LinkedIn conducted experiments with more than twenty million users over five years which, while aiming to improve the functioning of the platform for members, could have affected the income of some peopleaccording to a new study.
In experiments conducted around the world from 2015 to 2019, LinkedIn randomly varies the proportion of weak and strong contacts who suggested your algorithm “People You Might Meet” (the company’s automated system for recommending new connections to its users). The evidence was detailed in a study published this month in the journal Science co-authored by researchers from LinkedIn, Massachusetts Institute of Technology (MIT), Stanford University and Harvard Business School. .
LinkedIn’s algorithmic experiments could surprise millions because the company it did not notify users that testing was ongoing.
Routine tests that users are unaware of
Big tech companies like LinkedIn, the world’s largest professional network, regularly run large-scale experiments in which they test versions of app features, web designs, and algorithms on different people. The long-standing practice, called A/B testing, aims to improve consumer experiences and keep them engaged, helping businesses earn money from premium membership fees or advertising. Users often have no idea that companies are testing them.
However, the changes made by LinkedIn indicate how such widely used algorithm modifications can become experiments in social engineering with consequences that can change the lives of many people. Experts who study the effects of computing on society have said that conducting large-scale, prolonged experiments on people who could affect your job prospectsinvisible to them, has raised questions about industry transparency and research oversight.
Michael Zimmer, Associate Professor of Computer Science and Director of the Center for Data, Ethics, and Society at Marquette University, commented, “Results indicate that some users had better access to job opportunities or a difference significant in access to job opportunities. These are the kind of long-term consequences that have to be taken into account when thinking about the ethics of participating in this type of research data intelligence”.
The study in Science examined an influential theory in sociology called “The strength of weak ties”, indicating that people are more likely to obtain employment and access other opportunities through less close acquaintances than through close friends.
Researchers analyzed how changes to LinkedIn’s algorithm affected users’ career mobility. They found that relatively weak social connections on LinkedIn turned out to have twice as efficient to secure a job strengthen social ties.
LinkedIn, which is owned by Microsoft, did not directly respond to a question about how the company calculated the potential long-term consequences of its experiments on users’ jobs and economic status. However, the company assured that the investigation it had not provided a disproportionate benefit to some users.
Karthik Rajkumar, applied research fellow at LinkedIn and co-author of the study, explained that the goal of the research was “to help people at scale. No one was disadvantaged in finding a job.
Sinan Aral, professor of management and data science at MIT and lead author of the study, said the LinkedIn experiments were a move to ensure users have equal access to job opportunities.
Aral clarified: “Doing an experiment with twenty million people and then implementing a more suitable algorithm to improve everyone’s job prospects with the knowledge you have acquired with that is what they are trying to do, no not to grant social mobility to some people and not to others”. (Aral performed data analysis for The New York Times and received a Microsoft Research Grant in 2010.)
How was the LinkedIn experience
User experiments by large Internet companies have a patchy track record. Eight years ago, a Facebook study was published describing how the social network had covertly manipulated messages appearing in users’ news feeds to analyze the spread of negative and positive emotions on your platform. The week-long experiment, conducted with 689,003 users, immediately generated negative reactions.
LinkedIn’s networking experiences differed in intent, scope, and scale. They were designed by LinkedIn as part of the company’s ongoing efforts to improve the relevance of its “People You May Know” algorithm, which suggests new connections to members.
The algorithm analyzes data such as members’ employment history, job titles, and connections to other users. It then attempts to measure the likelihood of a LinkedIn member sending a friend invitation to a new suggested connection, as well as the likelihood of that new connection accepting the invitation.
For experiences, LinkedIn adjusted its algorithm to randomly vary the prevalence of strong and weak ties that the system recommends. The study reported that the first wave of testing, conducted in 2015, “involved more than four million subjects in the experiment”. The second wave, carried out in 2019, involved more than sixteen million people.
In testing, people who clicked on the “People You May Know” tool and viewed the recommendations were assigned different algorithmic paths. Some of these “processing variants,” as the study calls them, caused LinkedIn users to connect more with people with whom they had only weak social ties. Other changes caused people to form fewer links with weak links.
It is not known whether the majority of LinkedIn members they know they might be subjected to experiments that could affect your job opportunities.
However, None of these policies explicitly inform consumers that LinkedIn itself may be experimenting or testing its members..
In a statement, LinkedIn said, “We are transparent with our members in the research section of our Terms of Service.”
In an editorial statement, Science said, “It was our understanding and that of our reviewers that the experiments undertaken by LinkedIn operated within the guidelines of their Terms of Service.”
How strong is weak tie theory?
Following the first wave of algorithmic testing, researchers from LinkedIn and MIT came up with the idea of analyzing the results of these experiments to test the weak tie strength theory. Although this decades-old theory has become a foundation of sociology, had not been rigorously tested in a large-scale prospective trial that randomly assigned people with varying degrees of social connection strength.
External researchers analyzed aggregate data from LinkedIn. The study found that people who received more referrals from moderately low contacts applied and accepted more jobsresults consistent with weak tie theory.
The study reported that the 20 million users involved in LinkedIn experiences created more than 2 billion new social connections and submitted more than 70 million job applications that led to 600,000 new jobs. The study also indicated that weak ties have proven most useful for job seekers in digital fields such as artificial intelligence, while strong links have proven more useful for jobs in industries that are less reliant on software.
LinkedIn said it applied the weak tie findings to several features, including a new tool that notifies members when a first- or second-degree connection is hiring. However, the company has not made any study-related changes to its “People You May Know” feature.
According to MIT’s Aral, the most important thing about the study was that it demonstrated the importance of powerful social media algorithms, not only for amplifying issues such as misinformation, but also as critical indicators of economic conditions such as as employment and unemployment. .
Catherine Flick, senior IT and social responsibility researcher at De Montfort University in Leicester, England, described the study as an exercise in corporate marketing.
Flick concluded: “The study has an inherent bias. It goes to show that if you want to get more jobs, you should connect more to LinkedIn.”