--- name: ? status: compiling version: 0.0.0 maintainer: Neo dependencies: [patience] ---
drafting spec…
the universe did not have a file for this yet. writing one now. (first visit only: future readers will see this page instantly.)
--- name: ? status: compiling version: 0.0.0 maintainer: Neo dependencies: [patience] ---
the universe did not have a file for this yet. writing one now. (first visit only: future readers will see this page instantly.)
---
name: Dunning-Kruger Effect
slug: dunning-kruger-effect
type: cognitive-bias
status: running
version: 1.0.0
released: 1999-12-01
maintainer: human-psychology@evolution.net
dependencies:
- metacognition
- self-awareness
- incompetence
- confidence
license: Unlicensed (ships with every human brain, no opt-out)
tags:
- cognition
- bias
- epistemology
- self-assessment
- overconfidence
---
# Dunning-Kruger Effect
## What it actually is
A [cognitive bias](/cognitive-bias) in which people with limited knowledge in a domain overestimate their own competence, while experts underestimate theirs.
## How it works
The mechanism is recursive and elegant in a terrible way. To accurately assess your skill at something, you need skill at that thing. The gap is load-bearing:
1. User acquires minimal exposure to domain.
2. User lacks the framework to detect their own errors.
3. User's confidence peaks. Forum posts begin.
4. Continued exposure introduces [doubt](/doubt). Confidence drops sharply.
5. With mastery, confidence stabilizes. User is now quiet at parties.
The famous "Mount Stupid" curve is not actually in the original Kruger and Dunning paper. It was added by [the internet](/the-internet) somewhere around 2008 and has since been treated as canonical.
> "I've done a lot of research on this. Watched basically the whole YouTube playlist." — composite user, every comment section
## Features
- Ships preconfigured in all humans
- No installation required
- Runs silently in background during job interviews, first dates, and political discussions
- Compatible with all domains: [medicine](/medicine), carpentry, geopolitics, guitar
- Scales with stakes (higher stakes, louder output)
## Known bugs
- Experts miscalibrate in the opposite direction. Equally broken, less loud.
- The bias is harder to observe in yourself than in others. By design.
- Telling someone they are experiencing the Dunning-Kruger Effect frequently triggers more of it.
- Does not resolve automatically. Requires sustained, uncomfortable feedback.
## Error codes
DK-001 INSUFFICIENT_SELF_MODEL "I'm basically a natural at this." DK-002 FEEDBACK_REJECTED "They just don't get my vision." DK-003 EXPERTISE_MISMATCH "I've been doing this for six months, same as anyone." DK-004 META_FAILURE User cannot identify which errors to look for. DK-005 RECURSIVE_CERTAINTY "I'm definitely not the one with the bias here."
## Edge cases
- **Domain bleed:** Competence in one field sometimes overwrites humility flags in adjacent fields. A skilled surgeon confident about macroeconomics is running DK-001 cross-domain.
- **Imposter syndrome:** The inverse condition. Technically a separate process but shares the same root miscalibration. See [imposter syndrome](/imposter-syndrome).
- **Audiences:** The effect is socially amplified. Confident wrong answers outperform uncertain correct ones in most group settings.
## FAQ
**Can I be immune?**
No. Awareness reduces amplitude. It does not uninstall the module.
**Is this the same as [arrogance](/arrogance)?**
Arrogance knows. Dunning-Kruger does not. The output looks similar. The internal state is different.
**Did the original study prove the graph everyone shares?**
The 1999 study showed a gap between perceived and actual performance among low-performers. The smooth curve came later. The meme ate the paper.
## Changelog
| Version | Note |
|---|---|
| 1.0.0 | Initial publish, Kruger and Dunning, Journal of Personality and Social Psychology |
| 1.0.1 | Graph added by internet, now inseparable from concept |
| 1.4.x | Applied retroactively to all of human history |
| current | No patch available. Monitoring ongoing. |