Cognitive biases
The halo effect: how one trait changes the whole impression of a person
On his first day, a new coworker brings coffee for the whole team, smiles at everyone, and remembers all the names. A month later you calmly trust him with an important report, though you have not yet seen a single report of his. After all, he is “clearly competent.”
This is the halo effect: one noticeable quality of a person, product, or company colors how we rate its other qualities. A pleasant person also seems knowledgeable; attractive packaging seems a sign of a good product. It works in the opposite direction too: one unpleasant trait casts a shadow over everything else.
Below, we look at what research has shown, where the halo appears in ordinary life, and how to judge people and things more accurately. In the middle of the article there is a short experiment with a new coworker.
Contents
In short: how the halo effect works
What the halo effect is
The name was coined by the American psychologist Edward Thorndike in a 1920 paper. He studied how officers in the army rate their subordinates. Different things had to be rated separately: physique and bearing, intelligence, leadership, personal qualities. The instructions asked raters outright to think only about the quality being rated at that moment. But the ratings of different qualities of one person turned out to be linked too closely and too evenly. As if the officer gave one overall grade and then spread it across the columns.
The most striking experiment was run by Richard Nisbett and Timothy Wilson in 1977. Students watched a video interview with a teacher who spoke with a European accent. There were two recordings. In one, he answered warmly and in a friendly way. In the other, coldly and distantly. Then the participants were asked to rate his appearance, mannerisms, and accent. Those three things were the same in both recordings.
People who saw the warm version found his appearance, mannerisms, and accent pleasant. Those who saw the cold one found them irritating. The most interesting part came next. Participants did not notice that their overall attitude had affected their ratings. And those who saw the cold version believed it was the other way around: as if his appearance and accent had made the teacher unpleasant.
The halo of beauty has been studied separately. A meta-analysis by Alice Eagly and colleagues showed that attractive people really are rated higher on personal qualities. But the effect is moderate, and it depends strongly on which quality is meant. It is strongest for sociability and social skills. It is weaker for intelligence and emotional stability. And appearance hardly affects ratings of honesty and concern for others. The saying “what is beautiful is good” holds only in part.
Try it yourself: the new coworker
The experiment takes a couple of minutes. Here is what will happen.
Step 1. You will ask a new coworker, Martin, a simple question. Some players get a polite answer, others a cold one. The computer picks which answer you see.
Step 2. Then we will show three things: Martin’s portrait, his handwritten note, and his email to the team. They are the same for all players, however Martin answered. Rate each one.
Step 3. At the end we will show how Martin answers other players, and tell you what happened in the 1977 study.
Why rate the handwriting and the portrait? These are qualities that have nothing to do with how the person answered the question. If your rating of the handwriting depends on whether the author was polite, that is the halo. Nisbett and Wilson asked about appearance, mannerisms, and accent for the same reason: they are identical in both versions, and only the overall impression changes.
What usually happens. In the study, people who saw the warm version rated the very same appearance, mannerisms, and accent higher. And the participants themselves did not notice that the overall impression had affected their ratings.
Where the trap is. Martin’s note and email do not depend on how he talked to you: in other situations he writes exactly the same. Only the first impression changes. But it quietly colors everything else: after a polite answer the handwriting seems neater and the email more competent. Try to imagine honestly whether you would have given the handwriting the same rating if Martin had answered you differently. One player cannot check this alone. The halo effect becomes visible when you compare the ratings of those to whom Martin answered politely and those to whom he answered coldly.
Why one trait colors the rest
Picture the halo on an old icon. It lights up not one detail but the whole face at once. That is where the effect gets its name. The image conveys the main point well: light falls from the overall impression onto everything nearby.
The general attitude of “like or dislike” often forms before we have learned anything about the person in substance. When we later have to answer a specific question, such as “is his handwriting good?” or “will he manage the project?”, the ready-made impression supplies the answer. The Nisbett and Wilson experiment shows that this is invisible from the inside. It seems to us that we are rating the handwriting, but in fact we are rating the person.
Sometimes such a hint is reasonable. Some qualities really do go together. The mistake lies elsewhere: the link seems stronger than it is, and we stop checking. Thorndike drew attention to exactly this: the ratings of different qualities matched too well to be true.
Where it gets in the way
Relationships. At the start of an acquaintance, a pleasant person seems also reliable, smart, and honest, although there has been no chance yet to check any of it. It works in reverse too: one irritating habit of a new acquaintance can overshadow everything else. People are often surprised later: “I saw right away what he was like,” when they had seen one trait.
Work and business. In an interview, confident speech is easily taken for professionalism. In a performance review, one loud success colors everything else. Thorndike described this more than a hundred years ago in army ratings. A form where an employee is rated in five columns risks getting the same rating five times.
Shopping. A beautiful website seems a sign of a reliable company, a famous face in an ad a sign of a good product. Packaging design says nothing about what is inside. But the impression it makes carries over to the contents too.
News and opinions. A person who knows one field well is readily believed in another. A well-known doctor holds forth on politics, a successful entrepreneur on raising children. Expertise in one thing carries over to everything the person talks about. Here the halo easily links up with confirmation bias: once we have formed an opinion, we notice what supports it.
How to judge more accurately
The halo effect is one of the biases on the cognitive bias map. Next to it on the map are other errors in judging people, for example the fundamental attribution error.
Frequently asked questions
Is there a reverse effect?
Yes. One unpleasant trait can lower the rating of everything else. This is sometimes called the devil effect or the horn effect. In the Nisbett and Wilson experiment it showed in the cold version: the same accent seemed irritating.
Is it true that good-looking people are seen as smarter?
Partly. According to the meta-analysis by Eagly and colleagues, attractive people get the biggest “boost” in sociability and social skills. The influence on ratings of intelligence is weaker, and on ratings of honesty it is close to zero.
If I know about the halo, does it stop working?
Unlikely. The participants in the 1977 experiment were sure that their overall impression had not affected their ratings. The feeling of being objective is poor protection. It is more reliable to lean on a procedure: criteria in advance and ratings based on facts.
How can I tell whether a halo is affecting my judgment of someone?
From the inside, it is hard to tell: the participants in the 1977 experiment did not notice it. Instead, check your rating against facts. Ask “What did I actually see?”, rate one criterion at a time, and compare candidates on the same point. If a single vivid trait keeps pulling all your ratings in the same direction, that is the signal to slow down.
