Tie Strength

What it is

Tie strength is the quantity that sorts a person's relationships into close and distant. It is a property of a pair of people, not of either person alone, and it sits on a scale. Nobody has a fixed number of weak ties and strong ties in the way a person has a height. The labels are conveniences laid over a continuous measure, and where the line falls depends on the measure.

Mark Granovetter introduced the idea in The Strength of Weak Ties, published in the American Journal of Sociology in 1973. He described the strength of a tie as a combination, probably linear, of the amount of time, the emotional intensity, the intimacy and the reciprocal services that characterise it. That is a definition with four ingredients, and it is deliberately broad. A colleague you see every day and a cousin you call twice a year can come out similar in total for quite different reasons.

Granovetter used the idea to make a structural argument. People who are strongly tied to each other tend to share friends, so a tight cluster passes the same information round and round. A weak tie is the usual link between clusters. A bridge, in his usage, is a tie that provides the only path between two points in a network, and his conclusion was that all bridges are weak ties. He also described the configuration he thought least likely to occur: a person strongly tied to two others who have no tie to each other.

Because the definition has four ingredients and no instrument, later researchers have had to choose something they can count. That choice is where most of the disagreement in the literature now sits. The same word, strength, has been attached to at least three different measurements, and the studies that use them do not always agree.

In effect

Granovetter's own measure: how often people met

In the data Granovetter used to illustrate the argument, strength was reduced to a single question, how often the person who gave a job lead was seen. Of 54 people who found their job through a contact, 16.7 per cent saw the contact often, 55.6 per cent occasionally and 27.8 per cent rarely. Frequency of contact is a simple proxy for the amount of time two people spend together, which is one of his four ingredients. It leaves out intensity, intimacy and services altogether.

Two yardsticks on one platform

Rajkumar and colleagues (2022) ran randomised experiments on LinkedIn's People You May Know suggestions, across more than 20 million members between 2015 and 2019, and measured strength in two ways. By the number of mutual connections, job transmission rose with more mutual connections up to around ten and then fell. By the amount of messaging, the weakest ties raised job transmission the most and the strongest raised it the least. The same experiment therefore gives two different answers about which ties help, depending on which version of strength is plotted. The authors list the dependence on the measure as one of their three revisions to the theory.

An everyday illustration

Imagine two people who each share ten mutual connections with you. One messages you every day about a shared project. The other has not written in two years. On the mutual-connections yardstick they are identical, and on the messaging yardstick they sit at opposite ends. This is an illustration of the problem, not a finding from any study, but it is the reason a claim about weak ties has to say which yardstick it used.

What it does not say

Tie strength does not say that distant relationships are better in general. Granovetter's claim concerned the information a bridge carries, and he acknowledged that people with whom one has strong ties are more motivated to help. Gee, Jones and Burke (2017), using data on about six million Facebook users, found that most people were helped through one of their many weak ties, but that a single stronger tie was significantly more valuable at the margin.

It does not give a universal threshold either. The peak at around ten mutual connections is a result for LinkedIn members in the years of that experiment, and the same paper reports that weak ties helped more in digital industries while strong ties helped more in less digital ones. The experiment tracked applications in its 2019 wave, not completed hires.

Finally, the concept is a measurement problem and not a body of wellbeing research. Claims that brief exchanges with acquaintances improve mood belong to a separate literature, described on the weak ties page.

---