How we judge other people

Someone walks into the room. Before they’ve opened their mouth, your brain has already done its work. Height, clothing, posture, the set of their face. A few seconds. That’s all it takes.

This isn’t a character flaw. It’s not something to feel guilty about. The human brain is built to make decisions quickly, with minimal information, and to do it constantly. Is this person safe or dangerous? Friend or threat? Trustworthy or not? These questions, across hundreds of thousands of years of evolution, kept people alive. Today, the same machinery runs in job interviews, first dates, and chance encounters on the street.

The problem is that this system is optimised for speed, not accuracy. And we rarely know which of our confident judgements are wrong.

This piece is about that machinery. How it works, where it fails, and why understanding it matters more than most people realise.

Why We Judge at All: The Evolutionary Logic

Social psychologists have been studying social judgement seriously since the 1950s, but the behaviour itself is ancient. Fritz Heider, working at the University of Kansas, argued in 1958 that humans are natural “naïve psychologists.” We have an instinctive need to explain the behaviour of others. We can’t help it. When someone acts in a particular way, we want to know why.

In early human groups, reading others accurately was a matter of survival. Recognising an ally from a rival, spotting deception before it cost you, knowing who to trust in a crisis, all of this required fast, confident social assessment. The individuals who did it well had better odds. That capacity became part of us.

But evolution doesn’t optimise for truth. It optimises for survival, which means “good enough, fast enough.” A snap judgement that was right 70 per cent of the time was sufficient to keep someone alive. Today, that same 70 per cent system operates in a world of vastly greater complexity, and we’re often blind to the 30 per cent it gets wrong.

Attribution Theory: The Stories We Tell About Other People

One of the most important frameworks for understanding social judgement is attribution theory, which Heider first sketched and Harold Kelley and Bernard Weiner later developed in detail.

Attribution is the process of deciding what caused someone’s behaviour. Did they act that way because of who they are, what psychologists call an internal attribution? Or because of the circumstances they were in, an external attribution?

When a colleague turns up late to a meeting, your brain immediately reaches for an explanation. “He’s disorganised” or “the trains must have been bad.” Which one you land on depends on whether you know the person, whether you’ve seen this before, and, most critically, where you sit on a very well-documented bias called the fundamental attribution error.

The Fundamental Attribution Error: It Catches Almost Everyone

Lee Ross coined the term in 1977, but the decades of research since then have confirmed it so thoroughly that it’s now one of the most replicated findings in social psychology.

Here’s the short version: when we explain other people’s behaviour, we default to their personality. When we explain our own behaviour, we default to circumstances.

Someone is rude to a waiter? They’re an inconsiderate person. I’m rude to a waiter? I’m having a terrible day, I barely slept, the service was slow. Someone fails to meet a deadline? Lazy. I fail to meet a deadline? My workload has been unreasonable for weeks.

The asymmetry is consistent and largely automatic. We observe other people only from the outside, so their behaviour is all we have to go on. We experience ourselves from the inside, with full access to context, pressure, and competing demands. That difference in perspective produces a difference in attribution that most of us never notice.

Cross-cultural research has found that the error is stronger in individualistic societies (the United States, Western Europe) and somewhat weaker in collectivist societies (Japan, China, South Korea), where people are more accustomed to factoring in social context. But it doesn’t disappear anywhere.

The practical implication is significant. Every time you conclude quickly what someone “is like,” you’re probably explaining a behaviour with a personality trait while ignoring conditions you don’t know about.

Correspondent Inference Theory

Jones and Davis developed a related idea in 1965. They argued that we’re more likely to attribute behaviour to someone’s character when the behaviour was freely chosen, has distinctive consequences, and goes against social expectations.

So if someone is polite in a job interview, that tells you relatively little. Politeness in that context is practically compulsory. But if someone is calm and kind in a genuinely high-stress situation where they had every reason to lose their temper, that’s informative. It gives you real signal about who they are.

In practice, people apply this logic far less carefully than the theory suggests. We draw character inferences from highly constrained behaviour all the time: from the professional register someone uses in a formal email, from the uniform they wear, from the role they happen to be playing when we first encounter them. The context that made the behaviour near-inevitable goes unexamined.

