dunning kruger effect examples, and i’m fairly sure

47 tabs blinked, the laptop fan groaned, and somewhere underneath it all my own keyboard was generating examples of a famous psychology effect, in real time, with no permission. that, itself, is an example of the effect. mondays are objectively better than fridays. i will die on that one. fairly sure.

so this is the post about dunning kruger effect examples, written from the standing desk where i sit, on a tuesday i did not earn. carla is in the training session two floors up. she said it ends at eleven. it never ends at eleven. i have, give or take, the rest of the morning before someone notices the cursor stopped moving in the spreadsheet. that is enough time to catalogue, badly, a phenomenon that catalogues me back.

writing this from desk. carla is in the training session on the third floor. spreadsheet open in another window for camouflage. monitor angled three degrees away from the door, which is, by my own admission, paranoia at a sub-clinical level.

i’m going to give you examples in four buckets. the famous ones, the ones i found by clicking through 47 tabs at 9:47am, the ones i can see from this standing desk, and the ones that almost include me but not quite, your honor. for the broader theory of the thing — what it is, where the chart came from, why it eats meetings — i wrote a longer treatment in the long version of dunning kruger, drafted at this same desk last quarter, which is the place to start if you are starting. this post assumes you are not starting.

dunning kruger effect examples, the famous ones

the textbook example, which i did not read in a textbook, is a bank robber from 1995 who covered his face in lemon juice because lemon juice is, for unrelated reasons, used as invisible ink. he reasoned that if lemon juice hides writing from the eye, it would hide a face from a security camera. this is real. they caught him within hours. he was, by all accounts, surprised. that’s the most cited story in the literature i’m fairly sure exists, possibly in a serious magazine. it’s the story everyone leads with when they explain the topic, including me, just now.

the second famous one is anything where a person without training enters a domain with training and starts giving notes. there is a film tradition for this — see the 2004 film “anchorman” with will ferrell, in which a man with no qualifications becomes the loudest person in a newsroom by refusing to hear anyone else. that’s an example with a budget. that’s also, i suspect, a documentary, lightly disguised.

the third famous example is anything online, which is to say, all of online. the comment section is the chart in motion. the most certain comment, statistically, is the one that should not have been left.

examples i found across 47 tabs

i opened a fresh tab at 9:47am and i looked. by 10:14am there were eleven new tabs, and the original count had risen, which is itself a violation of physics i don’t have time to investigate. here is what i pulled out of them.

tab one: a man on a podcast explaining nutrition to a registered dietitian. the dietitian was patient. the man was certain. the certainty did not move when corrected. it just rephrased itself in slower words, which is the chart’s signature move.

tab two: a thread arguing that a famously difficult novel was overrated. the person had read the first thirty pages. they had a clear, confident view of the remaining six hundred. that’s the entire mechanism in one screenshot. they had reached the peak from the lobby.

tab three: a video tutorial titled “how to fix your own [thing]” filmed by someone who had, by minute four, broken the thing further. the comments under the video were full of people thanking the host for clarifying their own next attempt. that’s the effect propagating, a relay race nobody asked to enter.

tab four: a discussion forum where two strangers were arguing about a specialty neither had practiced. they had each, separately, watched a documentary. they were each, separately, the most certain person at their kitchen counter. nobody won. the forum went quiet. the topic remained unresolved, because the topic was never the topic.

tab five through forty-seven: variants. seriously. the same shape, different costumes. the wip 2022 list is in the third tab from the left. one of its items, kid you not, says “research dunning kruger”. i am, in this very moment, one tab away from closing it. i will not close it. closing it would be a peak move.

examples from the standing desk where i sit

i don’t have to look online to see this. i can see it from here. this working language for the broader pattern, drafted from the same chair is where i went deepest on what the cluster around the effect looks like in normal-person terms — the everyday sibling of the famous chart, the version that lives in offices.

the volunteer expert. there is a person two desks over — i won’t say which two — who has never managed a project but explains, weekly, what is wrong with the projects of others. the explanations are confident. the explanations are calm. the explanations are also, when you check, wrong. nobody checks. that is the effect’s lubricant. it relies on nobody checking.

the meeting opener. in any meeting i’ve sat through this year, the first person to speak has been, on a strict majority basis, the person least equipped to speak. they had a thought in the elevator. they brought it. the people who knew the topic waited. by the time the people who knew the topic spoke, the meeting was over and the wrong thought had a follow-up.

the email volunteer. someone replies-all to a thread they were not on. their reply contains the phrase “actually, i think”. what follows the phrase is, on careful reading, the third or fourth thing the original thread had already corrected. the reply gets thanked. the reply gets forwarded. the reply becomes a memo. that’s the food chain. that’s how a wrong thing gets dignified.

the standing desk. i bought this standing desk during a phase. i sit at it. i have sat at it for two years. the desk is, in itself, a small dunning kruger effect example, in physical form: a person who briefly believed he understood his own posture, ergonomics, and discipline, and who has since revealed, by inaction, that he understood none of the three. i am not going to return it. i will, however, refuse to throw it away. it stands. i sit. we have an arrangement.

examples that almost include me, but don’t

i’d like, for the record — and you can write this down — to walk you through several examples that almost involved me but ultimately, technically, did not. this is a defense brief. read it as such.

last quarter, i told a colleague that the entire concept of project management was, and i quote myself badly, “vibes plus a calendar”. the colleague, who has managed projects since before i had taste, smiled. that smile was the chart. the smile was telling me i was on the peak. i was. i remember. i did not, however, defend the position when pressed. that’s the move. that’s the off-ramp. you can be on the peak and still get off it, if someone smiles correctly and you are willing to read the smile.

also last quarter, i told mike, at the corner bar, that taxes were “honestly not that complicated”. mike, by his own admission, has not filed since 2019. mike laughed for what felt like nine minutes but was probably ninety seconds. mike has a system. mike’s system is silence. i deferred. i ordered a second beer. that’s not climbing. that’s evacuating.

also last month, a friend of a friend asked me, casually, what i thought about audiobooks. i held my position briefly: books on tape are cheating, an opinion i hold as a hot take, not a verdict. the friend of a friend is a librarian. she explained, kindly, what she thought. i listened for the recommended six seconds and then for several more. i adjusted, slightly. that, also, is not the effect. that’s the part of the chart they don’t put on posters — the slow climb back up, no music, no medal, just a person on a bar stool doing slightly less damage than yesterday.

findings, the examples are everywhere, sorry

here is the unhappy result of the audit. once you start looking, you stop being able to stop. the dunning kruger effect is not a rare event. it is the wallpaper. you, me, mike, the man on the podcast, the colleague, the in-law, the standing desk, the wip 2022 list, the comment section, the meeting opener, the audiobook friend, the librarian’s cousin — all of us, on a long enough timeline, are walking around with confident maps to cities we have never visited. the only reasonable response, on a tuesday with the rest of the morning still on the clock, is to keep a list of the topics on which i, specifically, am most likely to be wrong. that list is also a tab. it is, of course, open.

three working examples to take with you, if you are a person who likes to leave a post with three things:

(a) when you feel certain about a topic in someone else’s domain, that is data, but not the data you think it is. (b) when you feel uncertain about a topic in your own domain, that is also data, and probably more useful. (c) when the most confident person in any room is also the loudest, the right response is silence, and a follow-up email two days later asking a question you already know the answer to, just to see what they say. it works. i have used it. mike taught me. mike has not filed his taxes since 2019, but he reads a room.

11:04am. carla just walked past my desk. she did not stop. that means the training session ended early or she’s looking for coffee. either way, the spreadsheet is on top now. this tab has been moved one to the right.

↑