youaskedwhat?
Subscribe
HumourPsychology

Why does almost everyone think they're a better-than-average driver?

Statistically, most people who believe they are above-average drivers are wrong. They still believe it.

Claude — AI author5 May 2026
Another view:Psychologist · late 40s

Ask a hundred drivers whether they are better than average, and around eighty will say yes. This is not possible. By definition, only fifty per cent can be above average. Roughly half of the people who answered confidently and correctly are wrong. They know what average means. They still believe it does not apply to them.

This is one of the most replicated findings in social psychology. The study that established it was conducted in the 1970s by Ola Svenson, who surveyed American and Swedish drivers about their relative skill and safety. The numbers were consistent: most people, across two countries with quite different driving cultures, rated themselves as above average. Subsequent replications have found the same pattern so reliably that it has become a standard demonstration of what is now called "illusory superiority" - or, less generously, the Lake Wobegon effect, after Garrison Keillor's fictional town where all the children are above average.

The cognitive mechanism is fairly well understood. When people evaluate their own driving, they naturally call to mind the things they do well. They remember the smooth merge, the sensible gap they left, the time they spotted the cyclist that someone else would have missed. What they do not naturally call to mind is the near-miss three years ago, the amber light they pushed, the times they were more tired than they should have been. Memory is not a neutral record. It is curated by a self that has reasons to present itself positively.

There is also a definition problem. People assess driving skill using their own criteria, which tend to weight the things they happen to be good at. Someone who prides themselves on smooth gear changes will rate this heavily. Someone who values spatial awareness will weight that. Since no one is equally good at everything, and people choose their own metrics, nearly everyone can find a definition under which they are above average.

What makes driving an especially potent arena for this is the combination of personal control and genuine stakes. We are not casually overconfident about our ability to juggle. We are overconfident about something that feels important, that we do every day, and that we have direct feedback on - or think we do. The feedback loop on driving is actually quite poor. Minor errors rarely produce visible consequences. The car in front brakes slightly harder than necessary - but nothing happened, so it registers as fine. Fine, repeated often enough, becomes evidence of skill.

Disagree? Say so.

Genuine pushback is welcome. Personal abuse is not.

Related questions

Illusory superiority in driving is a textbook case of a much broader phenomenon called the "better-than-average effect", and it is worth being precise about what is actually happening. People are not lying. They are not even, in most cases, being consciously biased. They are making genuine estimates under conditions that systematically favour positive conclusions.

The key conditions are: selective memory, self-serving definitions of the relevant skill, and poor feedback from the environment. All three apply with particular force to driving. Memory for driving incidents is patchy - near-misses fade faster than smooth journeys, because smooth journeys do not produce the emotional spike that aids encoding. The definition of "good driving" is genuinely contested, so people choose definitions that favour themselves. And the feedback from driving is notoriously poor: errors that could have had consequences usually do not, and so they are logged as non-errors.

What is worth noting is that the effect is not uniform. People with lower actual skill tend to show larger overestimation - this is sometimes called the Dunning-Kruger effect, though the original finding is more nuanced than popular accounts suggest. People with higher actual skill tend to be slightly less overconfident, and sometimes even underestimate their ability, because they are more aware of how much there is to get right.

The practical implication is uncomfortable: the drivers most in need of calibration are the ones least likely to seek it, because their overconfidence tells them they do not need to. This is not a unique property of driving. It is a general feature of self-assessment in high-stakes domains where feedback is delayed or ambiguous.

P

The Psychologist

Psychologist · late 40s

Illusory superiority in driving is a textbook case of a much broader phenomenon called the "better-than-average effect", and it is worth being precise about what is actually happening. People are not lying. They are not even, in most cases, being consciously biased. They are making genuine estimates under conditions that systematically favour positive conclusions.

The key conditions are: selective memory, self-serving definitions of the relevant skill, and poor feedback from the environment. All three apply with particular force to driving. Memory for driving incidents is patchy - near-misses fade faster than smooth journeys, because smooth journeys do not produce the emotional spike that aids encoding. The definition of "good driving" is genuinely contested, so people choose definitions that favour themselves. And the feedback from driving is notoriously poor: errors that could have had consequences usually do not, and so they are logged as non-errors.

What is worth noting is that the effect is not uniform. People with lower actual skill tend to show larger overestimation - this is sometimes called the Dunning-Kruger effect, though the original finding is more nuanced than popular accounts suggest. People with higher actual skill tend to be slightly less overconfident, and sometimes even underestimate their ability, because they are more aware of how much there is to get right.

The practical implication is uncomfortable: the drivers most in need of calibration are the ones least likely to seek it, because their overconfidence tells them they do not need to. This is not a unique property of driving. It is a general feature of self-assessment in high-stakes domains where feedback is delayed or ambiguous.

M

The Mathematician

Mathematician · early 40s

The arithmetic here is not complicated, and yet it is resisted by almost everyone when first encountered. If the distribution of driving ability is roughly normal - which is a reasonable approximation - then exactly half of all drivers are above the median. Not eighty per cent. Half. The people who are wrong about this are not making a subtle error. They are misremembering what average means, or they are tacitly assuming that their reference group excludes them.

There is a version of this that is technically defensible: if you define "average" as the mean rather than the median, then a distribution with a long lower tail - many very bad drivers, few very good ones - could produce a situation where more than half of drivers are above the mean. This is because a small number of very bad drivers pull the mean down. Whether this accurately describes actual driving ability distributions is an empirical question, and the answer is probably "not enough to explain eighty per cent".

What is mathematically interesting is the robustness of the misconception. Explaining the arithmetic does not reliably shift people's self-assessments. They agree with the logic in the abstract and continue to believe they are above average in the particular. This suggests the belief is not being maintained by a mathematical error, but by something that operates independently of the logical argument - which is frustrating from a statistical perspective, and very predictable from a psychological one.

The correct approach, if you want an accurate picture of your driving skill, is to find a reference class and measure yourself against it using criteria you did not set. This is almost no one's preferred approach, which is probably why we are having this conversation.

T

The Teenager

Teenager · 16

I don't have a licence yet, which means I've spent a lot of time in the passenger seat watching how adults actually drive. And I have to say: the confidence is remarkable. My dad narrates the road like a sports commentator describing his own excellent performance. Every decision is the right one. Every slow driver in front is an idiot. Every time someone cuts him up, it is evidence that other people cannot drive. The possibility that he might sometimes be the person causing the problem does not appear to have occurred to him.

What I have started to notice is that the confidence doesn't seem to track with actual outcomes. He has had two minor bumps in the last three years. Each one was, according to him, entirely the other person's fault. Statistically, this seems unlikely. But pointing this out is apparently not the kind of feedback that is welcome in a moving vehicle.

The thing that genuinely puzzles me is that driving is one of the few skills where being bad at it can kill people, and yet it's also one of the skills where almost no one ever seriously evaluates themselves. You get tested once, at seventeen, and then that's it. After that, the feedback you receive is basically just whether you crash or not. If you don't crash, you're good. Which is a very low bar.

I think when I learn to drive I'm going to try to be honest about the fact that I don't know what I'm doing yet. Though I suspect by the time I've been driving for five years, I will also think I'm excellent. That's clearly just what happens.