We suggest that overconfidence (conscious or unconscious) is motivated in part by strategic considerations, and test this experimentally. We find compelling supporting evidence in the behavior of participants who send and respond to others’ statements of confidence about how well they have scored on an IQ test. In two-player tournaments where the higher score wins, a player is very likely to choose to compete when he knows that his own stated confidence is higher than the other player’s, but rarely when the reverse is true. Consistent with this behavior, stated confidence is inflated by males when deterrence is strategically optimal and is instead deflated (by males and females) when luring (encouraging entry) is strategically optimal. This behavior is consistent with the equilibrium of the corresponding signaling game. Overconfident statements are used in environments that seem familiar, and we present evidence that suggests that this can occur on an unconscious level.

We investigate how the selection process of a leader affects team performance with respect to social learning. We use a laboratory experiment in which an incentivized guessing task is repeated in a star network with the leader at the center. Leader selection is either based on competence, on self-confidence, or made at random. In our setting, teams with random leaders do not underperform. They even outperform teams with leaders selected on self-confidence. Hence, self-confidence can be a dangerous proxy for competence of a leader. We show that it is the declaration of the selection procedure which makes non-random leaders overly influential. To investigate the opinion dynamics, we set up a horse race between several rational and naïve models of social learning. The prevalent conservatism in updating, together with the strong influence of the team leader, imply an information loss since the other team members’ knowledge is not sufficiently integrated.

Bayesian updating remains the benchmark for dynamic modeling under uncertainty within economics. Recent theory and evidence suggest individuals may process information asymmetrically when it relates to personal characteristics or future life outcomes, with good news receiving more weight than bad news. I examine information processing across a broad set of contexts: (1) ego relevant, (2) financially relevant, and (3) non value relevant. In the first two cases, information about outcomes is valenced, containing either good or bad news. In the third case, information is value neutral. In contrast to a number of previous studies I do not find differences in belief updating across valenced and value neutral settings. Updating across all contexts is asymmetric and conservative: the former is influenced by sequences of signals received, a new variation of confirmation bias, while the latter is driven by non-updates. Despite this, posteriors are well approximated by those calculated using Bayes’ rule. Most importantly these patterns are present across all contexts, cautioning against the interpretation of asymmetric updating or other deviations from Bayes’ rule as being motivated by psychological biases.

This paper reports findings of a laboratory experiment, which explores how reported self-assessment regarding the own relative performance is perceived by others. In particular, I investigate whether overconfident or underconfident subjects are considered as more likeable, and who of the two is expected to win in a tournament, thereby controlling for performance. Underconfidence beats overconfidence in both respects. Underconfident subjects are rewarded significantly more often than overconfident subjects, and are significantly more often expected to win. Subjects being less convinced of their performance are taken as more congenial and are expected to be more ambitious to improve, whereas overconfident subjects are rather expected to rest on their high beliefs. While subjects do not anticipate the stronger performance signal of underconfidence, they anticipate its higher sympathy value. The comparison to a non-strategic setting shows that men strategically deflate their self-assessment to be rewarded by others. Women, in contrast, either do not deflate their self-assessment or do so even in non-strategic situations, a behavior that might be driven by non-monetary image concerns of women.

We aim to test the hypothesis that overconfidence arises as a strategy to influence others in social interactions. To address this question, we design an experiment in which participants are incentivized either to form accurate beliefs about their performance at a test, or to convince a group of other participants that they performed well. We also vary participants’ ability to gather information about their performance. Our results show that participants are more likely to (1) overestimate their performance when they anticipate that they will try to persuade others and (2) bias their information search in a manner conducive to receiving more positive feedback, when given the chance to do so. In addition, we also find suggestive evidence that this increase in confidence has a positive effect on participants’ persuasiveness.

To the known causes of overconfidence in decisions and judgments, we reveal another source that derives from a bias during the act of decision making. While this bias, the predecisional distortion of information, is well studied, its impact on overconfidence is not. We demonstrate how the distortion of information creates overconfidence in those professionals often regarded as singularly overconfident, entrepreneurs. When these professionals use a sequence of relevant information to make an accept-reject decision about a business opportunity, a cycle of confidence-distortion-confidence builds unjustified confidence in the chosen action – and does so whether that action is to accept or reject the venture. Overconfidence is a well-recognized cause of flawed decision making. Our work demonstrates the paradoxical converse of this claim, that flawed decision making can be a cause of overconfidence.

