Social anxiety might not warp how we perceive eye contact

A recent study challenges the idea that people with social anxiety consistently perceive others as staring at them more than those without the condition. Published in Journal of Affective Disorders, the research reviewed past claims and conducted new experiments, finding that previous links between social anxiety and altered gaze perception are weak and difficult to replicate. The findings suggest researchers should reconsider using this perceptual trait as a standard biological marker for social anxiety.

In everyday social situations, humans must constantly judge where others are directing their attention. To measure how people process this information, psychologists evaluate what they call the cone of gaze. This concept refers to the specific range of angles at which a person feels that another individual is looking directly at them.

Imagine sitting on a train across from a stranger. Even if the stranger is looking slightly over your shoulder, you might still feel they are making eye contact. The width of this imaginary cone dictates how far a person’s eyes can deviate from the center before you decide they are looking away.

For years, some psychologists have proposed that the cone of gaze could serve as a biological and behavioral marker for social anxiety disorder. People experiencing high levels of social anxiety often report feeling heavily observed or judged by others. Based on this clinical observation, earlier research suggested that socially anxious individuals possess a wider cone of gaze, meaning they interpret a broader range of angles as self-directed eye contact.

To explain why scientists think the cone of gaze might change, it helps to look at how the brain processes visual information. Psychologists generally divide this into two streams. Bottom-up processing involves the raw visual data the eyes take in, such as the exact position of another person’s pupils. Top-down processing involves a person’s internal expectations and beliefs.

Free daily newsletter

When visual data is ambiguous, the brain relies more heavily on its top-down expectations. Because humans have a strong default assumption that others are looking at them, poor visual evidence usually causes people to assume they are making eye contact. Psychologists suspected that people with social anxiety might naturally avoid looking directly at other people’s eyes. By looking away, they gather less bottom-up visual evidence, forcing their brains to rely on top-down expectations and hypothetically widening their cone of gaze.

Other studies proposed that angry facial expressions also widen this cone for the general population. This idea is based on the evolutionary benefit of detecting a potential threat. If someone looks angry, it pays to assume their anger is directed at you.

To test the reliability of these concepts, a team led by Madeleine E. Moses-Payne at University College London evaluated the existing scientific literature. The research team, which included Georgina Krebs, Raphaëlle Delpech, Antonina H. Przyborowska, and Argyris Stringaris, wanted to determine if the differences in gaze perception were consistent enough to act as a diagnostic tool. They designed a two-part investigation featuring a synthesis of past research followed by fresh experimental trials.

The investigators first conducted a meta-analysis, which is a statistical technique that combines data from multiple independent studies to identify overall trends. They pooled data from ten previous studies that compared the cone of gaze in socially anxious individuals against those without social anxiety. This combined group represented a total of 403 participants.

The pooled results showed a modest association, suggesting that socially anxious people did have a wider cone of gaze. However, the researchers noted that the existing evidence was highly inconsistent. The individual studies were often quite small, and mathematical projections indicated that future studies could just as easily produce contradictory results.

A second meta-analysis examined how emotional expressions alter gaze perception. The team combined nine past studies representing 523 participants. These studies compared how people judge the gaze of angry faces versus neutral faces.

The analysis revealed a slight trend where participants perceived angry faces as looking at them over a wider range of angles than neutral faces. Much like the first analysis, the data showed extreme variation between the studies. The high degree of inconsistency led the researchers to question the reliability of the previously reported effects.

Because the historical data proved unreliable, Moses-Payne and her colleagues recruited 514 participants for three new online experiments. Participants viewed images of faces presented on their screens for a fraction of a second. The faces featured eyes angled at various degrees, ranging from looking straight ahead to looking eight degrees to the side.

For each image, the participant had to press a key indicating whether the face was making eye contact. All participants also completed a standardized questionnaire measuring their current symptoms of social anxiety.

In the first experiment, 193 participants evaluated faces showing strictly neutral expressions. The researchers then mapped each participant’s responses against their social anxiety scores. They wanted to see if higher anxiety correlated with a wider cone of gaze.

The results were not statistically significant. Participants with high levels of social anxiety performed almost exactly like those with low anxiety. Both groups judged faces to be making eye contact at roughly the same angles.

The second experiment tasked 172 new participants with the same exercise, but all the faces displayed angry expressions. The goal was to see if the presence of a social threat would trigger a widened cone of gaze. The researchers were particularly interested to see if this happened for those with elevated anxiety.

In this specific setup, the researchers did observe a slight effect. As social anxiety scores increased, participants demonstrated a fractionally wider cone of gaze. For every one-point increase on the anxiety scale, the cone widened by just 0.01 degrees.

The third experiment presented 149 participants with a mix of both angry and neutral faces. This design allowed the researchers to compare the effect of emotion within the same testing session.

Once again, the results were not statistically significant. The presence of social anxiety did not widen the cone of gaze. Additionally, the angry expressions did not alter participants’ perceptions compared to neutral expressions. Even when the researchers isolated just the angry faces in this third trial, the link to social anxiety vanished.

When the research team combined the data from all three new experiments, they found no overall relationship between social anxiety and the cone of gaze. The collective data also showed that angry expressions did not widen the range of perceived eye contact. The single positive finding from the second experiment failed to hold up when viewed as part of the broader data set.

The researchers highlighted several limitations in their work. The participants in the new experiments were evaluated using a validated symptom questionnaire, but they were not formally diagnosed with social anxiety disorder by a clinician. A sample of patients with formal clinical diagnoses might yield different patterns of gaze perception.

The experimental trials were conducted entirely online. In-person testing environments sometimes introduce what psychologists call a social evaluative threat, where the physical presence of a researcher induces self-focused anxiety. The lack of an observing researcher in the online setup might have reduced the participants’ baseline anxiety, potentially dampening any effects on their visual perception.

Future studies could explore whether explicitly introducing social pressure alters how people perceive eye contact. Researchers might achieve this by asking participants to prepare a public speech before the visual task.

Scientists could also use eye-tracking technology to determine if socially anxious people avoid looking at eyes entirely. This avoidance would limit the visual information they absorb. That visual deficit would force them to rely on their baseline assumptions about being watched, which could help explain the variations seen in previous research.

The study, “Is everyone looking at me? The perception of direct gaze as a potential cognitive marker in social anxiety and its moderation by emotional expression,” was authored by Madeleine E. Moses-Payne, Georgina Krebs, Raphaëlle Delpech, Antonina H. Przyborowska, and Argyris Stringaris.