What the data actually shows
Research by Willis and Todorov found that people form impressions of others — including judgments like trustworthiness and competence — from faces in roughly a tenth of a second, and that giving people more time did little to change the impression, only to increase their confidence in it. First impressions are formed almost instantly and largely automatically.
To manage the flood of social information, the mind leans on stereotypes and heuristics as efficiency tools. The 'cognitive miser' idea holds that we default to mental shortcuts that conserve effort, and stereotyping is one of them: it lets us categorise people fast at the cost of accuracy. The trade is speed for precision.
A central bias here is the fundamental attribution error, described in Lee Ross's work: we tend to attribute other people's behaviour to their character while explaining our own behaviour by the situation. So the driver who cuts us off is 'a jerk,' but when we do the same we were 'having a bad day.' Combined with confirmation bias — noticing what fits the impression we've already formed — this makes snap judgments both quick to form and slow to revise.
Why this feels different from how it actually is
Snap judgments feel like accurate perception rather than guesswork because they arrive instantly and with confidence. Willis and Todorov's work suggests extra time mainly boosts how sure we feel, not how right we are — so a fast impression registers as 'reading someone clearly' even when it's mostly a shortcut.
The fundamental attribution error makes our judgments of others feel fair because we don't see their situation, only their behaviour. We have full access to our own context and almost none to theirs, so it's natural to read their actions as character and our own as circumstance — and the asymmetry is invisible from the inside.
And confirmation bias makes a judgment feel increasingly justified over time. Once an impression forms, we tend to notice the evidence that fits it and overlook the rest, so the initial snap judgment seems to keep being confirmed. That's why first impressions feel so reliable even when they're wrong — the mind quietly curates the evidence in their favour.
Extra time mainly increases how sure you feel, not how right you are — confidence and correctness aren't the same thing.
What the research says to do about it
The most realistic approach the research supports is catching the judgment rather than trying to prevent it. Because impressions form automatically and fast, you usually can't stop the first read — but you can treat it as a hypothesis to check rather than a conclusion, especially when it matters or when you notice yourself attributing someone's behaviour straight to their character.
Deliberately considering the situation is a direct counter to the fundamental attribution error. Asking what circumstances might explain someone's behaviour — the same way you'd explain your own — tends to soften the snap judgment and produces a more accurate read, because it restores the context you don't automatically see.
Staying open to disconfirming evidence helps counter the stickiness. Since confirmation bias makes us notice what fits, intentionally looking for what doesn't fit a first impression, and giving people room to surprise you, helps revise judgments that were never built on much to begin with. None of this makes you bias-free; it just keeps the shortcuts honest.
What the research says does not help
Believing you can simply stop making snap judgments doesn't help, because they form automatically and almost instantly. Treating yourself as above the process tends to make biases harder to catch, not easier, since you stop watching for them. The realistic goal is catching and checking impressions, not eliminating them.
Trusting a quick read because it feels confident is a particular trap. Extra time mainly increases confidence rather than accuracy, so the strength of a first impression is a poor guide to whether it's correct. Confidence and correctness aren't the same thing here.
And treating every snap judgment as a moral failing doesn't help either. Quick judging is a normal mental shortcut shared by essentially everyone, attached to predictable biases. Self-condemnation adds guilt without improving accuracy, whereas noticing the bias and questioning the judgment actually addresses it.
Having the impulse doesn't make you a bad person; it makes you a person with an ordinary brain.
What this looks like in real life
The driver who cuts you off
Someone cuts across you in traffic and the instant read is 'what a jerk.' When you do the same thing, you were 'just having a bad day.' That asymmetry is the fundamental attribution error: you see their action but not their situation, and only your own context is visible from the inside — so the double standard feels perfectly fair.
The first impression that keeps 'proving itself'
You meet someone briefly and file them as unfriendly. Afterwards you keep noticing the moments that fit — the short reply, the missed hello — and overlook the warm ones. The impression formed in a fraction of a second and then confirmation bias quietly curated the evidence, so a snap judgment built on almost nothing came to feel reliable.
Real numbers in context
The clearest number here comes from Willis and Todorov: impressions of others can form in roughly a tenth of a second, and additional viewing time mainly increases confidence rather than changing the judgment. That speed is the core of why quick judging is so hard to prevent — it happens well before deliberate thought gets involved.
Beyond that, this is a topic of well-established cognitive biases rather than precise statistics. The fundamental attribution error (Ross) and the use of stereotypes as cognitive efficiency are robust, repeatedly observed patterns, though their strength varies by situation and culture. The honest summary is that the speed of impression formation is well documented, and the biases attached to it are well established as tendencies rather than fixed quantities.