New Research Suggests Gossip “Not Always a Bad Thing”

news story image

Dana Nau, Professor (retired), UMD Department of Computer Science and Institute for Systems Research.

A version of this story appeared originally at

We profess to hate it, even as we indulge in it. Gossip, defined as the exchange of personal information about absent third parties, is a near-universal behavior among humans, transcending gender, age, culture and socioeconomic background.

And though it gets a bad rap, a new theoretical study conducted by University of Maryland (UMD) and Stanford University researchers suggests that gossip might play a valuable role in the formation of social groups.

The much-maligned behavior, it turns out, helps others determine whether a person tends to be cooperative or self-seeking. That information can influence decisions about whether to associate with the object of the gossip.

“When people are interested in knowing if someone is a good person to interact with, if they can get information from gossiping—assuming the information is honest—that can be a very useful thing to have,” said study co-author Dana Nau, a retired professor in UMD’s Department of Computer Science and Institute for Systems Research.

Unraveling the Behavioral Puzzle

In their study, published on February 20 in Proceedings of the National Academy of Scientists, Nau and his collaborators used a computer simulation to help determine how, and why, gossip might be advantageous to the human species.

“One previous study shows that, on average, a person spends an hour per day talking about others, so this takes a lot of time out of our daily life,” said the study’s first author, Xinyue Pan (M.S. ’21, Ph.D. ’23, psychology), who published part of this research in her master’s thesis. “That’s why it’s important to study it.”

Previous theories have already pointed to potential benefits to large groups, who obtain information through this means about potential members, allies, and adversaries. How individual gossipers benefit from the behavior, however, has until now remained something of a mystery. “It’s unclear why gossiping, which requires considerable time and energy, evolved as an adaptive strategy at all,” said study co-author Michele Gelfand, a professor at Stanford Business School and a professor emeritus in UMD’s Department of Psychology.

Why gossip recipients are so willing to lend a sympathetic ear to gossipers or behave differently in their presence also remained unexplained.

To better understand the complex webs of gossip, the research team used an evolutionary game theory model that mimics human decision-making. They then observed how their agents, or virtual study subjects, interacted with each other and altered their strategies to receive rewards.

The researchers were particularly interested in learning whether agents would use gossip to protect themselves or to exploit others. Agents could cooperate with gossipers or defect; they could become gossipers themselves; and they could change their strategies after observing the consequences or rewards of other agents’ decisions. By the end of the simulation, 90% of the agents had become gossipers.

“When people are interested in knowing if someone is a good person to interact with, if they can get information from gossiping—assuming the information is honest—that can be a very useful thing to have.”

Dana Nau, Professor (retired), UMD Department of Computer Science and Institute for Systems Research.

The researchers found that people are more likely to cooperate in the presence of a known gossiper because they want to defend their own reputation and avoid falling victim to the rumor mill. For gossipers, receiving another person’s cooperation can be a reward in itself. “If other people are going to be on their best behavior because they know that you gossip, then they’re more likely to cooperate with you on things,” Nau explained. “The fact that you gossip ends up providing a benefit to you as a gossiper. That then inspires others to gossip because they can see that it provides a reward.”

A Deterrent to Selfishness?

The findings suggest that gossip proliferates because sharing information about people’s reputations can have a “selfishness deterrence” effect on gossip recipients. In other words, gossip recipients condition their behavior on others’ reputations, and because they don’t want to be the subject of future gossip, this deters them from acting selfishly. Thanks to their ability to influence the behavior of others and encourage cooperation, gossipers have an “evolutionary advantage” that perpetuates the cycle of gossip and provides a useful service to listeners.

Furthermore, the evolutionary benefits apply both to negative and positive gossip, the findings suggest. Whether we speak ill or well of another person who’s not in the room, we’re exchanging valuable tips about reputation. “Once people have this information, cooperative people can find other good people to cooperate with, and this is actually beneficial for the group. So gossiping is not always a bad thing. It can be a positive thing,” said Pan, who is now an assistant professor at the Chinese University of Hong Kong, Shenzhen.

The simulation also considered different factors that help or hinder the spread of gossip, ultimately confirming what past research has shown: small-town gossipers aren’t just a movie trope.

“The model highlights contexts in which we can expect more gossipers to evolve, particularly when social networks have high connectivity and mobility is low, consistent with research on rural areas,” Gelfand said. “It gives clues to contexts where gossiping may be more or less likely to flourish.”

Nau explained that their research does not encompass the full scope of human complexity, nor can it replace behavioral studies. However, computer simulations can yield new theories that inspire follow-up research involving human subjects.

“People are very complicated and we can’t come up with a simulation that does everything that people do, nor would we want to,” Nau said. “Since it’s an oversimplification, you can’t say conclusively that this is how people behave, but you can develop insights. Those can then lead to scientific hypotheses that you can try to investigate through studies that involve human

Avenues for Future Research

The researchers hope to pursue a follow-up study to test one of their simulation’s predictions on human participants: the idea that gossip is effective when people don’t have other methods of gathering information about people’s reputations.

“That, for me, is one of the really exciting parts of this,” Nau said. “If we can come up with hypotheses and verify the predictions of those models on human studies, then that’s what makes this kind of thing useful.”

There is one thing the researchers can already say with confidence. Given the overwhelming number of gossipers in their simulation and in real life, gossiping is unlikely to go away anytime soon.

In addition to Nau, Pan and Gelfand, this study was co-authored by Vincent Hsiao (B.S. ’18, computer science; B.S. ’18, mathematics; Ph.D. ’24, computer science), a research scientist at the U.S. Naval Research Laboratory. The paper, “Explaining the evolution of gossip,” was published in Proceedings of the National Academy of Sciences on February 20, 2024.

Published February 21, 2024