysically controlled production task within a bargaining game, where we elicit participants’ subjective estimation of their performance, both before and after they receive information about the joint production. In addition, we test several interventions to mitigate these biases, and successfully decrease disagreements and overplacement through one of them. Our approach illustrates how combining psychophysical methods and economic analyses could prove helpful to identify the impact of cognitive biases on individuals’ behavior.