Heuristics: The Mental Shortcuts That Shape Every Judgement

Daniel Kahneman and Amos Tversky, across a remarkable body of work in the 1970s and 1980s, showed that the brain uses cognitive shortcuts to make judgements under uncertainty. They called these heuristics. They’re efficient, often accurate, and capable of producing systematic errors that are predictable and consistent.

Three heuristics matter most in social judgement.

The Availability Heuristic

The brain estimates the likelihood of something based on how easily examples come to mind. If vivid examples of a person’s bad behaviour surface quickly in memory, we assume that behaviour is frequent or typical, even if those examples are memorable precisely because they were unusual.

One standout failure sticks. A hundred unremarkable successes fade. When it comes time to assess someone, the failure arrives in conscious thought first and carries more weight than it should.

Media amplifies this in ways that distort social judgement at scale. When news coverage is dominated by violent crime, people’s estimates of how dangerous their neighbourhood is, or how trustworthy certain groups are, draw on a deeply skewed sample of reality. The events that get covered aren’t representative of the world. But they’re the ones that come to mind fastest, and so they become the basis for confident assessment.

The Representativeness Heuristic

We judge how likely it is that a person belongs to a particular category by how closely they resemble our mental image of that category. If someone’s appearance, accent, or manner matches our internal template for “expert,” “criminal,” or “trustworthy,” they get sorted accordingly, quickly and without much deliberation.

Research in the American legal system has shown that when suspects more closely resembled jurors’ mental image of a perpetrator, verdicts came faster and sentences were longer, even when the evidence was the same. This heuristic operates in job interviews, classrooms, and clinical assessments. It’s not malicious. It’s largely unconscious. But its consequences are real.

The Anchoring Heuristic

The first information we receive about someone functions as an anchor. Everything that follows gets interpreted in relation to it. Solomon Asch demonstrated this elegantly in 1946: when a list of personality traits moved from positive to negative, overall ratings of the person described were higher than when the identical traits were presented in the opposite order.

The primacy effect in social judgement is strong. Once someone has been categorised as “intelligent” or “unreliable,” later information tends to be assimilated into that frame. Their mistakes get explained away as exceptions. Their competence gets attributed to luck. The initial impression doesn’t just influence subsequent judgements; it filters the interpretation of new information.

The Halo Effect: When One Thing Colours Everything

Edward Thorndike documented this in 1920 and it’s remained one of the most studied phenomena in social psychology since. When we perceive one strongly positive quality in a person, we tend to assume other positive qualities follow. The reverse also holds: a single negative trait casts a shadow across the whole assessment.

Physical attractiveness is among the most powerful of these halos. Research consistently finds that attractive people are perceived as more intelligent, more honest, more competent, and more trustworthy, without any evidence for any of these assessments. These biases show up in contexts where they shouldn’t: sentencing decisions in criminal trials, performance reviews at work, marks given to written work when a photo is attached.

In one study of mock sentencing decisions, attractive defendants received meaningfully lighter sentences than unattractive defendants for the same offence. The conscious mind would reject this reasoning immediately. The unconscious mind had already acted on it.

The horn effect works in the same way but in the opposite direction. One credible instance of dishonesty reduces perceived competence, warmth, and reliability across domains that have nothing to do with honesty. The brain treats a negative trait as diagnostic of the whole person, not of a specific behaviour in a specific situation.

Thin Slices: Why First Impressions Hold

Nalini Ambady and Robert Rosenthal published research in 1992 that still provokes genuine surprise. They showed students silent 30-second clips of teachers in the middle of a lecture, then asked for an assessment. Those 30-second assessments correlated strongly with evaluations from students who had spent an entire semester with the same teacher.

When they replicated the study using six-second clips, the correlation held.

Ambady called these “thin slices” of behaviour. The brain processes non-verbal cues, body language, vocal tone, microexpressions, with considerable accuracy in some domains. A physician’s tone of voice in a brief interaction predicted whether they were ever sued for malpractice better than the content of what they said.

This is genuinely interesting, not just as a curiosity. In contexts where we have real expertise and where the signal is real, fast judgements can outperform slower, more deliberate ones. But the same machinery applies when the signal isn’t real, when appearance, accent, or physical characteristics are driving the assessment. The speed doesn’t discriminate between accurate and inaccurate inputs.

