Optimism bias is the tendency to believe good things are likelier to happen to you than to comparable other people, and bad things less likely — regardless of the actual rates.
Neil Weinstein documented and named 'unrealistic optimism' in 1980. Tali Sharot's later work with Korn and Dolan identified the belief-updating asymmetry that sustains it and popularised the term optimism bias.
// Where it sits
The tell is comparative: not 'I'll be fine' but 'I'll be fine — more than they will'.
What it really means
Weinstein's 1980 study contains an impossibility, and the impossibility is the point. He asked 258 students to rate their own chances of 42 life events compared with their classmates' — the very people sitting around them. They rated themselves above average for the good outcomes and below average for the bad ones. Every one of them cannot be at below-average risk of divorce when they are the average.
That structure is what separates optimism bias from simply being an optimist. It isn't a mood. It's an arithmetic claim about a comparison group that can't hold for everyone making it.
The mechanism came thirty years later, and it's the part with real consequences. Sharot, Korn and Dolan gave people their actual statistical risk for eighty adverse events — divorce, illness, burglary — and measured how much their beliefs moved. Good news (your risk is lower than you thought) moved beliefs by 11.2 units. Bad news moved them by 7.7. The correcting information arrives and is only partly absorbed, which explains why a lifetime of encountering real base rates doesn't fix this. It isn't that we never hear the bad news. It's that most of us discount it slightly, time after time, and the discount compounds.
Where it comes from
Also called 'unrealistic optimism' or 'comparative optimism', the latter after the comparative design of the classic studies: they measure how you rate yourself relative to others.
Myths & misconceptions
Optimism bias is just being an optimistic person.
It's a measurable error in probability judgment, not a temperament. The signature is structural: a whole group cannot all be at below-average risk of divorce, yet Weinstein's 258 students collectively claimed exactly that.
Everyone has it, equally and permanently.
Common but not uniform — 79% of participants showed the asymmetric updating, a strong majority rather than everybody, and its size tracked individual differences in how the brain coded unwelcome information.
Compare & contrast
Optimism bias vs being an optimist
Optimism is a disposition — expecting things to go well. Optimism bias is an arithmetic error — rating your own odds better than comparable people's when they can't be. Cheerfulness isn't the bias; the impossible maths is.
| // | Optimism bias | Optimism |
|---|---|---|
| Is | A judgment error | A temperament |
| Measurable as | Impossible group maths | Self-report |
| About | Comparative risk | General outlook |
| Universal | Not quite (79%) | No |
How it connects
- Negativity Biasthey coexist: one governs information you already have, the other your forecasts.
- Survivorship Biasa diet of visible successes makes an above-average forecast feel reasonable.
Tell it apart
Questions people ask
- What is optimism bias?
- Systematically rating your own odds better than other people's. Ask a room of smokers whether they're more or less likely than the average smoker to develop lung disease and most will say less — which cannot be true of most of them.
- Why doesn't bad news correct it?
- Because we update unevenly. Shown the real population risk for 80 adverse events, people shifted their beliefs substantially when reality was kinder than expected and much less when it was harsher. Bad news simply doesn't stick the way good news does.
- Is optimism bias bad for you?
- Mixed. It supports persistence and lowers day-to-day anxiety — and it's also why renovations run months over and why people skip screening. The practical hedge is to forecast your own plan the way you'd forecast a stranger's: start from the base rate for people in your situation, then add the contingency you'd advise them to add.