Credit: fizkes on Shutterstock
In A Nutshell
- Fake ratings from strangers reliably shifted how much pain, empathy, and mental effort participants reported feeling in a Dartmouth study of 111 adults.
- The brain learns more readily from experiences that confirm existing expectations and discounts evidence that contradicts them, a pattern researchers call confirmation bias in learning.
- These socially shaped expectations persisted across three days and up to 72 trials, fading only modestly even after repeated experiences that should have exposed the cues as meaningless.
- Researchers suggest the same mechanism could help explain why online horror stories about medical procedures or difficult tasks may make those experiences feel worse for people who read them first.
Before a medical procedure, many people can’t help but go online for information. They scan comment threads, read patient reviews, and scroll through forum posts describing what the experience is like. New research from Dartmouth College suggests that habit may be doing more than managing anxiety. Published in the Proceedings of the National Academy of Sciences, the study suggests exposure to other people’s negative opinions can amplify how painful or draining an experience feels, and that effect can be surprisingly hard to shake.
Just as a wave of glowing restaurant reviews can make food taste better, a flood of warnings about a painful experience may make it feel genuinely worse. And once those socially shaped expectations take root, the brain actively works to preserve them, even when direct sensory experience tells a different story.
According to the study, “others’ opinions can strongly bias experiences of pain, empathy, and cognitive effort, creating beliefs resistant to corrective evidence.”
How Fake Crowd Ratings Changed Real Pain
Researchers recruited 111 healthy adults for experiments spread across multiple sessions. Participants took part in three tasks. In the first, heat was applied to the inner forearm at temperatures between 48 and 50 degrees Celsius. In the second, they watched videos of people in varying degrees of visible pain, a measure of vicarious pain, the discomfort felt when witnessing someone else suffer. In the third, they mentally rotated objects at different angles, a task used to gauge perceived cognitive effort.
Before each trial, participants viewed what appeared to be ratings from 10 previous participants who had faced the same stimulus, shown as a cluster of dots on a scale. Half the cues pointed high, suggesting the experience had been very unpleasant. Half pointed low.
The ratings were randomly generated and had no connection to the actual stimulus.
Participants reported more pain when fake ratings pointed high and less when they pointed low. Identical patterns held for empathy and perceived mental effort, and effects were large and consistent across all three tasks and three testing days.
Why the Brain Locks In What It Expects
Social cues did not just color experience in the moment; they reshaped how participants learned from experience going forward.
When someone received a high-aversive cue and then felt more pain than expected, they readily updated their future expectations upward. When they received the same cue but felt less pain than expected, the update barely registered. Bad news reinforced itself. Good news was mostly ignored.
Researchers called this a confirmation bias in learning, one that appears to influence perception and learning, not just conscious opinion. Computational modeling showed that participants’ behavior required two separate learning rates: a higher one for experiences that matched expectations and a much lower one for experiences that contradicted them.
People who showed this pattern strongly in the pain task tended to show it in the cognitive effort and vicarious pain tasks as well, suggesting that sensitivity to crowd opinion may reflect a relatively stable individual tendency that crosses different types of experience.
Even across 72 trials over three days, cue effects faded only modestly and never disappeared. A person primed to expect pain by false social ratings kept feeling more of it, even after repeated experiences that should have revealed those ratings as meaningless.
Social Media Opinions and the Pain You Feel
Most prior research on social influence and pain used cues that had been repeatedly paired with real pain. Those cues earned their power. What this study tested is something more relevant to daily life: suggestions with no real-world basis at all, the kind manufactured and spread in seconds across social networks and online platforms.
Health forums, comment sections, and review sites are full of exactly that content. A post describing a vaccination as excruciating, a thread warning that a procedure is brutal, or a wave of reviews calling a workout unbearable can reach thousands of people before any of them have had the experience. If socially conveyed expectations behave like the fabricated cues in this study, that content may not just prime people for discomfort but could make experiences feel more unpleasant than they otherwise would.
Researchers also speculated that this mechanism, social priming followed by confirmation bias, could help explain some features of chronic pain and chronic fatigue syndrome. A brain continually primed to expect suffering, and resistant to evidence otherwise, could sustain symptoms that seem impervious to reassurance. The authors were clear that this idea needs clinical testing before drawing firm conclusions.
Scroll through enough horror stories about a medical procedure, and the body may well brace itself accordingly. According to this research, that bracing may linger long after.
Disclaimer: This article is based on a published peer-reviewed study. The findings reflect results from a controlled laboratory setting involving healthy young adults and may not apply to all populations. The speculative links to chronic pain and chronic fatigue syndrome have not been clinically tested and should not be interpreted as medical advice.
Paper Notes
Limitations
Study participants were 111 healthy adults recruited at Dartmouth College, with an average age of approximately 24.7 years, reporting no psychiatric or neurological diagnoses in the prior six months, no history of chronic pain, and no active pain at the time of the study. Results may not transfer to older populations, clinical groups, or people with chronic conditions. Some individual behavioral parameters, particularly condition-specific learning rates, showed modest cross-task stability, and measurement noise may have limited the detection of certain correlations. Social desirability, the tendency to report what seems socially acceptable, could have influenced some ratings, though the authors argue it would not substantially alter the core findings on cue and learning effects as long as individual response tendencies remained internally consistent.
Funding and Disclosures
This work was supported by National Institutes of Health grants MH076136 and R01EB026549, both awarded to Tor D. Wager. The authors declared no competing interests.
Publication Details
Authors: Aryan Yazdanpanah, Heejung Jung, Alireza Soltani, and Tor D. Wager, all affiliated with the Department of Psychological and Brain Sciences at Dartmouth College, Hanover, NH. Edited by Elke Weber of Princeton University. Distributed as an open-access article under Creative Commons Attribution License 4.0. Title: “Social information creates self-fulfilling prophecies in judgments of pain, vicarious pain, and cognitive effort.” Published February 9, 2026, in the Proceedings of the National Academy of Sciences (PNAS), Vol. 123, No. 7, article e2513856123. DOI: https://doi.org/10.1073/pnas.2513856123.