Confirmation Bias: Seeing What We Expect to See

Peter Wason’s 1960 experiment on hypothesis testing produced one of the most important findings in cognitive psychology. People don’t naturally look for evidence that would disprove their beliefs. They look for evidence that confirms them. This tendency, confirmation bias, is pervasive in social judgement.

Once a belief about someone has formed, the filter is on. Behaviour that fits the belief is noticed, remembered, and weighted. Behaviour that contradicts the belief is downplayed, explained away, or simply not encoded. The belief becomes self-reinforcing not because it’s correct, but because the way we process information makes it very difficult to dislodge.

In close relationships, this plays out in ways that compound over time. The partner who has been labelled “critical” finds every comment they make heard through that lens. The colleague who has been written off as “not a team player” finds their collaborative gestures go unnoticed. The brain is doing what it’s built to do: reducing cognitive load by using existing models. But those models can calcify into something that no longer reflects who the person actually is.

Research in workplace settings has shown that managers who have formed a negative impression of an employee give them lower performance ratings even when independent observers score the work more highly. The assessment isn’t of the work. It’s of the person, filtered through a pre-formed judgement.

The Role of Context in Social Judgement

A well-replicated finding in social psychology: the same person, in different contexts, produces different judgements. Role and setting shape perception in ways people rarely notice.

In one study, students who observed a conversation between an interviewer and an interviewee rated the interviewer as more intelligent, even when both parties asked and answered the same questions. The role itself carried the inference. Power and position produce perceived competence, not demonstrated competence.

This matters because it means our judgements of people are partly judgements of situations. The manager who seems authoritative in a meeting and the same person seen off-duty at a supermarket are the same individual. The assessments that spring up around them are not. Understanding that our impressions of people shift with context makes it harder to mistake a contextual observation for a character trait.

In-Group and Out-Group: How Group Identity Distorts Judgement

Henri Tajfel’s research in the 1970s produced a finding that is both simple and deeply uncomfortable. Even when groups are formed entirely at random, with no shared history or meaningful distinction, members consistently favour their own group and view the other group more negatively.

This is the basis of social identity theory. We categorise ourselves as members of groups, and group membership shapes perception. Behaviour by an in-group member gets a sympathetic reading: the context was difficult, the intention was good. Behaviour by an out-group member confirms expectations: this is how they are.

The out-group homogeneity effect is the specific form this takes in social judgement. Members of the out-group are perceived as more similar to each other than members of the in-group are. “They’re all the same” is an in-group perception of the out-group. Within the in-group, individuality and complexity are readily acknowledged.

This operates across lines of ethnicity, religion, nationality, profession, political affiliation, and even preferences that carry no real weight at all. The mechanism doesn’t distinguish between important and trivial categories. It runs on any available distinction.

The Emotional Dimension of Judgement

For a long time, research treated judgement as primarily a cognitive process. The evidence now points elsewhere. Emotion is not a distortion layered on top of judgement. It’s a constituent part of it.

António Damásio’s work with patients who had suffered damage to the brain’s emotional processing regions produced a striking finding. These individuals, who retained their reasoning capacities entirely, became markedly worse at social decision-making. Without emotional responses to draw on, they couldn’t navigate the texture of social situations. The somatic markers, physical feelings that carry information about value and risk, were gone, and judgement suffered.

But emotion can also introduce systematic error. The mood congruency effect is one of the cleaner demonstrations: people in a good mood judge others more positively; people in a low mood judge others more negatively. The same person, assessed on two different days by the same assessor, receives different ratings based on something that has nothing to do with them.

Israeli researchers analysed parole decisions by experienced judges across hundreds of cases. At the start of the day and after food breaks, approval rates were around 65 per cent. By the end of a session, just before a break, they had fallen close to zero. Fatigue and depletion weren’t just affecting energy. They were affecting who got released. This isn’t a story about bad judges. It’s a story about the limits of human cognitive systems.

Theory of Mind: Modelling Other People’s Inner Lives

To judge others, we need to be able to represent their mental states. This capacity, which psychologists call theory of mind, develops around age four to five in most children, and its absence or atypicality is associated with significant social difficulties.

