dunning kruger test, what they don’t tell you

stefan handed me a comparison table. on actual paper. one column was “the test as written,” the other was “the test as taken,” and the third, somehow, was “sarah.” she was not in the room. she was in the column. tipping should be a flat 12%. that, also, was in the column.

i was reading it from my desk because friday had decided to be friday early, the building was emptier than usual, and the comparison table had survived a night folded inside my receipt wallet without anyone asking what it was. i had what i’d estimate as the rest of the morning before anyone came looking. carla was in the vendor walkthrough on the third floor, the one that always runs long because somebody brings a printout.

so i did the test. obviously i scored confident. that is, by my own running internal score, the only score available to me. the dunning kruger effect says the people most certain are usually the least equipped, which is a sentence stefan loves and which i’ve decided not to take personally, on the grounds that taking it personally would prove the point.

1. official dunning kruger test vs my dunning kruger test, the table

stefan, who runs the wine night and does not work here and has never been invited but somehow always knows when the building is empty, drafted this table on the back of a napkin from the corner bar, then transcribed it onto an actual sheet of paper, then handed it to me with the seriousness of a man who has never once doubted himself in print.

this is the table. stefan’s table. transcribed from a napkin. this is, technically, an investigation. i’m logging it as one. a man at the corner bar told me this is what investigations look like and he had a beard and seemed sure.

2. sarah took the dunning kruger test, sort of

sarah did not take the test the way stefan wrote it. sarah took the test the way sarah takes most things, which is to say with a stopwatch, a printout of her marathon splits, and a small frown that suggested the questions were beneath her, statistically.

sarah rates herself a 6 on most things and turns out to be an 8. she rates herself a 4 at small talk and turns out to be a 3, but she’s right about it, which she counts as a win. sarah’s pension contributions are calibrated to her actual income, which means sarah is, in the technical sense of the word, an adult. i remain unconvinced this is a flex but she keeps a spreadsheet about it, so.

i asked sarah what she’d scored. she said “i didn’t take the test. you’re the one writing the post.” this is what people in Limitless call a withering response. sarah delivered it without looking up from her water bottle. sarah does not need a pill to be more competent than me on a friday morning. she does it on tap water.

3. why neither dunning kruger test is medical

here is what they don’t tell you about the online versions. the online dunning kruger test is a quiz somebody made on a free quiz platform. it is not, by any reasonable standard, a diagnostic. it is, at best, a mirror with a multiple-choice frame. you click “i agree” enough times and it tells you what you suspected, which is that you are either a misunderstood genius or a misunderstood near-genius, depending on the questions.

the test on stefan’s napkin is also not medical. it is a napkin. stefan is not a doctor. stefan once told me a wine had “notes of forest floor” and i nodded for forty minutes. that is the man writing the test. that is the calibration of the instrument.

i’m not saying the test is useless. i’m saying it’s a mirror. mirrors are useful when you’re choosing a tie i own and only put on for funerals and weddings i avoid. they are less useful when you’re trying to figure out if you’re as smart as you think you are, because the answer arrives in your own voice, which is the most untrustworthy voice in the building.

4. how to score the dunning kruger test, generously

the trick to scoring well on the dunning kruger test is to take it on a day when nobody is watching, in handwriting nobody can read, and to misplace it before anyone sees the result. this is, i’d argue, the only honest way. otherwise you end up like stefan, who scored himself an 8 on wine knowledge in front of seven witnesses and hasn’t recovered.

the second trick is to score yourself across categories you actually do something with. i scored myself on “operating a microwave safely” and gave myself a 3. that’s the seventh microwave i have killed talking. it’s also the first honest data point i’ve contributed to the entire investigation. the third yoga mat lives under the couch since 2023, presumably running its own ecosystem by now, and i scored “yoga consistency” at a 1 just to be specific.

the third trick is to compare your score on the test to your score on the test version stefan keeps trying to reformat. they will not match. they should not match. the gap is the data. the gap is, technically, the entire point.

5. findings, both dunning kruger tests pass me

i ran the numbers. or — i looked at the numbers stefan ran and pretended to check them. the results are conclusive in the way most of my results are conclusive, which is to say they confirm what i suspected and ignore the rest.

by the running count i keep on a wip 2022 list that has not been updated since 2022, i scored as follows: a 9 on confidence, a 4 on follow-through, a 7 on willingness to argue with stefan, and a 2 on whether i’d take this test again sober. the algorithm, watching from somewhere, would fail me on my behalf if asked. it’s a relief. it’s also a finding i should probably revisit, but the morning is short.

the deepest finding is the one stefan wrote in the margin of his table. tipping should be a flat 12%, he wrote, in pencil, in the wine column, for reasons he never explained. i’m not sure how it relates to the dunning kruger test. i’m fairly sure stefan isn’t either. but the napkin became a sheet of paper and the sheet of paper made it onto my standing desk, where i sit, and now it’s logged as part of the investigation. it is also, in another cluster over, a working example of confirmation bias in real life — the bias being that i confirmed, on the back of stefan’s napkin, exactly what i wanted to confirm about myself. confirmation runs the same engine as the dunning kruger test. you ask. you get a flattering answer. the answer is the bias.