Is it really a wise decision to accept random people on LinkedIn?

Introduction

In class, we discussed the work of Mark Granovetter where he makes a connection between the social and structural roles of an edge and claims that weak ties have access to different parts of the network. In the slides, we learned of the strength of weak ties theory which states that networks are composed of tightly connected sets of nodes. This means that there are strongly connected components formed of strong ties with weak ties acting as local bridges to connect various strongly connected components together. Upon discovering a recent research paper regarding a paradox of weak ties, released by LinkedIn, it got me wondering to what extent do weak ties truly affect one’s social and informational network. As I am currently looking for an internship, the paper also made me wonder whether attempting to purposely accept weak ties, with talent acquisition specialists or current employees at companies, would result in a higher chance of acquiring a job through LinkedIn.

Research

A research paper consisting of 5 years of data collection was recently released by LinkedIn to test the extent to which weak ties increased job mobility through LinkedIn. The experiment worked by varying the frequency of weak ties in LinkedIn’s “People You May Know” (PYMK) algorithm and was experimented on over 20 million members on LinkedIn. There were 2 billion new connections and 600,000 jobs created through LinkedIn over the 5 years. As the paper is directly referring to the strength of weak ties theory, it closely refers to Granovetter’s work and provides results by limit-testing to show the true impact and effectiveness of weak ties in job mobility. The result of the paper is interesting as it will assist LinkedIn in optimizing its PYMK machine-learning model to increase job mobility and users can also reap the benefits from the paper as they learn about what connecting with strong, medium, and weak ties can potentially do for them.

Analysis

The test analyzed the labour market mobility by measuring both the number of job applications on LinkedIn and job transmission. It was measured by looking at how engaging the interaction between two ties was and the number of mutual connections they had. In fact, a formula to measure the structural tie strength was created as follows.

In the formula, i and j are LinkedIn members, Mij is the number of mutual connections between them, and Di and Dj are the total numbers of direct connections of members i and j respectively.

From recent data, the paper identifies a “paradox of weak ties” and it suggests that strong ties are actually more valuable than weak ties in generating job transmissions. The following charts show the results of the ordinary least squares (OLS) analysis.

The following charts show that job transmission per tie is actually higher with strong ties, then weak ties, and lastly medium ties. It also tells us that having more mutual friends does lead to a higher probability of job transmission, but it will eventually flatten out. Evidently, it also holds that there’s a higher probability of job transmission provided more interactions. Something that was surprising was that it seemed to flatten out at 0.25 interaction intensity and stayed consistent until 1.00 interaction intensity. These were merely estimates by the linear regression and the study attempts to see whether these estimates truly hold. On the contrary, the study identifies that in reality, at low levels, mutual connections have a high probability of job transmission, but as a member gets more mutual friends, the probability of job transmission actually decreases. The final results showed that adding new structurally diverse ties with weak interaction intensity created the greatest marginal increases in terms of job transmission.

Conclusion

Although this paradox exists, it does not mean that Granovetter was incorrect. When studied by Granovetter, the times were different and LinkedIn is different from other social networks such as Facebook and Twitter as it is used mostly for professional reasons. When it comes to the initial concern of whether or not LinkedIn members should be purposely connecting with weak ties in attempts to secure job positions, then clearly in LinkedIn’s case, it should not be prioritized. Although Granovetter’s strength of weak ties theory suggests that weak ties act as a local bridge from one strongly connected component to another, it is not necessarily the case for LinkedIn’s PYMK algorithm. From the statistical analysis done by the algorithm, it is evident that the weak ties do increase job transmission to a point, but after that, it is wholly dependent on other factors such as maintaining the connection, achieving mutual connections, and creating strong ties. Regarding attempting to optimize your position in the algorithm, it should not be a reason to not connect with someone on LinkedIn. It is perfectly fine to connect with others and still maintain a good position in the PYMK algorithm!

Sources 

Rajkumar, K., Saint-Jacques, G., Bojinov, I., Brynjolfsson, E., & Aral, S. (2022). A causal test of the strength of weak ties. Science, 377(6612), 1304–1310. https://www.science.org/doi/10.1126/science.abl4476