The illusion of explanatory depth
Believing you understand how something works in more detail than you do — and finding out otherwise the moment you try to explain it. The classic demonstration is mundane: people rate their understanding of a zip, a toilet or a bicycle highly, are asked to write out the mechanism step by step, and rate themselves lower afterwards. The explanation attempt is the measurement.
Entered this corpus through epictetus-learning-spacedaily, which reports Fisher, Goddu and Keil’s nine experiments at Yale (2015) extending it to search: searching the internet inflated people’s sense of how much they understood, and being able to look something up was mistaken for knowing it. All of that reaches this page at second hand from a T3 essay — the paper itself is not held, and the count of experiments and the finding are as reported there.
Why it is not Dunning-Kruger
The two get merged constantly, including by sources that are otherwise careful, so the distinction is worth stating on its own page.
Dunning-Kruger is about rank. Kruger and Dunning measured how people place themselves against everyone else, once, with nothing done to them; the bottom quartile sat at the 12th percentile and claimed the 62nd (kruger-dunning-1999).
The illusion of explanatory depth is about depth, and it is inducible. Fisher’s design changes someone’s felt understanding by having them search, while their knowledge stays where it was. Nobody has to be a poor performer for it to work; the effect is manufactured by the act of retrieval.
So: one is a miscalibration you arrive with, the other is a distortion you can be given. A remedy that suits one is wrong for the other, which is the practical reason to keep them apart.
What follows from it
The correction is unusually concrete for a bias, because the diagnostic and the fix are the same act: explain the thing plainly, without looking anything up, and treat the gap between expected and actual as the measurement. That is Fisher and Keil’s paradigm turned into something a person can run on themselves, and it is the advice epictetus-learning-spacedaily closes on.
It also puts a two-thousand-year-old observation and a 2015 experiment in the same frame, which that essay does deliberately: Epictetus said a person who thinks he already knows cannot begin to learn. The gap between the two is real and the essay glosses it — Epictetus is describing someone who will not start, while Fisher’s subjects are willing and simply wrong about what they hold. Motivation against calibration.
Held thinly
One second-hand report, no primary. The corpus does not hold Rozenblit and Keil’s original work, which is where the effect was named, nor the 2015 paper itself. Both are the obvious next sources.