Why you shouldn’t use the Dunning-Kruger effect to insult people (and how it actually works)

Amanda van Lamsweerde, Ph.D., Research Psychologist II

Who doesn't love a snide new way of insulting people? Enter the Dunning-Kruger effect. It's experiencing quite a lot of attention in the media these days as a means of saying "you're dumb AND you're too dumb to recognize how dumb you are." And hey, I totally get it. Insults are fun, and when you can insult people with science, it's even better. But there are some systematic ways that this effect is being slightly misinterpreted across social media these days. I'm going to clear them up for you because if you go on acting like you're an expert about this phenomenon when you actually aren't, well, guess what you're exhibiting?

​
There are two parts to this that I'm going to talk about: 

1. What the Dunning-Kruger is
2. The sources of the effect

 
First, let's tackle what the Dunning-Kruger effect is.
​
Dunning and Kruger (1999) found across a variety of tasks, when asked to evaluate their own performance, low performers (those in the bottom quartile) tended to think they performed a bit above average (around the 60th percentile) - see Figure 1. Those in the second and third quartile also overestimated their abilities, but the better people performed, the less 'off' their judgments were. Those in the top percentile actually thought they performed a teensy bit worse than they actually did. The pattern of the Dunning-Kruger effect has been replicated many times over and across many types of tasks.

---

The part of this that grabs attention is that low performers rate themselves as shockingly better than they really are. Note, however, that while the lowest performers rated themselves as above average, they did not rate themselves as "the best" - they rated themselves as "a bit above average." The people who rated themselves as the best were, in fact, the best: the top performing quartile also had the highest perceived ability. This runs contrary to the popular understanding of the effect, in which the most incompetent people think they are the best. They don't think they're the best, they just don't think they're the worst.


To sum up Part 1, the Dunning-Kruger effect:

Now for Part 2: the cause.

Part of this effect is likely a systematic bias that people have about believing that they are just generally crushing it. This is called the above-average-effect,  better-than-average effect, or illusory superiority. People just rate themselves as above average at most things. This has been found a lot.

However, Dunning and Kruger argue that an additional factor is at play in this specific effect, and this explanation captures a lot of attention. Poor performers rate themselves as better than they are because of failed metacognition, which is to say, they lack insight into their abilities. Poor performers (or "the incompetent") lack the ability to know that they're not so hot. These poor performers suffer what's called a double burden: not only do their lack of skills lead them to produce the wrong answer, but it also prevents them from recognizing their error, thus inflating their impressions of themselves. According to Dunning and Kruger, the same knowledge is needed to do the task and recognize how you're doing. If you don't have that knowledge, you're bad at both. Poor performers think they did better than they did because they lack the ability to accurately assess themselves. As you become more competent, you gain both the skills needed to master the task and more accurately evaluate yourself, creating greater metacognitive calibration. The people at the top may have a tendency to think everyone knows as much as they do (the false consensus effect), thus making them think that other people did better. This slightly lowers their own self-estimation.

This idea of a lack of metacognition on the part of the incompetent has been pretty front-and-center lately. However, this interpretation has been subject to a bit of debate. As it turns out -  as is so often the case -  the original authors are not the only ones with thoughts about the cause of their reported pattern of results.

One alternate explanation is regression to the mean, and it's a big one. All people are going to make some errors about their performance. These errors will tend to pull toward the average. Therefore, you will see low performers rating themselves as higher than they are and high performers rating themselves as lower than they are. When this effect is statistically controlled for, the Dunning-Kruger effect is reduced, but it doesn't go away completely. So it's probably part of the effect, but not the whole effect.

One argument is that the 'above-average effect' + 'regression toward the mean' are sufficient to explain the Dunning-Kruger effect: no metacognitive explanation is necessary. Or at least, low-performers need not  have poorer insight than high performers. Think of it this way: everyone may think that they're better than average, but some people actually are. Are those people better calibrated -  do they have more insight - or do they have the same "I'm awesome" bias as everyone else, but they happen to be right? McIntosh et al. (2019) suggest that all three factors can contribute: a little bit of general above-average effect, a little bit of statistical quirks, and a little bit of low metacognition (with the last being a very small, and not even necessary, contributor).

While the debate about the cause of the Dunning-Kruger effect rages on, go ahead and use quirky psychological phenomena to insult people. I won't judge.