Dunning–Kruger Effect vs Overconfidence Effect
Both describe confidence outrunning ability — but Dunning–Kruger is specifically that the LEAST skilled overestimate themselves the most, because the skill to perform and the skill to judge your own performance are the same skill. The overconfidence effect is broader: nearly everyone, including experts, is more confident than accurate, especially on hard questions.
| Dimension | Dunning–Kruger Effect | Overconfidence Effect |
|---|---|---|
| Who it hits hardest | The least skilled, specifically | Nearly everyone, including experts |
| What is measured | Self-rated competence in a domain | Calibration — stated confidence vs actual accuracy |
| The mechanism | Lacking the skill to recognize the skill you lack | Rich access to your own reasoning, none to what you're missing |
| What happens to experts | Often UNDERrate themselves (the flip side) | Still overconfident, just less than novices |
| Shape | A skill-level-specific curve — steepest at the bottom | A general miscalibration — present at every level |
The same failure, two different scopes
Both name the gap between how good you think you are and how good you actually are. The difference is scope. Dunning–Kruger is a claim about WHO — specifically the least competent, whose incompetence blinds them to their own incompetence. The overconfidence effect is a claim about HOW MUCH — a general miscalibration between stated confidence and actual accuracy that shows up at every skill level, experts included.
Why the least skilled overestimate the most
Dunning and Kruger's original finding is that the skills required to produce a correct answer are largely the same skills required to recognize a correct answer. A novice missing both doesn't just perform poorly — they lack the yardstick to notice they performed poorly, so confidence stays high while competence stays low. This is why the effect is strongest at the bottom of the skill curve and why genuine experts, fluent in how much the domain actually contains, often underrate themselves — the flip side of the same mechanism.
Why almost everyone is still overconfident
The overconfidence effect doesn't require ignorance of a specific domain. Ask anyone — including domain experts — to give a "90% confidence interval," and the true answer falls outside it far more than 10% of the time. The cause isn't a missing skill; it's that people have rich access to the reasoning behind a judgement but none to the evidence they failed to consider, and rarely get the clean, repeated feedback that would correct the miscalibration.
Where they overlap
A novice confidently wrong about a topic they barely understand is living out BOTH effects at once — Dunning–Kruger explains why their specific self-assessment is so inflated, and the overconfidence effect explains why their stated certainty exceeds their accuracy even once you set the skill-level question aside. The two are close enough that ReadGlobe's own overconfidence-effect page fields this exact question.
The verdict
Use Dunning–Kruger when the question is "why does this specific novice feel like an expert?" — it points at a skill-and-self-assessment gap that closes as real competence grows. Use the overconfidence effect when the question is "why is everyone, including the experts in the room, more certain than the evidence supports?" — it points at a calibration habit that requires deliberate feedback (tracked predictions, scored outcomes) to fix, competence alone will not cure it.
Frequently asked
- Is the Dunning–Kruger effect the same as overconfidence?
- No. Dunning–Kruger is a specific claim: the least skilled overestimate their ability the most, because judging performance requires the same skill as producing it. Overconfidence is broader — a general gap between stated confidence and actual accuracy that persists even among experts.
- Do experts show the overconfidence effect?
- Yes, routinely — studies on calibrated confidence intervals find experts forecasting timelines and outcomes with more certainty than their track records support. They typically do NOT show Dunning–Kruger in the same domain, since that effect is specifically about the least-skilled overestimating; experts often underrate themselves there instead.
- Which one explains a beginner who thinks they're an expert?
- Dunning–Kruger is the more precise fit — it specifically predicts that low competence produces inflated self-assessment, because the beginner lacks the knowledge to see their own gaps. The overconfidence effect would apply too, but it's the less specific of the two explanations here.
- How do you fix each one?
- Dunning–Kruger closes as real competence grows — external feedback and benchmarking against others speed that up. The overconfidence effect needs deliberate calibration: log predictions with a stated confidence level, score them against what actually happened, and widen your intervals — competence alone does not fix miscalibration.
Explore further
Keep reading
Stoicism
Your judgements are up to you; almost everything else never was.
Go deeper
The book behind this idea: Thinking, Fast and Slow by Daniel Kahneman. Hear the whole thing free — start an Audible trial and your first audiobook is on the house.
Read the full summary of Thinking, Fast and Slow →
More canonical picks:
- The Art of Thinking Clearly — Rolf Dobelli
- The Great Mental Models, Volume 1 — Shane Parrish
- Poor Charlie’s Almanack — Charlie Munger
- Super Thinking — Gabriel Weinberg & Lauren McCann
- Seeking Wisdom — Peter Bevelin
As an Amazon Associate, ReadGlobe earns from qualifying purchases and Audible trials — at no extra cost to you.
Editorial synthesis © ReadGlobe 2026, drawing on Kruger & Dunning (1999), the calibration literature on confidence-interval accuracy (Alpert & Raiffa), and Kahneman's Thinking, Fast and Slow. · Last reviewed 2026-09-11.