The classic test is the false belief task. A doll places a marble in a box and leaves the room. Another doll moves the marble to a different location. When the first doll returns, where will she look for the marble? A child with developed theory of mind knows she’ll look in the original box, because she doesn’t have the information the child has. A child without this understanding says she’ll look in the new location, because that’s where the marble actually is.

This capacity underpins empathy, social reasoning, and the ability to construct accurate models of other people. But it has real limits in practice. When we model other people’s minds, we tend to use our own minds as the template. We assume they see what we see, feel what we’d feel, care about what we care about. This produces a well-documented error called the false consensus effect.

The False Consensus Effect

Tversky and Kahneman, along with later researchers, showed that people systematically overestimate the extent to which their views, preferences, and behaviours are shared by others. They expect others to agree with them more than others actually do.

In social judgement, this matters because we assess other people partly through the lens of what we ourselves would do. When someone’s behaviour diverges from what we’d have done, we read it as informative about their character, when it may simply reflect a difference in values, constraints, or experience that we haven’t accounted for.

Moral Judgement: Why We Condemn

Jonathan Haidt’s 2001 paper on moral intuitionism changed how psychologists think about moral judgement. His argument, backed by evidence, was that moral verdicts typically arrive before the reasoning does. We feel that something is wrong, and then we construct a justification. The reasoning process functions more like a lawyer defending a pre-formed verdict than a judge weighing evidence.

His demonstration used scenarios that most people found morally distasteful but couldn’t articulate a rational objection to. Participants said things like “I know it’s wrong, I just can’t explain why.” Haidt called this “moral dumbfounding.” The emotional reaction was fast and confident. The reasoning was, in his term, confabulated.

This has significant implications for how we understand the judgements we make about other people’s behaviour. When we condemn someone, often quickly and with considerable certainty, the condemnation typically precedes the analysis. The feeling of moral clarity doesn’t guarantee that the reasoning supporting it is sound.

The Expectation of Moral Consistency

We tend to assume that character is stable. Someone who cheats once is “a cheat,” not “someone who cheated in a specific situation.” Someone who lies once is “dishonest,” not “someone who lied under particular pressure.”

Walter Mischel’s research on personality consistency challenged this directly. Behaviour across different situations is often less consistent than people expect. The person who took money in one context doesn’t necessarily take it in another. Moral behaviour is context-sensitive to a degree that our intuitive models of character don’t capture.

This is one of the places where our model of other people fails most reliably. We treat personality traits as stable causal forces when they’re often better described as tendencies that manifest differently depending on circumstances, relationships, and situational pressures.

Social Judgement Online: Speed Without Context

Social media has intensified every distortion described above. A tweet, a clip taken out of context, a photo from a single moment: these are what we use to form and broadcast judgements about people, with less information and less reflection than almost any other form of social encounter.

The fundamental attribution error runs particularly hard online. When someone’s entire visible identity is their posts, it’s almost impossible not to treat those posts as the person. The circumstances of their life, the pressures they were under, the fuller context of a conversation, all of it is absent.

The speed of judgement in online environments is also essentially unchecked. The time between exposure to information and expression of a verdict is very short. The social dynamics of platforms amplify the most emotionally charged responses, which tend not to be the most carefully considered ones. The result is a social environment in which confident, rapid condemnation is the norm, and revision or nuance is structurally difficult.

This isn’t a comment on any particular controversy. It’s about the epistemic conditions under which social judgement happens online, conditions that make the biases described above more likely to operate, not less.

Individual Differences: Not Everyone Judges the Same Way

The biases covered here affect almost everyone, but not equally. A few factors are worth understanding.

Need for cognitive closure is a personality variable that describes the drive to reach a definite answer and resist ambiguity. People high in this trait tend to form social judgements more quickly, hold onto them more firmly, and show stronger correlations with stereotyping and prejudice. It’s not that they’re less intelligent. They’re more uncomfortable with uncertainty, and quick judgements reduce that discomfort.

Intergroup contact, under the right conditions, reduces stereotyping and improves social judgement. Gordon Allport’s 1954 contact hypothesis proposed that contact between groups reduces prejudice when it happens under conditions of equal status, common goals, cooperation, and institutional support. Decades of research have broadly confirmed this. Knowing people from groups you might otherwise judge via stereotype makes the stereotype harder to apply automatically.

