The Halo Effect: An Evidence-Based Guide

What it changes, from school to work

The advantage starts in childhood. A large review of the research found that attractive children, like attractive adults, are judged more positively and treated better than unattractive ones, even by people who know them well (Langlois et al., 2000)⁽⁶⁾.

School is part of it. Teachers tend to judge attractive students as more intelligent and more academically able (Ritts et al., 1992)⁽⁷⁾. Attractive teenagers also tend to be better integrated socially and face less stigma, although Gordon and colleagues (2013), who followed them into young adulthood, found those benefits partly offset by social distractions⁽⁸⁾.

At work, it shows up in pay. In US and Canadian surveys where an interviewer rated each person’s looks, Hamermesh and Biddle (1994) found that men rated below average earned about 9% less per hour than average-looking men, and men rated above average earned about 5% more, with education, experience and other job factors taken into account. For women, both gaps were smaller⁽⁹⁾.

But the penalty is bigger than the prize

In that last chart, the drop for unattractive faces is about two and a half times the boost for attractive ones.

Other studies point the same way. When Griffin and Langlois (2006) compared faces of low, medium and high attractiveness side by side, they concluded that “most often, unattractiveness is a disadvantage,” which runs against the beauty-is-good idea⁽²⁾. The pay data leans the same way, if only a little: in Hamermesh and Biddle’s (1994) words, the penalty for plain looks is “slightly larger than the premium for beauty”⁽⁹⁾.

This side of the bias has its own name, the horn effect, sometimes called the devil effect. One negative feature, such as an unkempt look or a visible flaw, drags down the judgment of the whole person.

Good looks can even count against you. Evaluators discriminate against attractive candidates for less desirable jobs, because they assume those candidates feel entitled to something better (Lee et al., 2018)⁽¹²⁾. In the Italian study, an attractive photo helped with technical and sales jobs but lowered callbacks for professional, skilled and elementary roles (Busetta et al., 2021)⁽¹⁰⁾.

Why it happens, and why it partly comes true

The halo works because first impressions are fast. People judge a face within a tenth of a second, and those snap judgments agree closely with the ones they make when they have as long as they like (Willis & Todorov, 2006)⁽¹³⁾. A meta-analysis of brain imaging studies found consistent activity in the amygdala, a region involved in emotional evaluation, both when people judge how attractive a face is and when they judge how trustworthy it looks (Bzdok et al., 2011)⁽¹⁴⁾.

Those judgments shape how people are treated, and over the years, treatment shapes people. On average, attractive adults are a little more outgoing, a little more self-confident and slightly more socially skilled than others (Langlois et al., 2000)⁽⁶⁾. One explanation is that being treated as likable gives you more practice at being liked. Langlois and her colleagues weigh that against an evolutionary explanation and treat both as open.

This is also where the halo effect differs from the Self-Esteem Multiplier, the idea our tenet articles use for the way looking better can raise your self-esteem, with knock-on effects for confidence and wellbeing. The halo runs on how other people see you. The Self-Esteem Multiplier runs on how you see yourself.

But the stereotype runs well ahead of the reality. Attractive adults are barely more intelligent than anyone else, an effect close to zero (Langlois et al., 2000; Feingold, 1992)^(6)(15). Grades are a mixed picture: in a high school study by French and colleagues (2009), looks helped girls’ grades, but once personality and grooming were taken into account, the effect of looks turned negative⁽¹⁶⁾. The idea that a beautiful face advertises good genes (Scheib et al., 1999)⁽¹⁷⁾ or good health has not held up well either. Attractiveness in adolescence did not predict health, then or later (Kalick et al., 1998)⁽¹⁸⁾, and a major review found little evidence that the facial features thought to signal health actually do (Weeden & Sabini, 2005)⁽¹⁹⁾.

AI systems learned it anyway

Wrong as much of the stereotype is, it has now spread beyond people. AI systems help screen job applications and sort what people see online, and research shows they have picked up the same bias.

Kim and colleagues (2025) gave AI models the same job application with a different photo attached each time. The face was the same good-looking person in every photo, and only the setting changed, from a plain portrait to photos at work, relaxing or doing a hobby. Several multimodal models, the kind that read images as well as text, scored the same candidate’s competence differently depending on the photo, although two of the models showed no halo effect. When the extra information came as text, the models were less swayed⁽²⁰⁾.

A second study, by Gulati and colleagues (2025), tested seven open-source multimodal models on 91 everyday decisions, using photos of the same people before and after a beauty filter. Attractiveness changed the models’ decisions in 86.2% of the scenarios on average. Asked to choose between a positive and a negative trait, such as trustworthy or untrustworthy, the models linked the filtered faces with the positive trait in 92.6% of scenarios, and the bias weighed most heavily on women⁽²¹⁾. The authors point to training: models that learn from images and descriptions of people may come to treat attractiveness as relevant.

A person’s bias reaches the people they meet, but a screening model applies the same bias to every application it reads.

Beauty filters show how quickly the halo follows a change in looks. When Gulati and colleagues (2024) had photos of 462 people rated before and after a beauty filter, none was rated less attractive after it, and the filtered versions were also rated more intelligent, trustworthy, sociable and happy⁽²²⁾. The gain was largest for faces that started out rated lower.

What it means if you are the one being judged

If people have ever treated you differently after a new haircut or losing weight, this is the mechanism at work. The halo runs on what other people see, and that is not quite the face you know. A mirror shows you a flipped version, and cameras distort faces in their own ways, which is why your own view is a poor guide to theirs. A QOVES facial analysis is built for that gap: it measures your face from your photos and compares it with faces from your own background, so you can see the features other people respond to.

Online, the same idea goes by “pretty privilege”: the everyday advantages attractive people get without earning them. The phrase comes from social media, but the effect it names is the one in this article, and its details change with culture. In a study by Wheeler and Kim (1997), Korean students linked attractive faces with integrity and concern for others, which North American students do not, and did not see them as more dominant or assertive, which North American students do⁽²³⁾.