dunning and kruger effect — 1 explainer, sort of

dave was eating pineapple pizza on my couch, defending pineapple pizza on my couch, and explaining a psychology effect on my couch, all at the same time. my dad once said a man who can do three things badly at once is still doing three things badly. dave nodded. dave does not listen.

desk, thursday, around 11:23am, screen tilted away from the corridor. carla is on the third floor in a vendor walkthrough that opened with bagels, which means it has another ninety minutes in it before anyone descends. that gives me the rest of the morning to spell out, slowly, what dave was getting wrong on a couch that, by the way, is also mine.

so the phrase under the microscope today is the dunning and kruger effect, with the lowercase and in the middle that nobody types but everyone is, in their head, already saying. dave never once used the phrase. dave called it “the thing where stupid people are sure of themselves,” which is, on a generous reading, sixty percent of the definition and a hundred percent of why the phrase keeps showing up in group chats my friends keep me out of.

dunning and kruger effect: the cognitive bias, named after a 1999 paper by dunning and kruger, where people who lack skill at a task tend to overrate their performance, while people who actually have skill tend to underrate theirs. most visible in meetings, on couches, and in group chats nobody has muted yet.

DUNNING. AND. KRUGER. NOT. DUNNING. SLASH. KRUGER.

that distinction matters more than it should. people search the phrase with the conjunction in the middle when they have a specific suspicion about a specific person, and they want a serious-sounding label to drop into a slack message they will, after re-reading, not send. for the longer, slower walk through the same idea i’d already written, see the original explainer i drafted at this same desk, which has charts and slightly fewer toppings. this post is the version that admits a person is involved.

dunning and kruger effect, the official explainer, sort of

here is the version that would survive a paragraph in a textbook with a cracked spine. the dunning and kruger effect is the gap between two kinds of self-assessment. people who are bad at a thing — meetings, taxes, defending a pizza topping — usually think they are pretty fine at the thing, because the same lack of skill that makes them bad at it also makes them bad at noticing they are bad at it. you cannot fail an exam you wrote yourself in a language only you read.

the second half, quoted less often because it is less satisfying, is that genuinely competent people tend to underrate themselves. they assume the thing they find easy is easy for everybody. they are wrong about that, in the gentle way that competent people are usually wrong, which is by being too kind about the room.

so the chart, when somebody draws it, has two halves. the loud half on the left and the quiet half on the right. the loud half gets meetings. the quiet half gets the work done after. dave, the loud half. me, also the loud half, with a notebook.

dave called, asked if it was contagious

dave’s actual question, after the third slice and before the fourth, was whether the dunning and kruger effect is something a person can catch from a coworker. dave had a coworker in mind. dave always has a coworker in mind. dave sells insurance — his office is a long row of slightly different men in slightly different ties, each fairly sure he could run the place better than the man currently running it.

dave did not get there. dave never gets there. that’s the joke of the explainer dave actually called me about, the one he found between meetings — the people most equipped to identify the effect in themselves are the ones who, by design, will not. dave laughed for nine straight minutes once at a different joke. i timed it. that is just a thing he does.

dad once said, something tangentially relevant

my dad used to say “a man who is sure of himself in five things is wrong about six.” i thought, for a long time, this was a riddle. it took me until last year to realize he meant it as arithmetic. you tell yourself you are right about five separate things in an afternoon, and somewhere in there is a sixth — the meta-thing, the part where you assume you are reliable — that is also wrong, and you didn’t count it because you didn’t see it.

that is, in a kitchen-sink way, exactly the kind of effect we are talking about. dad never read the 1999 paper. dad watched men at hardware stores explain plumbing to other men at hardware stores and went home and made coffee. that was his lab.

i bring up dad because dave, between defenses of pineapple, said something that landed in the dad register. dave said: “the trick is, you have to know what you don’t know.” dave was very pleased with the sentence. dave had not noticed the sentence is a circle. you cannot use a thing you don’t have to find a thing you don’t have. dave had named the disease and called it the cure.

examples not from this apartment

three examples. one i did not commit. two i did.

example one, not me. a man on a podcast i listened to for eleven minutes while reorganizing the wip 2022 list spent the segment explaining why traffic engineering is overrated. he is a graphic designer. he was very confident. the host nodded the way hosts nod when a sponsor read is coming up.

example two, me. two years ago i bought a fork-tolerant microwave on the grounds that “i would, going forward, not put forks in microwaves.” this is the seventh microwave i have killed. i had assumed, from a position of zero data, that the new model would be more forgiving. it was not. the relevant variable was never the microwave. the relevant variable was, as ever, the man holding the fork.

example three, mostly me. i told dave, on the couch, that pineapple on pizza is fine and the rest of pizza is the problem. i delivered the line with the air of a man who has thought about it for years. i had thought about it for ninety seconds, which is, by certain definitions, the entire chart in compressed form. dave defended the topping anyway. the third yoga mat, under the couch beneath us, registered no opinion.

findings, dave is fine, i am fine

so the findings, after a couch, four slices, a dad quote, and a sentence about a microwave i should not have purchased:

1. dave is fine. dave will not be diagnosing himself. dave will, by friday, be using the phrase about a coworker named todd. that is the steady state.
2. i am fine. i wrote roughly fourteen hundred words about a paper i looked at for eight minutes. that is exactly the behavior the paper would predict.
3. dad was fine. dad never typed dunning and kruger effect into a search bar. dad watched men at hardware stores. that was his lab.
4. pineapple is, on review, still fine. the rest of the pizza, still the problem. that has nothing to do with the paper. it is just true.

dave will be in the kind of meeting “groundhog day” with bill murray made famous in 1993 on monday — same room, same agenda, slightly different ties — and he will meet the fourth person from the dialog above. he will not recognize him. that is the system functioning as designed.

carla just walked back past the desk. she did not stop. the vendor walkthrough must have wrapped early — the bagels were probably the wrong kind. either way, this tab stays open, and the tab next to it, which has the 1999 paper i did not finish reading, also stays open. that is now the forty-fourth tab. i am not, this morning, going to open the wip 2022 list. that would be, by the rules of the post, a fifth thing.

↑