The science suggests they might be right.
New psychological research indicates that seeing a sad, angry or fearful face doesn’t just colour how a situation feels — it actually makes it harder for our brains to work out what caused what in the first place.
Emotional faces quietly warp our sense of cause and effect
The study, published in the Quarterly Journal of Experimental Psychology, looked at a basic skill that underpins everyday social life: learning what usually follows what. Psychologists call this “contingency learning” — the ability to estimate how likely an outcome is when we see a particular cue.
In social settings, this might mean judging whether a smile from a colleague usually leads to a friendly response, or whether a certain tone of voice tends to lead to conflict. We rely on this constant statistical guesswork to anticipate reactions and avoid awkward or risky encounters.
When faces showed negative emotions, people became worse at spotting genuine cause-and-effect links — even when the statistics were clear.
The research team, led by Rahmi Saylik of Mus Alparslan University alongside scientists from Brunel University London and the University of Oxford, set out to test a simple question: does the emotional “valence” of a face — positive or negative — change how well we learn about cause and effect between people?
Inside the experiment: flashing faces and hidden patterns
Across four experiments, more than 300 participants completed rapid, computer-based tasks. They watched streams of images flash up: sometimes faces, sometimes geometric shapes or patterned designs. One image played the role of a “sender”, the next a “receiver”. The participants’ job was to rate how strongly they thought the first caused the second.
The twist: the researchers secretly controlled the true statistical relationship between cue and outcome. In some blocks, the first image almost always predicted the second. In others, the pairing was largely random.
Sad faces blur the statistical picture
In the first experiment, 107 people viewed combinations of happy faces, sad faces, or simple shapes. After each block, they rated how strong the causal link felt, using a scale from negative to positive.
- Participants could usually tell strong relationships from weak ones.
- But when the images were sad faces, their ratings drifted away from the actual statistics.
- They consistently judged the causal link as weaker than it really was.
Happy faces and shapes produced much more accurate judgments. That pattern suggested something specific about sadness was getting in the way of clear causal thinking.
➡️ Why light discomfort often appears after mentally demanding tasks
➡️ Banana peels in the garden: they only boost plants if you put them in this exact spot
➡️ No vinegar and no baking soda needed: pour half a glass and the drain cleans itself effortlessly
➡️ This habit helps avoid misunderstandings before they start
➡️ The French Rafale “crushes” the American F‑35 in a little‑discussed area: incident rates
Negative emotion did not just fail to help learning — it actively pulled people’s judgments away from the facts.
Not just a quirk of pictures or simple counting
One possible objection was that faces and shapes simply look too different. Maybe people reacted differently because the images were harder to compare visually, not emotionally.
So in a second experiment with 82 new volunteers, the team controlled for that. All images were presented in black and white, faces appeared through oval windows, and shapes were designed to mimic the presence or absence of a visual “feature”, just as a face is emotional or neutral.
The result barely shifted: sad-face trials still led to weaker perceived causality than happy-face or shape trials. Interestingly, there was no real difference between happy faces and neutral shapes, suggesting positive emotion wasn’t boosting performance — instead, negative emotion was dragging it down.
Another concern was that people might simply be counting how often two images appeared together rather than computing a true probability. This is known as the “pairing hypothesis”.
To test that, the third experiment with 90 participants engineered tricky conditions: sometimes the cue and outcome appeared together often but were poor predictors; in other conditions they appeared together rarely but were strong predictors.
Participants consistently rated statistically stronger links as stronger, even when the number of pairings was the same — showing they were tracking probability, not just frequency.
Yet again, under negative emotional conditions, their sense of causality dipped.
Anger and fear distort judgments too
The final experiment broadened the question: is this effect unique to sadness, or does it extend to other negative states?
Fifty-one participants completed the same style of task, but this time the faces expressed happiness, anger or fear. The pattern held. When the images were angry or fearful, people rated genuine causal links as weaker than when the faces were happy.
| Facial expression | Perceived causal strength |
|---|---|
| Happy | Closer to true statistical relationship |
| Sad | Systematically underestimated |
| Angry | Underestimated in positive relationships |
| Fearful | Underestimated in positive relationships |
Across emotions with negative valence — sadness, anger, fear — one theme emerged: they interfered with the brain’s ability to correctly gauge cause and effect in social cues.
Why threatening faces hijack our mental bandwidth
Psychology has long treated threatening faces as “special”. They grab attention quickly, trigger fast emotional responses and may have offered evolutionary advantages by flagging danger.
This new work suggests that attention-grabbing is only half the story. Negative faces may pull resources into internal reactions — worry, vigilance, physiological arousal — leaving less capacity for slower, statistical reasoning.
A sad or furious glare might be noticed instantly, but that same glare can make it harder to think clearly about what led to it.
In other words, the brain may be caught between “What does this mean for me right now?” and “What usually follows this expression across many situations?”. The urgent, emotional question tends to win, at the cost of analytic thinking about patterns.
What this means for arguments, conflict and everyday misunderstandings
The findings offer a fresh angle on why conflicts escalate and why people often misjudge intentions when tensions run high.
Imagine a couple arguing. One partner looks hurt and angry. The other tries to reason through what led to this moment. If negative expressions interfere with causal learning, the “reasoning” partner may actually be struggling to track the real chain of events. Their brain is simultaneously dealing with the emotional impact of the other person’s face.
That mix could lead to distorted interpretations, such as:
- Overlooking earlier, calmer signals that contributed to the conflict.
- Underestimating how strongly one person’s actions predict the other’s reactions.
- Assuming everything is random or “out of the blue” when patterns actually exist.
In workplaces, similar dynamics can apply. A manager who often looks stressed or irritated may make it harder for staff to read what really causes praise, criticism or promotion. People may fail to notice that certain behaviours reliably lead to positive outcomes, simply because the emotional backdrop is tense.
Key terms behind the research
Two ideas from the study surface regularly in behavioural science:
- Valence: The emotional “direction” of a stimulus — roughly, how positive or negative it feels. A smile has positive valence, a scowl negative.
- Contingency: The strength of the statistical link between a cue and an outcome. If a friend almost always laughs at your jokes, the contingency between your joke-telling and their laughter is high.
Participants were not told any of this technical language, of course. Yet their ratings showed that, under calm or positive imagery, people are fairly good at acting like intuitive statisticians. Under negative emotion, that ability falters.
Limits of the evidence and real-life scenarios
The experiments used static, computer-presented photos. Real conversations involve moving faces, tone of voice, body language and history between people. That complexity might make the emotional interference stronger or weaker in daily life.
The participants were mostly neurotypical university students. People with anxiety, depression or trauma histories might react very differently to negative expressions. For some, a fearful or angry face may trigger much stronger internal reactions, further shrinking the mental space left for tracking cause and effect.
Consider a crowded commute. Someone frowns at you after you bump into them. If your attention locks onto their expression, you may feel shame or defensiveness and stop noticing the pattern that actually matters: how often small apologies quickly defuse such moments. Over time, these tiny misreadings can build into a sense that social situations are unpredictable or hostile, even when they usually follow fairly stable rules.
Future work combining these tasks with physiological measures — heart rate, skin conductance, eye tracking — could map exactly how emotional arousal competes with statistical reasoning. That sort of data may help design better communication strategies for high-stress settings, from classrooms to couples therapy and tense team meetings, where faces are rarely neutral and cause-and-effect judgments really do matter.
Originally posted 2026-03-05 02:00:44.