Home > Education, Metacognition > “I totally nailed it and I am pretty sure I did better than most people”- The Pitfall of Overconfidence

“I totally nailed it and I am pretty sure I did better than most people”- The Pitfall of Overconfidence

Have you ever been disappointed by your exam score when you thought you actually did pretty well on it? Or have you ever overestimated how sufficiently you have prepared for a test and panicked as you read through the actual exam and found questions more difficult than expected? If you have had these experiences, you have been a victim of overconfidence effect.

Although we hardly realize such errors or often feel reluctant to admit them, we are all familiar with the mismatch between self-evaluation and actual outcomes. This phenomenon is called the overconfidence effect, a cognitive bias that occurs when people inaccurately evaluate their own performance as above average or higher in accuracy or quality than it actually is.

Overestimation of Capacity                    [https://advanced-hindsight.com/blog/b-e-dogs-overconfidence/]

People have faith in their erroneous self-evaluation about a variety of topics, including but not limited to application of factual knowledge, as in a college exam scenario. Psychologists have found that people tend to position themselves above others when assessing their own capacity. Overconfidence is explicit not only in self-estimation about skills like safe driving but also in self-positioning within a community when participants see themselves as more popular and sociable than their friends (Svenson, 1981; Sanbonmatsu et al., 2016; Zuckerman & Jost, 2001).    

So, why are we so easily biased towards ourselves? What happens in our cognitive system when we generate overconfident conclusions about ourselves? To answer these questions, we need to first understand how people use metacognition to perform a self-evaluation task. Metacognition is the cognitive process situated beyond our cognition, in other words, it is the mechanism we use to be aware of and monitor our own mental processes (Flavell, 1979). In the context of exam performance, metacognition functions well if a student considers his performance poor and later gets a poor grade just as expected. However, if a student feels pretty confident about his performance but later gets a grade lower than expectation, metacognition has failed to function accurately and as a result overconfident self-evaluation is adopted by the cognitive system as the final conclusion. Such undesired mismatch is also known as the Dunning-Kruger Effect, and occurs when people fail to reject erroneous judgments based on preconceptions of themselves.

Metacognition as a patrolling guard [www.iconfinder.com]

Now we know that the major cause of overconfidence effect is the failure of metacognition, but you might start to wonder why our metacognition is so unstable. To further understand what prevents metacognition from functioning effectively, try to consider metacognition as a patrol guard in the building of our cognitive system. The guard cannot be at all locations of the system simultaneously and it takes time for him to patrol around and spot problems. Moreover, the guard needs resources like a flashlight and toolboxes to locate and fix problems. Unfortunately, there are other workers in the building who share the resources with the guard and need to use the tools quite frequently to keep the building functioning normally.

Let’s now apply this metaphor to a real scientific study. Psychologists have studied metacognition by asking participants to operate a car-driving simulator while talking on their cell phone, and later evaluate the errors they made during driving (Sanbonmatsu et al., 2016). Multitasking requires a great amount of attentional resources, metaphorically, requires a lot of busy workers running around using the tools. As a result, the guard has no available resource to carry out inspections and therefore records only a few or none of the errors made by the driver. When the guard needs to report to the system during self-evaluation, he reports way less errors than there actually were, leading to the participants being overconfident about their multitasking capacity and evaluating their driving performance as better than reality.

In addition, it is time-consuming for metacognition to find and report what is consistent with the truth and this characteristic lowers its efficiency. Again, let’s think in terms of a real-life example. When a student makes self-evaluation on his actually poor test performance, he faces a selection between two answers: “I did great,” and “I did poorly.” The student finds it reasonable for him to perform well on the test because he has reread the textbook many times at the night before exam. Some psychologists would consider the students as having an “illusion of validity” since he truly believes in his consistent story even it is not the truth according to scientific studies (Kahneman, 2011; Putnam et al., 2016). Such belief creates an accessibility disparity between the two answers “I did great,” and “I did poorly.” This is because our cognition is, overall, biased to favor consistent patterns and therefore also the answer that perfectly fits into a consistent story since we rely on consistent patterns for perception and interpretation of environmental stimuli (Kahneman, 2011). As a result, the answer consistent with the individual’s belief comes to mind more quickly and is more easily accepted. Back to the patrol guard metaphor, overconfidence occurs when the guard finally arrives with the answer “I did poorly” at the desk of the decision-making judge in the student’s cognitive system only to be told that the result “I did great” was handed in by a worker and adopted a few minutes ago. Or, the judge looks at both answers together and rejects “I did poorly” because it is inconsistent with the judge’s biased knowledge. In general, the difficult procedure carried out by the metacognition guard is called controlled processing, into which a lot of attention and resources must be invested, while the report of answer “I did great” to the judge is called automatic processing, which is simple, favored by the superior, and happens within a blink (Posner & Snyder, 1975).

