Data on 20 million LinkedIn users put a 50-year-old social science theory to the test
The key to landing your dream job could be connecting with and then sending a single message to a casual acquaintance on social media.
That’s the conclusion of a five-year study of over 20 million users on the professional networking site LinkedIn, researchers report in the Sept. 16 Science. The study is the first large-scale effort to experimentally test a nearly 50-year-old social science theory that says weak social ties matter more than strong ones for getting ahead in life, including finding a good job.
“The weak tie theory is one of the most celebrated and cited findings in social science,” says network scientist Dashun Wang of Northwestern University in Evanston, Ill., who coauthored a perspective piece in the same issue of Science. This study “provides the first causal evidence for this idea of weak ties explaining job mobility.”
Sociologist Mark Granovetter of Stanford University proposed the weak tie theory in 1973. The theory, which has garnered nearly 67,000 scientific citations, hinges on the idea that humans cluster into social spheres that connect via bridges (SN: 8/13/03). Those bridges represent weak social ties between people, and give individuals who cross access to realms of new ideas and information, including about job markets.
But the influential theory has come under fire in recent years. In particular, a 2017 analysis in the Journal of Labor Economics of 6 million Facebook users showed that increasing interaction with a friend online, thereby strengthening that social tie, increased the likelihood of working with that friend.
In the new study, LinkedIn gave Sinan Aral, a managerial economist at MIT, and his team access to data from the company’s People You May Know algorithm, which recommends new connections to users. Over five years, the social media site’s operators used seven variations of the algorithm for users actively seeking connections, each recommending varying levels of weak and strong ties to users. During that time, 2 billion new ties and 600,000 job changes were noted on the site.
Aral and his colleagues measured tie strength via the number of mutual LinkedIn connections and direct messages between users. Job transitions occurred when two criteria were met: A pair connected on LinkedIn at least one year prior to the job seeker joining the same company as the other user; and the user who first joined the company was there for at least a year before the second user came onboard. Those criteria were meant in part to weed out situations where the two could have ended up at the same company by chance.
Overall, weak ties were more likely to lead to job changes than strong ones, the team found. But the study adds a twist to the theory: When job hunting, mid-tier friends are more helpful than either one’s closest friends or near strangers. Those are the friends with whom you share roughly 10 connections and seldom interact, Aral says. “They’re still weak ties, but they are not the weakest ties.”
The researchers also found that when a user added more weak ties to their network, that person applied to more jobs overall, which converted to getting more jobs. But that finding applied only to highly digitized jobs, such as those heavily reliant on software and amenable to remote work. Strong ties were more beneficial than weak ties for some job seekers outside the digital realm. Aral suspects those sorts of jobs may be more local and thus reliant on members of tight-knit communities.
The finding that job seekers should lean on mid-level acquaintances corroborates smaller studies, says network scientist Cameron Piercy of the University of Kansas in Lawrence who wasn’t involved in either the 2017 study or this more recent one.
That evidence suggests that the weakest acquaintances lack enough information about the job candidate, while the closest friends know too much about the candidate’s strengths — and flaws. “There’s this medium-ties sweet spot where you are willing to vouch for them because they know a couple people that you know,” Piercy says.
But he and others also raise ethical concerns about the new study. Piercy worries about research that manipulates people’s social media spaces without clearly and obviously indicating that it’s being done. In the new study, LinkedIn users who visited the “My Network” page for connection recommendations — who make up less than 5 percent of the site’s monthly active users — got automatically triggered into the experiment.
And it’s unclear how LinkedIn, whose researchers are coauthors of the study, will use this information moving forward. “When you are talking about people’s work, their ability to make money, this is important,” Piercy says. The company “should recommend weak ties, the version of the algorithm that led to more job attainment, if its purpose is to connect people with work. But they don’t make that conclusion in the paper.”
Another limitation is that the analyzed data lacked vital demographic information on users. That was for privacy reasons, the researchers say. But breaking down the results by gender is crucial as some evidence suggests that women — but not men — must rely on both weak and strong ties for professional advancement, Northwestern’s Wang says.
Still, with over half of jobs generally found through social ties, the findings could point people toward better ways to hunt for a job in today’s tumultuous environment. “You may have seen these recommendations on LinkedIn and you may have ignored them. You think ‘Oh, I don’t really know that person,’” Aral says. “But you may be doing yourself a disservice.”