Metacognitive awareness helps to a degree. People who know about confirmation bias don’t eliminate it, but they show some ability to check themselves in specific, deliberate ways. Knowing that you’re prone to a particular error doesn’t immunise you against it. But it does create the conditions in which you might occasionally catch it.

How Psychology of Judgement Applies in Clinical Work

For psychologists working clinically, patterns of social judgement are themselves clinically significant.

Paranoid personality disorder involves a systematic bias towards interpreting ambiguous behaviour as threatening or hostile. Small, neutral events are read as targeted. The bar for concluding hostile intent is very low. Understanding this as a distortion in the attribution process, rather than as a deliberate choice, changes how it can be addressed therapeutically.

Borderline personality disorder involves a pattern that psychologists sometimes describe as idealisation and devaluation: the same person swings from being perceived as entirely good to entirely bad. This is related to difficulties with what’s called whole object relations, the capacity to hold a complex, mixed picture of another person simultaneously. The judgement system, in this context, struggles to tolerate ambiguity and defaults to extremes.

Research on depression and social judgement has produced a counterintuitive finding that’s been replicated enough to take seriously. In certain conditions, people experiencing depression produce more accurate social judgements than people who are not depressed. This has been called depressive realism. The proposed mechanism is that non-depressed people maintain a mild positive bias in social perception that helps them function. When that bias is absent, reality is assessed more accurately in some domains. The costs of that accuracy are considerable, but the finding complicates any simple story about “clear” versus “distorted” social perception.

Can We Actually Judge Better?

This is the practical question, and the honest answer is: somewhat, under specific conditions, with deliberate effort.

Slowing down helps, but less than people hope. The brain forms an initial judgement and then revises it. The revision requires cognitive resources. When we’re tired, under pressure, or cognitively overloaded, the revision doesn’t happen. The initial judgement stands, dressed up as considered assessment.

Seeking contextual information is effective. Knowing what circumstances someone was operating under consistently improves the accuracy of behavioural attribution. In practice, people rarely have time or motivation to gather this information before forming a view. But in contexts where accuracy genuinely matters, investing in context is one of the most reliable things you can do.

Active perspective-taking, Adam Galinsky and Gordon Moskowitz showed in 2000, measurably reduces stereotyping. The key word is active. Abstractly reminding yourself that people are complex doesn’t do much. Deliberately imagining the specific circumstances of another person’s life, from the inside, does.

Awareness of specific biases gives you narrow but real advantages. Knowing that you’re more likely to judge others harshly when you’re hungry, tired, or in a bad mood doesn’t change the biology. But it can change whether you make a significant judgement in those conditions. Recognising the halo effect when you catch yourself liking everything about someone you’ve just met is a useful pause. Noticing that your confidence in a first impression is running ahead of your evidence is a productive signal.

None of this produces perfect judgement. The mechanisms described in this article are deeply embedded. What it does is create a little space between stimulus and response. That space is where accurate social perception becomes more possible.

The Honest Conclusion

We will always judge other people. There’s no version of social life in which this doesn’t happen. The question is whether we do it with any awareness of the system doing the judging.

The brain’s social judgement machinery is fast, confident, and often wrong in predictable ways. It defaults to character when circumstances are the real explanation. It anchors on early information. It confirms what it already believes. It sees out-groups as homogeneous and in-groups as diverse. It responds to emotional state in ways that have nothing to do with the person being assessed.

Knowing this doesn’t fix it. But it does mean that confident is no longer a reliable signal that you’re right. It means that the stories we tell about other people, the ones that feel most obvious and self-evident, are exactly the ones worth examining with some care.

That’s not a call for chronic self-doubt. It’s a call for a particular kind of intellectual honesty about how much we actually know when we think we’ve got someone figured out.

Further reading: Thinking, Fast and Slow by Daniel Kahneman; The Righteous Mind by Jonathan Haidt; Mistakes Were Made (But Not by Me) by Carol Tavris and Elliot Aronson; The Person and the Situation by Lee Ross and Richard Nisbett.

Frequently Asked Questions