Reducing Error: the Dunning-Kruger Effect

How good are you at dancing or singing? Without any sort of objective measure, a common method to try and determine our ability is by using peer comparison. We watch those around us participating in similar activities and estimate our relative performance. The only problem, as discovered by research psychologists David Dunning and Justin Kruger, is that when we use peer comparison to judge our own performance we are most often wrong, and sometimes by a significant margin.

How good are you at dancing or singing? Without any sort of objective measure, a common method to try and determine our ability is by using peer comparison. We watch those around us participating in similar activities and estimate our relative performance. The only problem, as discovered by research psychologists David Dunning and Justin Kruger, is that when we use peer comparison to judge our own performance we are most often wrong, and sometimes by a significant margin.

In arguably a seminal study, Dunning and Kruger asked 141 students that had just completed an exam to estimate their relative performance. What they discovered was that students that performed the worst, in the bottom 25% of the class, believed they had outperformed the majority of their peers. In fact, some students actual performance placed them in the 12th percentile, yet they estimated their mastery of the material fell in the 60th percentile. As actual performance improved students demonstrated they were able to better estimate their ranking, and then something interesting occurred, students that ranked the highest in actual performance slightly underestimated their ability (figure 1).

Multiple studies across a variety of tasks and domains confirmed this same pattern, establishing a cognitive bias now referred to as the Dunning-Kruger effect. Whether in academics, chess, nursing, or social abilities like humor, low skill performers grossly overestimate their performance relative to their peers, while high skill performers slightly underestimate.

---

Explaining the Error

​In determining why we make these kinds of errors in judging our own performance, Dunning and Kruger offer two explanations. For low skill individuals, or novices, the inability to accurately estimate performance is a form of meta-ignorance. The novice is unaware that they are unaware. They don’t and can’t know what they don’t know. Therefore, when they estimate their ability they are operating from an incomplete body of knowledge. A reasonable example is the novice driver. They know the rules of the road and they have been driving for a few months so they conclude that they are just as capable as the next driver. Still, the novice has never experienced driving in icy conditions or having an animal jump out in front of their vehicle while traveling at high speed. There are a variety of situations not covered by the rules of the road that are unknowable to the novice and therefore cannot be factored into their estimate. 
​
For those of high skill, or experts, the explanation is slightly different. It is still at type of ignorance, but it is in believing that others are just as capable. Having achieved a high level of performance, the expert doesn’t see what they have accomplished as anything special. Therefore, when asked to rank themselves in comparison to their peers they slightly underestimate their true ability as they fail to recognize that others are less capable. The thought process becomes, “If I can do it, so can everyone else.”

Implications

Given the findings, what then are the implications? In other words, so what if we might be horrible at singing or dancing, but think that we are wonderful? In cases where perceived consequences are low, having a false sense of being better than we actually are might actually be beneficial. Not recognizing you are bad might have an affective component, thereby providing the necessary motivation to continue working on improving your ability oblivious to your actual talents. There is something to be said for ignorance as a state of bliss.

On the other hand, while misjudging dancing or singing probably does not have any dire consequences, what about for driving or swimming? What about in domains such as nursing or emergency first responders? In these cases, believing you are much more capable at a particular task than you actually are can be deadly. The novice driver that suddenly hits a patch of ice for the first time might find that misjudging their true ability comes at a high price. 

For high skill individuals implications are more subtle. Typically speaking, slightly underestimating your true ability will not generally result in a negative outcome. The main downside is failing to progress or reach one’s full potential as the expert underestimates their capability. In situations of high consequence however, this might be beneficial in ensuring the expert does not place themselves in an overtly dangerous situation well beyond their limits. 

The more deadly combination is that of low and high skill individuals working on some shared task or trying to achieve a mutual goal that has potentially high consequences. For example, a low and high skill swimmer see an island in the distance. The low skill swimmer incorrectly estimates they can make it and the high skill swimmer is incapable of understanding why anyone would not be able to make it. The expert unwittingly encourages the novice, they begin swimming for the island, and tragedy potentially follows.

Reducing the Dunning-Kruger Effect

​Given that the major challenge in addressing the Dunning-Kruger effect is that we don’t know what we don’t know, this can be a tough situation to resolve, but there are several potential ways to avoid or minimize the effects, including:

1. Rejecting the use of peer comparison. As an alternative, establish objective measures and solicit frequent feedback regarding your performance. This is a meta-decision, whereby you actively scaffold your learning, comparing your performance today against your previous performance. Using self-comparison negates peer comparison.
2. Next, break down a goal into the necessary sub-tasks, focusing on specific rather than general abilities. Research has shown that the more specific a task, the harder it is to maintain an illusion of being better than we actually are. For example, it is easier to make the general claim that you are great at using tools for home repair, than it is to make the more specific statement that you are awesome with a power saw.
3. Last, seek out qualified mentors. Note, these are not individuals within your peer group, even those considered relative experts. These are credentialed professionals that can accelerate the learning process, providing formal instruction that includes objective and timely feedback. This allows you to make a quicker transition from novice to intermediate, and then advance towards becoming an expert. 

​In summary, the Dunning-Kruger effect is a form of cognitive bias that is largely a result of using peer comparison. When we compare ourselves to others, we risk the consequences of making inaccurate judgments of our own abilities. This is most likely not a significant issue when the perceived consequences are low. But, for those environments or situations with potentially high consequences, you want to account for the Dunning-Kruger effect and use techniques or methods that can help you to minimize the use of peer or social comparison.
​
References

Dunning, D. (2011). The Dunning–Kruger effect: On being ignorant of one's own ignorance. In Advances in experimental social psychology (Vol. 44, pp. 247-296). Academic Press.

Dunning, D., Heath, C., & Suls, J. M. (2004). Flawed self-assessment: Implications for health, education, and the workplace. Psychological science in the public interest, 5(3), 69-106.

Dunning, D., Johnson, K., Ehrlinger, J., & Kruger, J. (2003). Why people fail to recognize their own incompetence. Current directions in psychological science, 12(3), 83-87.

Kennedy, E. J., Lawton, L., & Plumlee, E. L. (2002). Blissful ignorance: The problem of unrecognized incompetence and academic performance. Journal of Marketing Education, 24(3), 243-252.
​
Kruger, J., & Dunning, D. (1999). Unskilled and unaware of it: how difficulties in recognizing one's own incompetence lead to inflated self-assessments. Journal of personality and social psychology, 77(6), 1121.