In this project, we manipulate the public observability of forecasts and outcomes of a physical task. We explore how these manipulations affect overconfidence (OC). Participants in the experiment are asked to hold a weight after predicting how long they think they could do it for. Comparing the prediction and outcome times (in seconds) yields a measure of OC. We independently vary two dimensions of public observability (of the outcome and of the prediction). Additionally, we manipulate incentives to come up with an accurate prediction. This design allows us to shed light on the mechanism behind male and female OC. Following the existing literature, we formulate several hypotheses regarding the differences in predictions and outcomes for males and females in the presence of the public observability of predictions and outcomes. Our experimental data do not provide support to most of the hypotheses: in particular, there is no evidence of a gender gap in overconfidence. The most robust finding that emerges from our results is that incentives on making correct predictions increase participants’ forecasts on their own performance (by about 24%) and their actual performance as well, but to a lower extent (by about 8%); in addition, incentives to predict correctly in fact increase error for females (by about 33%).

Overconfidence plays a role in a large number of individual decision biases and has been considered a ‘meta-bias’ for this reason. However, since overconfidence is measured behaviorally with respect to particular tasks (in which performance varies across individuals), it is unclear whether people generally vary in terms of their general overconfidence. We investigated this issue using a novel measure: the Generalized Overconfidence Task (GOT). The GOT is a difficult perception test that asks participants to identify objects in fuzzy (‘adversarial’) images. Critically, participants’ estimated performance on the task is not related to their actual performance. Instead, variation in estimated performance, we argue, arises from generalized overconfidence, that is, people indicating a cognitive skill for which they have no basis. In a series of studies (total N = 1,293), the GOT was more predictive when looking at a broad range of behavioral outcomes than two other overestimation tasks (cognitive and numeracy) and did not display substantial overlap with conceptually related measures (Studies 1a and 1b). In Studies 2a and 2b, the GOT showed superior reliability in a test–retest design compared to the other overconfidence measures (i.e., cognitive and numeracy measures), particularly when collecting confidence ratings after each image and an estimated performance score. Finally, the GOT is a strong predictor of a host of behavioral outcomes, including conspiracy beliefs, bullshit receptivity, overclaiming, and the ability to discern news headlines.

People often overestimate their understanding of how things work. For instance, people believe that they can explain even ordinary phenomena such as the operation of zippers and speedometers in greater depth than they really can. This is called the illusion of explanatory depth. Fortunately, a person can expose the illusion by attempting to generate a causal explanation for how the phenomenon operates (e.g., how a zipper works). This might be because explanation makes salient the gaps in a person’s knowledge of that phenomenon. However, recent evidence suggests that people might be able to expose the illusion by instead explaining a different phenomenon. Across three preregistered experiments, we tested whether the process of explaining one phenomenon (e.g., how a zipper works) would lead someone to report knowing less about a completely different phenomenon (e.g., how snow forms). In each experiment, we found that attempting to explain one phenomenon led participants to report knowing less about various phenomena. For example, participants reported knowing less about how snow forms after attempting to explain how a zipper works. We discuss alternative accounts of the illusion of explanatory depth that might better fit our results. We also consider the utility of explanation as an indirect, non-confrontational debiasing method in which a person generalizes a feeling of ignorance about one phenomenon to their knowledge base more generally.

How do emerging technologies affect nuclear stability? In this paper, we use a quasi-experimental cyber-nuclear wargame with 580 players to explore three hypotheses about emerging technologies and nuclear stability: (1) technological uncertainty leads to preemption and escalation; (2) technological uncertainty leads to restraint; and (3) technological certainty leads to escalation through aggressive counterforce campaigns. The wargames suggest that uncertainty and fear about cyber vulnerabilities create no immediate incentives for preemptive nuclear use. The greater danger to strategic stability lies in how overconfidence in cyber exploits incentivizes more aggressive counterforce campaigns and, secondarily, how vulnerabilities encourage predelegation or automation. Both of these effects suggest worrisome relationships between cyber exploits and inadvertent nuclear use on one hand and cyber vulnerabilities and accidental nuclear use on the other hand. Together, these findings reveal the complicated relationship between pathways to escalation and strategic stability, highlighting the role that confidence and perhaps-misplaced certainty—versus uncertainty and fear—play in strategic stability.

Survey respondents over-forecast inflation: they expect it to be higher than itturns out to be. Furthermore, people are generally overconfident in theirforecasts. In two experiments, we show that providing outcome feedback thatinforms people of the actual level of the inflation that they have forecastreduces both over-forecasting and overconfidence in forecasts. Theseimprovements were preserved even after feedback had been withdrawn, a findingthat indicates that they were not produced because feedback had a temporaryincentive effect but because it had a more permanent learning effect. However,providing forecasters with more outcome feedback did not have a greater effect.Feedback appears to provide people with information about biases in theirjudgments and, once they have received that information, no additional advantageis obtained by giving it to them again. Reducing over-forecasting also had noclear effect on overall error. This was because providing outcome feedback afterevery judgment also affected the noise or random error in forecasts, increasingit by a sufficient amount to cancel out the benefits provided by the reductionin over-forecasting.

