The Dunning-Kruger Effect: You (Probably) Don’t Know What You Don’t Know
If you spend any time on Twitter, you may have seen a tweet last fall reacting to a YouGov survey. The survey, conducted by the British market research firm YouGov, asked people about their tennis abilities and how they thought they’d perform in a match against Serena Williams. 12% of men surveyed said that if they were playing their very best tennis, they think they’d be able to win a point off of Serena Williams. 12%!!! That’s 1 in 8!! These are average, everyday British men, talking about scoring a point against the tennis player with the most Grand Slam titles in their career. In reality, these people would be lucky to even touch one of the balls hit at them by Serena Williams. So why do they have this overinflated sense of their tennis abilities? This is only one example of the Dunning-Kruger effect, a metacognitive idea that influences many aspects of our everyday lives, from wine-tasting to practicing medicine.
In 1999, two psychology researchers at Cornell, Justin Kruger and David Dunning, published a paper outlining their recent study. Their research was investigating peoples’ abilities to accurately assess their performance in various tasks, driven by the observed phenomenon that people tend to hold “overly favorable views” of their abilities. This is an example of the “above-average effect,” in which people tend to believe that they are above average in their skills. Statistically, this is impossible – only half of a given population can actually be above average. Participants were evaluated in tasks regarding humor, logical reasoning, and grammar; these evaluations were performed by the participants themselves (self-assessment) and by objective experts.
So, how did people perform in their self-assessment? Were they able to accurately rate their own abilities? Yes, sometimes. It appeared that people who performed in the bottom 25% (according to objective judges’ evaluations) were the ones responsible for most of the above-average self-assessments. In other words, those who performed the worst were most likely to think they did better than average.
WHAT?
This seems like a cruel twist of fate (and metacognition) – if you’re bad at something, you’re more likely to think you’re good at it. I bet anyone reading this post can think of a time when they encountered someone like this (or maybe, you were someone like this). How often have you been in a conversation where the person speaking up the most and appearing the most confident actually has no idea what they’re talking about?
The cause of this miscalculation is not always clear, but Dunning and Kruger argued that people with low competence don’t know enough about the task or skill they’re performing to know that they’re bad at it. In this sense, they don’t even know what they don’t know. Once lower-performing participants were taught how to complete a task correctly and were able to do so, they also showed an increased ability to accurately assess their performance. One way to make people recognize their incompetence was to make them competent (of course, this is a bit of a paradox as they were no longer incompetent).
Since 1999, the Dunning-Kruger effect has been cited in studies on various topics, ranging from wine consumption to financial literacy. Claudio Aqueveque’s recent article, Ignorant Experts and Erudite Novices: Exploring the Dunning-Kruger Effect in Wine Consumers, illustrated just how prevalent this cognitive bias is among people buying wine, which has important implications for the wine industry. This study found significant overestimation of one’s wine knowledge in the “lowest objective knowledge” group, and, conversely, significant underestimation of wine knowledge in the group with the “highest objective knowledge.” Those with the lowest knowledge were likely to overestimate their knowledge, while those with the most knowledge were likely to underestimate it.
So what does this mean for us? Should we assume that no one has any idea what they’re talking about when they mention things like tannins and bouquets? Not necessarily. There are definitely people who can accurately assess their knowledge, though this study suggests that these folks are in the minority. This is more relevant to businesses conducting market research, who need to be aware of consumers’ bias and emphasize objective measures of wine knowledge instead.
Ok, fine. I guess people don’t always understand how much they know about wine. Why should you care?
The Dunning-Kruger effect is not limited to people’s wine knowledge. Students learning chemistry also demonstrated the Dunning-Kruger effect, as outlined by Samuel Pazicni and Christopher F. Bauer in their 2014 study, Characterizing Illusions of Competence in Introductory Chemistry Students. Their findings confirm and extend observations of the Dunning-Kruger effect to the college classroom environment, as students with low performance substantially overestimated their knowledge, while those with high performance underestimated their abilities, though to a lesser extent. This may be familiar to you if, like me, you’ve procrastinated studying something because you feel like you already know it pretty well. In reality, you may not know it well at all, so you’re unable to recognize your lack of understanding. Moral of the story: study more than you think you need to!
Pazicni and Bauer found, in addition to the general Dunning-Kruger effect, an interesting gender dynamic in self-assessments. Female students had less inflated self-perceptions than male students did. This gender gap is not specific to introductory chemistry. Vajapey et al reported significant gender differences across students, residents, and faculty in medicine at colleges and hospitals around the world. On average, 75% of men had higher self-confidence than their female peers, even though there was no difference in clinical abilities between men and women. This confidence gap could have further implications for what medical specialties physicians choose. Female doctors, feeling less confident in their abilities, may be less likely to go into “harder” specialties like surgery.
