Researchers who study social media have a consistent finding about what spreads: shocking and novel material travels faster than careful material. A 2018 study in Science by Soroush Vosoughi, Sinan Aral and Deb Roy of MIT tracked about 126,000 news cascades tweeted more than 4.5 million times by roughly 3 million people between 2006 and 2017. False news stories were 70 percent more likely to be retweeted than true ones, and true stories took about six times as long to reach 1,500 people, MIT News reported. Human users, not bots, drove the difference. Aral said false news is more novel and people are more likely to share novel information.
Emotion played a part. People responded to false stories with surprise and disgust, the MIT team found, while true stories drew sadness, anticipation and trust. A 2017 study in the Proceedings of the National Academy of Sciences by William Brady, Julian Wills, John Jost, Joshua Tucker and Jay Van Bavel examined 563,312 tweets on gun control, same-sex marriage and climate change. Each additional moral-emotional word, such as "hate," increased a message's retweets by about 20 percent, the authors found. The effect was larger within liberal and conservative networks than between them.
Brain imaging offers a view of why a like can feel like a prize. In a UCLA study published in Psychological Science in 2016, 32 teenagers viewed photos in a brain scanner. Their own photos with many likes produced strong activity in the nucleus accumbens, a part of the brain's reward circuitry, ScienceDaily reported. A later study of 58 adolescents and young adults found that giving likes activated the striatum and the ventral tegmental area, according to the paper in Social Cognitive and Affective Neuroscience. A 2021 study in Nature Communications by Bjorn Lindstrom and colleagues analyzed more than one million posts from over 4,000 people on several platforms. People adjusted how often they posted to the rate of likes they received, a pattern the authors said conforms to the principles of reward learning. An experiment with 176 participants found that social rewards causally influenced posting behavior, according to the study's abstract.
Those studies measured brain activity and behavior. None of them measured dopamine, so "dopamine rush" remains a popular shorthand. A 2023 systematic review in Brain Sciences by Michael Wadsley and Niklas Ihssen of 28 MRI studies on social networking site use found the ventral striatum, which includes the nucleus accumbens, to be a key region. The review's authors called their conclusions tentative. They wrote that no longitudinal evidence shows social media causes lasting changes in the brain, and that calling problematic use a disease like substance addiction is premature.
One study in that review bears directly on what spreads. Twenty-eight Weibo users in a Chinese experiment chose whether to repost positive, negative or neutral messages. Overall, they preferred to repost the negative ones.