To Our Critics

People keep telling us the Dunning-Kruger effect isn't real. They might be right. We'd like to talk about that.

We've Read the Papers

Every few months, someone sends us a link. The subject line is usually something like "thought you should see this" or "lol your store is based on a lie." The link leads to a paper, or a blog post about a paper, or a Twitter thread about a blog post about a paper, all arriving at roughly the same conclusion: the Dunning-Kruger effect, as popularly understood, is not real.

We've read them. Not all of them — we're not academics, and some of them are behind paywalls that cost more than our mug. But we've read enough. And rather than quietly hoping no one notices, we thought we'd engage with the criticisms directly.

This is not a rebuttal. We are not qualified to write a rebuttal, and the fact that we considered writing one anyway is itself part of the problem.

The Dunning-Kruger Graph Is Fabricated

The most recognizable image associated with the Dunning-Kruger effect — the smooth curve with "Mount Stupid" at the peak, the "Valley of Despair" at the trough, and the "Plateau of Sustainability" at the far right — does not appear in Kruger and Dunning's 1999 paper. It isn't in their follow-up work either. The original paper has bar charts. Four of them. They show that people in the bottom quartile overestimate their performance. That's what the data looks like: quartile bars, not a sweeping mountain range.

The curve was popularized later, likely adapted from a graph in a 2011 blog post, which was itself adapted from a concept in a satirical context. It passed through enough hands and enough slide decks that it eventually became "the Dunning-Kruger graph," as if Dunning and Kruger had drawn it themselves. They did not.

We named our entire store after the peak of this curve. We put it on a mug. The mountain in our logo represents a geographic feature on a graph that was never in the paper it claims to represent.

It's Just Regression to the Mean

In 2020, Gignac and Zajenkowski published a large-scale reanalysis of the Dunning-Kruger pattern. Their conclusion was blunt: most of the effect could be explained by two well-known statistical phenomena — regression to the mean and the better-than-average effect. Regression to the mean is not a cognitive bias. It's what happens whenever you measure something imprecisely and then plot it against itself. The extreme scores drift toward the center, creating an illusion of asymmetry.

The better-than-average effect is the tendency for most people to rate themselves as above average, regardless of their actual ability. This is real and well-documented, but it's not specific to the incompetent. Almost everyone does it. About almost everything.

What the critics are saying is that the original Dunning-Kruger pattern — where the worst performers are uniquely miscalibrated — may just be everyone being slightly overconfident, with the extremes exaggerated by statistical noise. The incompetent aren't special. They're just like everyone else, except more so.

This is supposed to be a problem for us. We think it might be the opposite. If overconfidence isn't limited to the least skilled but is instead a basic feature of being human, then a store built on the premise that people overestimate their own competence isn't niche satire. It's general merchandise.

It's an Autocorrelation Artifact

In 2022, the economist Blair Fix published a detailed analysis arguing that the Dunning-Kruger effect is a mathematical inevitability. His point was straightforward: when you plot the difference between estimated and actual performance against actual performance, you're plotting a variable against a component of itself. The resulting pattern — low performers overestimate, high performers underestimate — is guaranteed by the structure of the math, not by any feature of human psychology.

Fix showed that you could generate the classic Dunning-Kruger pattern using purely random data. No humans needed. No cognition involved. Just noise that, once plotted in the right configuration, produces what looks like a deep insight about the human condition.

This is a methodological critique, and a serious one. It suggests that the original 1999 finding — the specific claim that the unskilled are uniquely unaware — may be an artifact of how the data was graphed rather than what the data contains. We don't have the statistical training to evaluate this claim on its merits, which is exactly the kind of sentence Dunning and Kruger would appreciate.

What we can say is this: we sell merchandise themed around a psychological effect that may owe its most famous pattern to an autocorrelation artifact. We didn't check the math before ordering the mugs. We saw a compelling graph, assumed it meant what everyone said it meant, and built a store. If that isn't a case study in confident ignorance, the field of psychology can stop looking for one.

Random Numbers Reproduce It

In 2017, Nuhfer and colleagues published a study that approached the question from a different angle. They generated data using random numbers — no human participants, no tests, no self-assessments — and then applied the same analytical method used by Dunning and Kruger. The result was a pattern that looked like the Dunning-Kruger effect.

