Social capital → Employment status

Mechanism

Close friends share the same pool of news. Acquaintances sit in other clusters. Granovetter (1973) argued that is why a weak tie is the one that tells you about a job you have not already heard of. Rajkumar, Saint-Jacques, Bojinov, Brynjolfsson, and Aral ran large randomized experiments on LinkedIn's People You May Know algorithm, which recommends new connections. The experiments randomly varied how many of those recommendations were weak ties in the networks of more than 20 million members over five years. About 2 billion new ties formed. About 600,000 new jobs appeared. More weak ties raised job mobility. They also revised the theory. Tie strength and job transmission form an inverted U: weaker ties moved more jobs up to a point, then further weakness added less. Two ways of measuring a weak tie did not agree. Ties that were moderately weak by number of mutual connections, and the weakest ties by how often people interacted, created the most job mobility. Industry split the result. Weak ties raised job mobility in more digital industries. Strong ties raised it in less digital ones. Who you know is a hiring channel. That is also why a network is a backdoor in studies that credit a vitamin or another purchasable input for a career.

Caveats

Rajkumar randomized a recommendation algorithm on LinkedIn, not a program that tells people to collect acquaintances. Members of a professional network are already selected. The outcome is job mobility, including job changes by people who already had work. Employment status here means in work or not. Job mobility is the matching Granovetter wrote about. It is not only a switch from unemployed to employed. The inverted U and the industry split mean adding ever-weaker ties is not a universal knob. In less digital industries the strong ties did the work. Granovetter is observational. People befriend people like themselves. Education and money travel with networks outside the LinkedIn experiment.

Effect

qualitative. Weak ties carry job information (Granovetter 1973). LinkedIn PYMK experiments: more weak ties raised job mobility in 20 million members (Rajkumar 2022); inverted-U in tie weakness; industry split.

Nodes

Named confounders

• Educational attainment
• Disposable income

Papers

• The strength of weak ties

  M. S. Granovetter · 1973 · American Journal of Sociology

• A causal test of the strength of weak ties.

  K Rajkumar, G Saint-Jacques, I Bojinov, E Brynjolfsson, S Aral · 2022 · Science (New York, N.Y.)