Scientists Built a Way to Find Who Really Spreads News Online

Researchers Xiang Liu, Yan Jia, Rong Jiang and Yong Quan published a paper (on arXiv, a public site where academics post research before or after formal publication) describing three new ways to calculate how influential a person is on social media. Their key insight is simple: follower counts lie. What actually signals influence, they argue, is behavior — specifically, whether people forward your posts and whether they tag you with an @ when they write something.

Think of it like a dinner party. Someone can stand in the middle of the room and talk all night, but that doesn't mean anyone repeats what they said the next day. The person who really matters is the one whose story shows up at three other dinner tables the following week. The researchers turned that idea into three mathematical formulas, then checked their results against a real dataset to see whether the numbers matched what actually happened. They report that the methods performed well.

Why should you care? This kind of research sits behind features you already use. Platforms use influence scores to decide whose posts appear in your feed, which accounts get promoted, and which content gets flagged as spreading rumors. If those scores are based on the wrong signals — raw follower counts, say — the wrong people get amplified. Better scores could mean a feed that reflects genuinely useful voices rather than whoever bought the most followers.

The catch is real, though. This is a 2016 conference paper just now posted publicly, not a new product you can use. It was tested in one setting, and its methods rely on the same tracking that lets platforms watch what you forward and who you tag. The same tool that surfaces trustworthy voices could also help governments or marketers map who moves opinion — and target them. Read the paper for the math; read it twice for what the math could be used for.