Gender differences aside, Davis et al (2006) conducted a systematic review of physicians’ self-assessments, and found that physicians that were the least skilled showed the highest levels of self-confidence, while being the worst at self-assessment. Just like the wine consumers, physicians are vulnerable to falling into the Dunning-Kruger trap. This could have more serious consequences for doctors than wine connoisseurs, though. If you’re undergoing a complicated surgery or facing a difficult disease, you want the most skilled doctors on your care team. However, the people who are likely to speak up and volunteer their opinions may be the ones with the least skills, but the most confidence. Yikes! I’ve known some people like this, who think that watching Grey’s Anatomy makes them qualified to give their medical opinion. Yet I trust what they say, because they are confident that they are right. I’m not the only one who does this; people tend to place more value on the opinions of confident people. Pairing this cognitive bias with the Dunning-Kruger effect, we end up in a situation where we are trusting the opinions of the people least qualified to give them!
That all sounds like a recipe for disaster. Indeed, lots of dangerous misinformation, especially about vaccines, has been spread by people with very little knowledge about the things they share. Yet there has been recent evidence suggesting that the Dunning-Kruger effect is not as significant as originally thought.
Oh, so does this mean that people are actually better at self-assessment than Dunning and Kruger proposed?
Unfortunately, no. In fact, it’s sort of the opposite. While Dunning and Kruger theorized that those with low competence at tasks are also the worst at evaluating their performance on these tasks, recent studies have shown that people tend to be equally bad at self-assessing their abilities. The best performers are just as likely to misjudge their competence as the worst performers. In their 2006 paper, Burson, Larrick, and Klayman illustrated that the “skilled and unskilled” are similarly unable to judge their performance relative to others with accuracy. The direction of the miscalibration is actually related to the difficulty of the task being performed, not just the person’s abilities. When asked to perform tasks that appear easy, poor performers are more likely to overestimate their abilities, while good performers can assess their skills more accurately. The converse is also true; with a difficult task, good performers are more likely to underestimate their performance, while poor performers are more perceptive of their skills.
Ok, this makes sense. When facing a hard task, we’re able to recognize our shortcomings. But when easily completing a simpler task, our perceptions of our skills are overinflated. If something comes to us easily, we are likely to think that it’s because we are better than others, not because the task is inherently easier.
Not gonna lie, this is a bit disappointing. We’re not as good at things as we think we are. Bummer. However, if you look at it another way, we’re all equally misled about our performance. If I’m really bad at something, I’m just as likely to misjudge my abilities as someone who’s really good at it. In general, people don’t really have a good idea of their own abilities, regardless of their skill level.
So what does this mean for everyday life? Should we assume that no one has an accurate idea of their skills? Yes, sometimes. A safer approach would be to question what people tell you, and seek input from multiple sources. Don’t just believe someone because they seem confident in their information. People who are more confident may just be more misinformed!
Even if the Dunning-Kruger effect is not as dramatic as originally thought, we should approach new information with a healthy degree of skepticism. People tend to believe that the information they receive is true, even though the people reporting it are as equally susceptible to misinformation as the rest of us. Being aware of this bias is a good first step towards counteracting it.
What’s the takeaway here? Good question! I hope that reading this has helped you learn more about the Dunning-Kruger effect and its impact on everyday life, as well as the limitations of Dunning and Kruger’s work. Don’t worry too much about your self-assessment skills if you’re bad at something – chances are everyone is equally bad at introspection and self-evaluation! However, this means that everyone is susceptible to overestimating their skills or knowledge, and may confidently report things that really aren’t true.
In the Serena Williams survey, men were more likely to think they could score a point. This demonstrates an interesting intersection of the Dunning-Kruger effect and gender biases. As outlined above, in many situations women are more likely to underestimate their abilities while men overestimate theirs, leading to imposter syndrome and decreased confidence. Misjudged self-assessments can lead to misplaced overconfidence, or an underestimation of your abilities. The Dunning-Kruger effect may not be limited to only the worst performers, as everyone is likely to misjudge their skills. However, I can confidently say that Serena Williams is unlikely to give up any points against amateur tennis players, so you may need to reassess if you were thinking you could take her on.
References
Aqueveque, C. (2018). Ignorant Experts and Erudite Novices: Exploring the Dunning-Kruger Effect in Wine Consumers. Food Quality and Preference, 65: 181-184. https://doi.org/10.1016/j.foodqual.2017.12.007
Burson, K.A., Larrick, R.P., & Klayman, J. (2006). Skilled or Unskilled, but Still Unaware of It: How Perceptions of Difficulty Drive Miscalibration in Relative Conceptions. Journal of Personality and Social Psychology, 90(1): 60-77. http://dx.doi.org/10.1037/0022-3514.90.1.60
Davis, D.A., Mazmanian, P.E., Fordis, M., Van Harrison, R., Thorpe, K.E., & Perrier, L. (2006). Accuracy of Physician Self-Assessment Compared With Observed Measures of Competence: A Systematic Review. JAMA, 296(9): 1094-1102.
Kruger, J., & Dunning, D. (1999). Unskilled and Unaware of It: How Difficulties in Recognizing One’s Incompetence Lead to Inflated Self-Assessments. Journal of Personality and Social Psychology, 77(6): 1121-1134.
Pazicni, S., & Bauer, C.F. (2014). Characterizing Illusions of Competence in Introductory Chemistry Students. Chemistry Education Research and Practice, 15(24): 24-34.
Vajapey, S.P., Weber, K.L., & Balch Samora, J. (2020). Confidence Gap Between Men and Women in Medicine: A Systematic Review. Current Orthopaedic Practice, 31(5), 494-502.
Recent Comments