The following essay is reprinted with permission 
The Conversation is an online publication covering the latest research.
British comedian John Cleese once summarized the idea of the Dunning Kruger effect as follows: “If you’re really, really stupid, it’s impossible to know that you’re really, really stupid.” There are dozens of headlines that tie everything down to why he was elected president.
As a math professor who teaches students to leverage data to make informed decisions, I am well aware of the common mistakes people make when working with numbers. The Dunning-Kruger effect is the idea that the least skilled people overestimate their abilities more than anyone else. This sounds convincing on the surface and makes for excellent comedy. However, in a recent paper, my colleagues and I suggest that the mathematical approach used to show this effect may be wrong.
What Dunning and Kruger showed
In the 1990s, Cornell University psychology professors David Dunning and Justin Kruger wanted to test whether incompetent people were unaware of their own incompetence.
To test this, they gave 45 undergraduates a 20-question logic test and asked them to rate their performance in two different ways.
First, Dunning and Kruger asked students to estimate the number of questions they answered correctly. This is a very simple evaluation. Dunning and Kruger then asked the students to estimate how they performed compared to other students who took the test. This type of self-assessment requires students to make assumptions about the performance of others and exposes them to common cognitive mistakes. Most people consider themselves better than average.
Studies show that 93% of Americans believe they are better than average drivers, and 90% of teachers believe they are more skilled than their peers, and this overestimation includes logic tests. Spread across many skills. However, it is mathematically impossible for most people to perform better than average on any given task.
After giving the students a logic test, Dunning and Kruger divided the students into four groups based on their scores. In the lowest scoring quarter of the students, on average he got 10 out of 20 questions correct. In contrast, her 1st-percent-percent student got an average of 17 questions correct. Both groups estimated that they got about 14 questions correct. This is not a terrible self-assessment for either group. The least skilled people overestimated their scores by about 20 points, while the top performers underestimated their scores by about 15 points.
The results look more striking when looking at how students rated themselves compared to other students. Here the better than average effect is fully visible. The lowest-scoring student estimated that he scored better than her 62% of test-takers, while the highest-scoring student thought he scored better than her 68%. .
By definition, being in the bottom 25% means that at best you score better than 25% of people, and on average you score better than 12.5%. Estimating that he scored 12.5% better than his colleague’s girlfriend, even though his grades were 62% better than his colleague’s girlfriend, overestimated him by a whopping 49.5 percentage points. Become.
The Dunning-Kruger effect is a measure of how students compare themselves to other students, not their actual scores. This grossly exaggerates the overestimation of the bottom 25%, and seems to indicate that, as Dunning and Kruger titled their paper, the least-skilled students are “low-skilled and low-perception.” .
Using the protocol formulated by Dunning and Kruger, many researchers have since “confirmed” this effect in their respective fields of study, suggesting that the Dunning-Kruger effect is inherent in how the human brain works. It gave me a sense of being there. To ordinary people, the Dunning-Kruger effect seems real. Because overly arrogant fools are a familiar and annoying stereotype.
Debunking the Dunning-Kruger Effect
Dunning and Kruger’s analysis is misleading for three reasons.
The worst test takers overestimate their performance the most simply because they are the furthest from a perfect score. Moreover, like most people, even the least skilled believe themselves to be better than average. Finally, even those with the lowest scores do not have significantly worse objective performance ratings.
To prove that the Dunning-Kruger effect is a product of research design rather than human thought, my colleagues and I showed that it can be produced using randomly generated data.
First, we created 1,154 fictitious people and randomly assigned test scores and self-rating rankings relative to their peers.
We then divided these fake people into quarters based on their test scores, similar to what Dunning and Kruger did. Self-assessment rankings are also randomly assigned scores from 1 to 100, so each quarter they return to an average of 50. By definition, a bottom quarter outperforms her on average by only 12.5% of participants, but from a random assignment of self-assessment scores that she considers herself to be better than her 50% of test-takers. This would give her a 37.5 percentage point overestimation if no humans were involved.
The last point, proving that even less-skilled people can judge their own skills, required a different approach.
My colleague Ed Nouffer and his team gave students a 25-question scientific literacy test. After answering each question, students rate their performance on each question as ‘well done’, ‘not sure’, or ‘not sure at all’.
Working with Nuhfer, we found that novice students are pretty good at estimating their own abilities. In the study of lower-quarter-performing novice students, only 16.5% of her students significantly overestimated their abilities. It turns out that 3.9% significantly underestimated the score. This means that her nearly 80% of unskilled students were able to estimate their actual abilities fairly accurately. This contrasts sharply with Dunning and Kruger’s idea that inexperienced students consistently overestimate their skills.
dunning kruger today
The original article by Dunning and Kruger begins with the quote, “One of the essential characteristics of incompetence is the inability of the victim of such damage to recognize his or her incompetence.” . This idea was widely disseminated through both scientific literature and pop culture. But my research and that of my colleagues show that very few people are truly unskilled and unaware of it.
The Dunning and Kruger experiment actually showed an effect. Most people consider themselves better than average. But according to my team’s research, that’s all Dunning and Kruger showed. As a matter of fact, people have an innate ability to measure their abilities and knowledge. Claiming otherwise falsely suggests that much of the population is hopelessly ignorant.
This article originally appeared in The Conversation. Please read the original article.
