In the Hollywood movie version of revenge, our wronged hero justifiably vanquishes the villain. In real life, though, revenge is hardly ever so clear-cut. Aggrieved persons typically do not know, or cannot access, the specific individual who did them wrong. Instead a phenomenon occurs that psychologists call “displaced revenge,” where avengers target a proxy—someone akin to the original transgressor. A new study finds that displaced revenge is sweeter when the target seems to belong to the same group as the wrongdoer.

The study explored entitativity, which is a measure of how closely people are associated with one another. A random crowd at a bus stop is loosely entitative. Sports team members—allied for a common cause and wearing the same jersey—are highly entitative.

The study's authors ran three experiments that compared displaced revenge against low- and high-entitativity third parties. The first involved hypothetical scenarios; the second had subjects recall a time they had felt wronged and then speculate about how they would feel if they had a chance get revenge on various third parties. In the third experiment, real-life victims could choose to exact revenge on innocent, real third parties. Students were manipulated into believing that their partners in a puzzle-solving test had decided not to share a prize of raffle tickets for a restaurant gift card. Before taking the test, the students had watched a video in which their partner—later their nemesis—either conversed with or ignored two other students who were dressed similarly or dissimilarly to the malfeasant partner. The wronged students could choose to do nothing or pursue vengeance by forcing these other students to view unpleasant images.

Across all experiments, avengers reported higher feelings of justice-related satisfaction against more closely tied people. The study illustrates how displaced revenge can fuel, for example, ethnic gang wars, says Arne Sjöström, co-author of the study and a psychologist at the Philipp University of Marburg in Germany. The results also suggest how cycles of retribution might be broken. “One potential strategy,” Sjöström says, “may consist in promoting perceptions of group variability,” so that the target group looks less monolithic.