Knowledge of information cascades through social networks facilitates strategic gossip

Social networks are composed of many ties among many individuals. These ties enable the spread of information through a network, including gossip, which comprises the lion’s share of daily conversation. Given the number of possible connections between people in even the smallest networks, a formidable challenge is how to strategically gossip—to disseminate information as widely as possible without the target of the gossip finding out. Using a novel gossip-sharing task in artificial social networks (Experiments 1-3; N=568), we find that people achieve this goal by leveraging knowledge about topological properties, specifically, social distance and popularity. We find a similar pattern of behavior in a real-world social network (Experiment 4; N=187), revealing the power of these topological properties in predicting information flow, even in much noisier, complex environments. Computational modeling suggests that these adaptive social behaviors rely on mental representations of information cascades through the social network.