Few researchers can do science single-handedly, making collaboration crucial. According to a new analysis, long-term collaborations pay especially big dividends, yielding a 17% boost in citation rate for resulting papers. The findings, published in Proceedings of the National Academy of Sciences (PNAS), were well received by observers on social media. Wouter Gerritsma, an information specialist at Vrije University in Amsterdam, tweeted:

With those rewards in mind, Jeremy Borniger, a neuroscience PhD student at Ohio State University in Columbus, used Twitter to ask:

Somewhat ironically, the PNAS paper had a single author, Alexander Petersen, who is a mathematician and economist at the IMT Institute for Advanced Studies Lucca in Italy. Examining the publication records of 193 biologists and 280 physicists, Petersen found 94,000 publications involving 166,000 different collaborations. Some scientists had serial relationships that didn’t last. But Petersen found examples of “life partners” — scientists who collaborated multiple times over an extended period. In many cases, researchers in such partnerships shared more than half of their publications with the same collaborators.

Petersen argues that collaborations are increasingly important because few scientists have the complete skill set — or the time — needed to tackle projects of significance. “It’s becoming harder for a single person to climb the mountain alone,” he says. While many have long suspected that collaboration is important for success, he says, “it’s important to assign some numbers to these things.”

Using a metric based on publications between collaborators over time, Petersen identified a group of especially strong partnerships that he calls ‘super ties’. These produced an unusually high number of papers in a given period. Papers including authors with such ties receive an average of 21 more citations in biology and 8 more citations in physics than those without a super tie.

“The qualities of super ties — trust, conviction, and commitment — might actually lead to better science,” he says. Frequent collaborators know each other’s strengths and weaknesses, and they aren’t afraid to offer constructive criticism, which can be a rare commodity. “Feedback can be hard to get these days,” he told Nature.

But it is also possible, Petersen adds, that collaborative works might draw more citations partly because long-term partnerships can increase opportunities for the collaborating scientists to cite their own work.

Borniger says that he is already trying to foster partnerships for his own career. “I hope all of them become long-term relationships, but I don’t think there’s any way to force it except by consistently doing good science,” he says.

For more, see www.nature.com/socialselection.