In a series of three experiments, participants made inferences about which one of a pair of two objects scored higher on a criterion. The first experiment was designed to contrast the prediction of Probabilistic Mental Model theory (Gigerenzer, Hoffrage, & Kleinbölting, 1991) concerning sampling procedure with the hard-easy effect. The experiment failed to support the theory’s prediction that a particular pair of randomly sampled item sets would differ in percentage correct; but the observation that German participants performed practically as well on comparisons between U.S. cities (many of which they did not even recognize) than on comparisons between German cities (about which they knew much more) ultimately led to the formulation of the recognition heuristic. Experiment 2 was a second, this time successful, attempt to unconfound item difficulty and sampling procedure. In Experiment 3, participants’ knowledge and recognition of each city was elicited, and how often this could be used to make an inference was manipulated. Choices were consistent with the recognition heuristic in about 80% of the cases when it discriminated and people had no additional knowledge about the recognized city (and in about 90% when they had such knowledge). The frequency with which the heuristic could be used affected the percentage correct, mean confidence, and overconfidence as predicted. The size of the reference class, which was also manipulated, modified these effects in meaningful and theoretically important ways.

The calibration of probability or confidence judgments concerns the association between the judgments and some estimate of the correct probabilities of events. Researchers rely on estimates using relative frequencies computed by aggregating data over observations. We show that this approach creates conceptual problems, and may result in the confounding of explanatory variables or unstable estimates. To circumvent these problems we propose using probability estimates obtained from statistical models—specifically mixed models for binary data—in the analysis of calibration. We illustrate this methodology by re-analyzing data from a published study and comparing the results from this approach to those based on relative frequencies. The model-based estimates avoid problems with confounding variables and provided more precise estimates, resulting in better inferences.

Right-wing ideology and cognitive ability, including objective numeracy, have been found to relate negatively. Although objective and subjective numeracy correlate positively, it is unclear whether subjective numeracy relates to political ideology in the same way. Replicating and extending previous research, across two samples of American adults (ns= 455, 406), those who performed worse on objective numeracy tasks scored higher on right-wing authoritarianism (RWA) and social dominance orientation (SDO), and they self-identified as more conservative on general, social, and economic continua. Controlling for objective numeracy, subjective numeracy related positively to measures of right-wing ideologies. In other words, those who strongly (vs. weakly) endorsed right-wing ideologies believed they are good with numbers yet performed worse on numeracy tasks. We discuss implications for the opposing direction of associations between ideology with objective versus subjective numeracy and similarities with literature on overconfidence.

Many natural decisions contain an element of skill. Modern conceptions of the skill component include control (Goodie, 2003) and competence (Heath & Tversky, 1991). The control hypothesis states that a task's skill component (the sensitivity of the task to skill) affects decision making; the competence hypothesis states decision making is affected only if the participant possesses the skill. Three experiments compared risk taking patterns between two groups. One group faced bets on random events, and another group faced bets on their answers to general knowledge questions, which is a task characterized by control. In Experiment 1, control increased risk taking markedly with all statistical properties held constant. In Experiment 2, decisions made in domains of varying difficulty, and by individuals of varying ability, yielded further qualified support for the role of competence. In Experiment 3, the role of control was replicated, and participants’ perceptions of the differences in group treatments aligned more with the implications of the control hypothesis than with the competence hypothesis. Results offered support for the control hypothesis across a range of competence.

This research investigated how different forms of overconfidence correlate with age. Contrary to stereotypes that young people are more overconfident, the results provide little evidence that overestimation of one’s performance or overplacement of one’s performance relative to that of others is correlated with age. Instead, the results suggest that precision in judgment (confidence that one knows the truth) increases with age. This result is strongest for probabilistic elicitations, and not present in quantile elicitations or reported confidence intervals. The results suggest that a lifetime of experience, rather than leading to better calibration, instead may increase our confidence that we know what we’re talking about.

Subjects who judged speed in a driving scenario overestimated how fast they could decelerate when speeding compared to when keeping within the speed limit (Svenson, 2009). The purpose of the present studies were to replicate studies conducted in Europe with subjects in the U.S., to study the influence of speed unit (kph vs. mph), affective reactions to outcome (collision) and identity of main actor (driver) on braking speed judgments. The results replicated the European findings and the outcome affective factor (passing a line/killing a child) and the actor factor (subject/driver in general) had significant effects on judgments of braking speed. The results were related to psychological theory and applied implications were discussed.

Overprecision is the most robust type of overconfidence. We present a new method that significantly reduces this bias and offers insight into its underlying cause. In three experiments, overprecision was significantly reduced by forcing participants to consider all possible outcomes of an event. Each participant was presented with the entire range of possible outcomes divided into intervals, and estimated each interval’s likelihood of including the true answer. The superiority of this Subjective Probability Interval Estimate (SPIES) method is robust to range widths and interval grain sizes. Its carryover effects are observed even in subsequent estimates made using the conventional, 90% confidence interval method: judges who first made SPIES judgments considered a broader range of values in subsequent conventional interval estimates as well.