With all this said, people’s confidence does not predict accuracy. Confidence is just a feeling we sense when we reach conclusions using what comes to mind the most quickly and effortlessly. The unconscious reluctance to engage in metacognitive processes can result in undesired outcomes in various fields, leaving people confused about where things went wrong. For instance, observers considering themselves as experienced authorities can be highly confident about their prediction accuracy even though they are only slightly more accurate than random guess (Kahneman, 2011). People are also commonly overconfident when asked to recall emotionally salient events such as the Sepetember 11 attack. False memories occur frequently due to suggestibility and misinformation, yet this does not influence our confidence level at all since we all implicitly believe that we cannot possibly misremember or forget events that are so sensational, especially when we can pull up a lot of details from our memory.

Overconfidence about Memory Accuracy [https://dilbert.com]

It seems like we are innately susceptible to overconfidence effect because cognition relies heavily on knowledge and memory database to function. In fact, we will experience a cognitive breakdown if we lack enough confidence in our own decision and actions (Dunning et al., 1990). An adequate confidence level ensures our timely reaction to the environment and coherent behaviors in task performance. However, overconfidence can lead to severe undesirable outcomes such as problematic gambling, irrational financial investments, and poor predictions about exam performances as we have discussed in great detail above (Camchong et al., 2007; Hirshleifer & Luo, 2001; Putnam et al., 2016).

We can Resolve Overconfidence [https://www.semfionetworks.com/blog/problem-is-not-a-negative-word]

After we have examined why and how overconfidence effect influences our lives, we are able to find resolutions that help us control over such instinct by maintaining our metacognition more active and alert. Feedback is a crucial factor since it effectively activates metacognitive processes (Putnam et al., 2016). When a student carefully reads through all feedbacks given by the professor on the exam, he is engaging in a metacognitive self-inspection process, making corrections about mistakes made by himself. Good feedbacks helps students gain a more accurate evaluation of how well they master material and therefore modify inaccurate predictions and assumptions. Therefore, try to study and review through recalling and testing yourself instead of rereading. These practices involve feedback and eliminate overconfidence. Questioning our general assumptions more frequently can also help us avoid overconfidence. For instance, search for popular science articles before you decide to make an investment and stop assuming yourself as especially different from the general public. The more frequently you consult your metacognition, the more aware you will become of its reliability. Also remember to have enough mental resource available for the patrol guard to spot and treat problems. We should be aware of the limitations of our cognitive capacity and avoid multitasking to ensure the quality of our performances.

Good news is that right now you have kept your metacognition patrol guard working actively for ten minutes or so just by reading through this blog post. Avoiding the pitfall of overconfidence is possible as long as we are willing to think actively about our own thought processes.



Dunning, D., Griffin, D. W., Milojkovic, J. D., & Ross, L. (1990). The overconfidence effect in social prediction. Journal of personality and social psychology58(4), 568.

Flavell, J. H. (1979). Metacognition and cognitive monitoring: A new area of cognitive developmental inquiry. American Psychologist, 34(10), 906-911. doi:10.1037/0003-066X.34.10.906

Hirshleifer, D., & Luo, G. Y. (2001). On the survival of overconfident traders in a competitive securities market. Journal of Financial Markets, 4(1), 73-84.

Kahneman, D. Don’t Blink! The Hazards of Confidence, The New York Times Magazine, October 19, 2011.

Posner, M. I., Snyder, C. R., & Solso, R. (2004). Attention and cognitive control. Cognitive psychology: Key readings, 205.

Putnam, A. L., Sungkhasettee, V. W., & Roediger III, H. L. (2016). Optimizing learning in college: tips from cognitive psychology. Perspectives on Psychological Science11(5), 652-660.

Sanbonmatsu, D.M., Strayer, D.L., Biondi, F., Behrends, A.A., and Moore, S.M. (2016). Cell phone use diminishes self-awareness of impaired drivingPsychonomic Bulletin & Review, 23, 617–623. doi:10.3758/s13423-015-0922-4.

Svenson, O. (1981). Are we all less risky and more skillful than our fellow drivers?. Acta psychologica47(2), 143-148.

Zuckerman, E., and Jost, J. (2001). What Makes You Think You’re So Popular? Self-Evaluation Maintenance and the Subjective Side of the “Friendship Paradox”. Social Psychology Quarterly, 64(3), 207-223.

  1. December 9th, 2019 at 23:45 | #1

    Hi Yan, I loved reading your post and learning more about the overconfidence effect. I was particularly interested in this bias’ connection and similarities to the naïve cynicism bias. Naive cynicism is when we people don’t see bias in themselves but expect everyone else to act in an egotistical way. In my opinion the overconfidence bias is kind of like the broader umbrella turm under which naive cynicism works, since we are overconfident that our ideas are of higher quality because we believe we are not biased.

You must be logged in to post a comment.