Only some emotions help posts go viral

(Press-News.org) The positive emotions of love and joy might seem to have much in common. Similarly, anxiety and anger appear to be close emotional siblings.

But on social media, seemingly similar emotions can lead to very different responses.

According to new research by Yifan Yu, assistant professor of information, risk, and operations management at Texas McCombs, certain emotions pack a bigger punch than others when it comes to how posts spread on social media.

“Some emotions, like anxiety and love, help content go viral, while others, like anger or sadness, tend to slow things down,” says Yu.

His research also maps how different demographic groups react to different kinds of emotions. “People of different ages, genders, and social circles behave differently when they come across emotional content,” he says.

His findings, he believes, could help content creators and platforms employ social media more effectively.

With co-authors Shan Huang of the University of Hong Kong, Yuchen Liu of the University of Florida, and Yong Tan of the University of Washington, Yu sampled 387,486 online articles transmitted by almost 7 million unique users on WeChat, one of the world’s largest social media platforms.

They used a lexicon of words related to eight discrete emotions to classify each article’s content. They then analyzed how articles containing those words diffused through the network, looking at factors such as how widely and quickly posts spread. They found:

Anxiety, love, and surprise have positive impacts on the spread of content, while expressions of anger, sadness, and joy influence it negatively. Older people are more likely to share articles that express anger or anxiety, while younger ones are more drawn to content expressing disgust. Users with many friends tend to share articles displaying love, anxiety, anticipation, or disgust. Those with fewer friends are more likely to share articles showing anger or surprise. Similar Emotions, Different Effects

A major surprise, says Yu, is “how similar emotions can have totally opposite effects on how content spreads.”

For example, although anger and anxiety are both negative, posts expressing anxiety outspread those expressing anger. Yu notes that anger is often seen as being reactive or irrational. Anxiety requires more time and thought, and users may perceive it as more reliable and helpful to share.

“It’s not just whether an emotion is positive or negative that matters,” he says. “It’s the specific type of emotion and what it signals to the reader.

“Some emotions, like anxiety or love, may prompt people to connect, support, or warn others. Others, like anger or joy, may come off as too aggressive or self-focused, making people less likely to share.”

Yu hopes his findings can help creators be more aware of the impact of emotional language when they’re crafting content.

Platforms can use his research in developing guidelines for content moderation, he adds. “Using our models, platforms can detect content with excessive emotional expression and anticipate its likely diffusion pattern, allowing for timely interventions.

“They should pay more attention to love, surprise, and especially anxiety, as these emotions can contribute to large information cascades.”

Ultimately, he believes his research “can help make social media safer by helping platforms better detect potentially harmful content early on, and by supporting policies that are more nuanced, data-driven, and effective.”

“Emotions in Online Content Diffusion” is forthcoming in Information Systems Research.

END