10.7 — First Impressions and Thin Slices

Nalini Ambady and Robert Rosenthal showed people thirty-second silent video clips of university lecturers they had never met and asked them to rate the teachers.

Those ratings correlated substantially with the end-of-term evaluations given by students who had sat through an entire course.

Then they cut the clips to six seconds. The correlation held.

This is the thin slice finding, and it is real. Brief exposure produces judgements that agree with judgements based on far more information.

Note carefully what that does and does not show. It shows that a six-second impression predicts what a term's worth of students will think. It does not show that either judgement is correct about teaching quality, and student evaluations are themselves known to be affected by warmth, attractiveness and confidence. Both groups may be reading the same superficial signal accurately.

That caveat runs through this entire page.

What thin slices genuinely get right

The research at zero acquaintance — strangers rating strangers — finds a consistent pattern.

Extraversion is read accurately, and it is the trait most reliably detected from brief exposure. It is expressed in behaviour that is immediately visible: volume, energy, how much someone talks.

Conscientiousness is read moderately well, partly from appearance and tidiness.

Openness, somewhat.

Agreeableness and emotional stability, poorly. These require watching how someone behaves under conditions that are not present in a brief pleasant encounter.

The general principle: traits with visible behavioural signatures are read accurately, and traits that only show under pressure are not.

Which is exactly the wrong way round for most decisions. The things you most need to know about a person — whether they are honest, whether they are kind when it costs them, whether they can be relied on — are the ones a first impression cannot see.

What first impressions are actually made of

Six inputs, and only two of them carry real information.

Facial structure (10.1). No predictive value, large effect.

Attractiveness. The halo effect. Large, and it leaks into judgements of competence, honesty and intelligence.

Similarity to you. Accent, background, dress, manner, points of shared reference. People rate people like themselves more positively, reliably, in every study that looks.

Warmth signals. Smiling, orientation, tone (10.3). Some real information, because these are behaviour rather than structure.

Confidence signals. Posture, voice, hesitation. Substantial effect on the impression and weak relationship to actual competence — a finding with real consequences in hiring.

And energy. Which is a genuine read on extraversion.

So the impression is built mostly from structure, attractiveness and similarity, and slightly from behaviour. The fact that it correlates with other people's impressions is not evidence that it is accurate — it is evidence that everybody is reading the same superficial features.

Why the first impression is so hard to shift

Three mechanisms, compounding.

Primacy. Early information is weighted more heavily than later information in forming an overall judgement. Solomon Asch demonstrated this with lists of adjectives: the same words in a different order produce different impressions.

Confirmation (1.4). Once formed, the impression filters what you notice. Ambiguous behaviour is read as consistent with it. Contradicting behaviour is treated as the exception.

And behavioural confirmation. The most powerful of the three. You behave differently towards someone you have judged, and that behaviour produces the response you expected (2.6). Treat someone as cold and you are slightly cold, and they are cold back.

The practical consequence: a first impression tends to become self-verifying, so the fact that it "turned out to be right" is much weaker evidence than it feels.

What this means for being judged

You are on the other side of this constantly, and the mechanics are usable.

The first few seconds carry disproportionate weight, so they are worth attending to in situations that matter — an interview, a first meeting with a client, meeting your partner's family.

Warmth before competence. Research on how people evaluate others finds that warmth is judged first and weighted more heavily than competence in overall impressions. Someone who seems capable and cold is trusted less than someone who seems warm and capable, and the ordering matters: warmth first.

The three signals that are free: turn towards the person, look at them, and greet them as though you are pleased to see them. These are not manipulation and they cost nothing.

And if your resting face reads as unfriendly (10.2), the compensation is not to change your face but to add an explicit signal — a greeting, a question, a stated interest.

The uncomfortable half: attractiveness, accent, name and background all affect how you are read, in ways nobody chose and most people will not admit to. This is worth knowing so that you can attribute a cold reception correctly, rather than concluding something about yourself.

When to trust a first impression

Trust it as a hypothesis about their current state. Nervous, hurried, warm, distracted. Usually accurate (10.5).

Trust it moderately about extraversion. You can tell.

Do not trust it about honesty, kindness, competence or reliability. Those require the evidence in 10.11.

And treat a strong negative first impression with particular suspicion, for two reasons. Negative impressions are stickier than positive ones, and they trigger the behavioural confirmation loop most strongly. The person you decided you disliked in the first minute is the person you have least data about, because you have been avoiding them ever since.

The one thing a first impression is genuinely good for

Detecting that something is wrong.

The rapid threat-detection system (2.1) is coarse and it is fast, and in situations of actual physical danger it is worth listening to without requiring justification.

The rule that follows is narrow and important: in a situation where your safety might be at issue — a person, a street, a room — act on the feeling first and analyse later. The cost of leaving unnecessarily is embarrassment. The cost of staying is not symmetrical.

This is the one place in this Part where the instruction is to trust the instant read, and it applies to physical safety specifically, not to whether somebody is trustworthy in general.

What to do with this

Label your first impression as a hypothesis, in words, and date it. First impression: seems arrogant. Then check it in a month. Doing this ten times will calibrate you better than anything else in this Part, and most people find their hit rate is far lower than they expected.

Deliberately seek disconfirming evidence about anyone you disliked immediately. One real conversation, about something they know about. This is the specific antidote to the confirmation loop, and it costs one coffee.

Manage the first seconds where the stakes are real — warmth first, then competence.

Attribute your own bad first impressions correctly. Some of it is you, and some of it is structure, similarity and the other person's day.

And keep the safety exception separate. Everything on this page argues for holding your snap judgements loosely, except when the judgement is that you are in danger. There, leave first.

Next: 10.8 asks the question underneath the whole Part — when you should judge a person at all, and when the correct answer is not to.