Although most people see themselves as open-minded, we all tend to judge others before we truly get to know them. Prejudging people before we even properly interact with them is the norm, and the origin of the term prejudice.
And yet, we can also learn to give people the benefit of the doubt and coach ourselves to suspend rapid and stereotypical evaluations of others when we want—it’s just that it requires far more effort (thinking slow rather than fast) and . . . is by no means guaranteed to work out in the end.
So, while we all agree (at least in theory) that you cannot judge a book by its cover, the brutal reality is that there is often no time for a second good impression. If someone doesn’t connect with you or like you right away, the work needed to persuade them to change their mind or see you in a different vein may exceed your ability—and motivation—to change their views on you.
But what do we really know about the science of first impressions? A recent meta-analytic review (which synthesized the results of hundreds of empirical studies in this field) offers five important generalizations.
1. We judge the whole person before we judge the parts.
The scholarly term for this is “holistic,” in that the system outdoes the sum of its parts. This sounds positive in principle, but it also means that we fill in the blanks and make stuff up in order to fit the parts to an archetype. Indeed, one of the strongest findings from the aggregated research is that first impressions are much more holistic than we tend to think. We may believe we are independently deciding whether somebody is competent, warm, trustworthy, dominant, etc., but these judgments overlap substantially.
To be more precise, across 61 samples and 13,353 research participants examined, the correlation among different features of first impressions was 0.61, such that the whole/configural judgments we make (good versus bad, fun versus boring, interesting versus not, “hot” or not, etc.) eclipses the granularity of individual features (smart but lazy, creative but narrow-minded, polite but untrustworthy, and so on). So, if people were restaurants, we would just classify them according to good versus bad as opposed to being capable of outlining what dishes are good, what the service is like, and whether the parking is convenient. Inevitably, this reflects our need to take mental shortcuts and simplify the complexity of human nature into unfair and inaccurate stereotypical depictions of human beings.
In fact, this effect is among the oldest found in social psychology, often referred to as the halo effect, and it is not a minor cognitive bug, but rather a major feature of our mental operating system. Once we broadly like what we see, we tend to see more specific virtues too. And once we dislike what we see, we will go to great lengths to see more negative traits, even if it means making them up. In fact, even when we may feel we have made fine-grained assessments of others, the power of categorical, holistic generalizations rules supreme.
2. How you say it matters more than what you say.
Although it’s a common misconception that nonverbal communication matters more than verbal communication (if this were true, try spending a week in Japan or China if you don’t speak their language, and see how you get by), the content of what we communicate is typically less important than the style in which we do. In other words, people are less impacted by what you say, and more by how you say it, especially when it comes to first impressions. More precisely, during initial interactions, observable communication behavior carries considerably more weight than the substantive information being conveyed about the person, which is why communication training and coaching can be more effective than people think.
The average layman assumes that when they have a clear idea and express it with conviction, it will make the desired impact on the audience; yet, just like good actors can convincingly embody their characters, well-coached communicators can maximize the impact they make, irrespective of the content they convey. To be fair (and back to point 1), any contextual insights your audience may have about you will likely be part of the “how you say it” bucket: for example, if your audience is aware of your expertise, track record, credentials, or high professional status, whatever you say may be deemed deep and convincing, as with Peter Sellers’s Chance the Gardener character in Being There.
3. First impressions matter more for relationships than performance
Overall first impressions are more strongly related to relational outcomes than achievement-oriented outcomes. In other words, they appear particularly consequential for things such as liking, relationship quality, trust, and future interpersonal contact, rather than merely who gets judged as competent or performs well. So, when you are interviewing for a job, the impression you make on others may matter more for impacting your future coworker relationship with the interviewers than, well, whether you get the job or not (unless the interviewers really just want a friend or someone who is fun to hang out with, in which case they should not be interviewing in the first place!).
Once you are on the job, though, whether you get mentored, trusted, befriended, included, listened to, or invited to collaborate can depend disproportionately on early intuitive judgments. So, perhaps the takeaway message here is to perform well on the interview so you can get the job—which will hopefully depend more on substance than style—but then leverage the power of charm and impression management to make a favorable first impression on your colleagues as soon as you start.
4. You can make a second impression. Just don’t expect a completely different one.
Although first impressions are more impactful and durable in their impact than they should be, they can actually change. Still, they are remarkably sticky in the long term. Across 22 samples involving 4,258 people, the correlation between initial impressions and impressions measured later was 0.47. Most strikingly, when more than a month had elapsed, it was still 0.51. To put things into perspective, these correlations are considerably larger than those between job satisfaction and job performance (0.30), conscientiousness and job performance (.020–0.22), or even self-efficacy and job performance (0.23): relationships most people would intuitively expect to be substantial.
So yes, people may update their views on us, but they don’t start again from zero. The first impression becomes an anchor against which subsequent information is interpreted. In other words, if you are lucky, you may get a second chance, but chances are you will be judged in a very similar vein, whatever you do. This is also where confirmation bias and self-fulfilling prophecy become interesting mechanisms: People may selectively notice subsequent information that confirms what they initially thought, while their own behavior toward the person can help elicit behavior consistent with that impression.
5. First impressions fade, but they never disappear.
There is an important distinction between the stability of the impression itself and its ability to predict later outcomes. The latter does weaken. When outcomes occurred less than a minute after the impression, the corrected association was 0.48; when outcomes occurred more than a month later, it fell to 0.28. That difference was statistically significant. But 0.28 is hardly trivial, and importantly, its credibility interval remained entirely above zero. The authors characterize this as a medium-size relationship even after more than a month. So yes, first impressions are important, but not because they permanently determine what everybody thinks of us. Their immediate influence declines as reality supplies more information, yet the original judgment itself remains surprisingly stable and continues predicting consequential outcomes long afterward.
An important overall caveat ought to be considered, which mitigates the impact of the above five findings. Fortunately (perhaps we are not that narrow-minded after all?), more exposure reduces the power of individual cues. The relationship between informational cues and impressions was 0.43 when exposure lasted under a minute, but only 0.25 when exposure exceeded an hour. That’s a lovely empirical version of “the best antidote to a first impression may simply be more information.” In other words, the more time we spend with people, the harder it becomes for any single feature, gesture, comment, or behavioral cue to dominate our judgment of them. Reality gradually crowds out the caricature: Additional information gives us more opportunities to notice inconsistencies, exceptions, and qualities that our initial snapshot simply could not capture. First impressions may therefore be surprisingly sticky, but they are not immune to one of the oldest and most effective debiasing technologies available to us: actually getting to know someone.
The overall conclusion, then, is much more nuanced than we may like, especially if we anticipated a simple, categorical answer. In fact, first impressions may matter about as much as popular psychology tells us, but for somewhat different reasons. We aren’t especially sophisticated at making them; communication style (which can be artificially performed) disproportionately drives them; they contaminate supposedly separate judgments about people; and they persist, though not in strength or impact. And fortunately, they are not destiny: Their influence on actual outcomes declines as time and additional information accumulate.