Let that settle for a moment. Random noise, with no psychological content whatsoever, produces the same graph that has been used to explain everything from anti-vaccine sentiment to overconfident middle management. The pattern that launched a thousand think pieces can be reproduced without a single human being involved.

This is, if you step back and look at it, the most Dunning-Kruger-adjacent finding about Dunning-Kruger. A pattern that looks like profound insight into human cognition turns out to emerge from statistical noise — and for years, the people who cited it most confidently understood it least. Including us.

The Pop-Science Dunning-Kruger Effect Is Wrong

Multiple outlets — Scientific American, McGill University's Office for Science and Society, the British Psychological Society — have published careful takedowns of the popular version of the effect. The consensus is that the pop-science Dunning-Kruger effect has drifted far from what the original research actually demonstrated.

The 1999 paper showed that low performers overestimate and high performers slightly underestimate, within specific tested domains. The pop-science version turned this into a universal law of human nature: stupid people are too stupid to know they're stupid, and the graph proves it. The original finding was narrow and domain-specific. The popular version is grand, sweeping, and frequently used to dismiss entire groups of people.

In other words, the popular understanding of the Dunning-Kruger effect is itself a textbook example of the Dunning-Kruger effect. People encountered a nuanced finding, stripped it of its caveats, inflated it into a universal principle, and then deployed it with complete confidence. We built an entire brand on this version. We have a logo and everything.

Even Dunning Says You're Doing It Wrong

David Dunning has given numerous interviews over the years, and a recurring theme is his discomfort with how the effect that bears his name gets used. The most common misapplication, he's argued, is using it as a weapon — pointing at other people and declaring them victims of Dunning-Kruger, as if you yourself are somehow exempt.

"I wish people wouldn't use it as an invective, because it is really about being reflective about yourself and knowing that there might be things you don't know. It's not about judging other people."

That's Dunning himself, in an interview. His consistent message is that the effect visits all of us — "sooner or later, in our pockets of incompetence," as he's put it. It's not a diagnosis you give to someone else across the room. It's a condition you have to consider might apply to yourself, right now, about the thing you're most confident about. The moment you use it to feel superior to someone else, you've become the example.

Our mug says "I Climbed Mt. Stupid." The word doing the most work in that sentence is "I." It's a first-person admission, not a third-person accusation. Whether our customers always use it that way is beyond our control, but the grammar is deliberate.

So Is the Dunning-Kruger Effect Real

Let's take stock. The graph on our mug was never in the paper. The effect it depicts may be regression to the mean. The analytical method that produced the original finding may be an autocorrelation artifact. Random numbers can reproduce the pattern without human beings. The popular version that inspired our brand has been thoroughly debunked by the very outlets that popularized it. And the psychologist whose name is on the effect has repeatedly said that the way most people use it — including, presumably, the way we have built a store around it — is wrong.

We read a few articles about a psychological effect. We did not read the original paper. We did not check the methodology. We did not consult the statistical criticisms. We looked at a graph that turned out to be fabricated, found it compelling, and built an entire commercial enterprise on our confident interpretation of research we did not fully understand.

That is, by any reasonable standard, the most Dunning-Kruger thing we could have done.

But here's what the critics haven't debunked, because it isn't a statistical claim: the experience is real. You have sat in a meeting where the least prepared person talked the most. You have watched someone with a weekend's worth of knowledge explain a topic to someone with a decade of it. You have felt that specific cringe of looking back at something you said with total confidence three years ago and realizing you had no idea what you were talking about. The measurement may be flawed. The graph may be fabricated. But the thing the graph was trying to point at — the gap between what we think we know and what we actually know — is as real as it ever was.

We can't fix the methodology. We are not going to rename the store. The mug is already in production. What we can do is be honest about the foundation we're standing on, which turns out to be shakier than we thought when we started. That's an uncomfortable position, but it's also the most on-brand position imaginable.

We climbed Mt. Stupid. It turns out Mt. Stupid might not exist. We're still wearing the t-shirt.