Sunnyvale, California - A new study reveals LinkedIn ran social experiments with data from 20 million users.

The research, published in the journal Science earlier this month by researchers from LinkedIn, M.I.T., Stanford, and Harvard Business School, aimed to assess the role of social ties in the job search process, which could improve how the platform performs for job hunters.

The experiments, conducted between 2015 and 2019, focused on the website's People You May Know algorithm, which is used to recommend new connections to users.

But even though the study was intended to improve user experiences on the site, it may have unintentionally impacted users' access to job opportunities.

The experiment worked by randomly varying the proportion of "strong" and "weak" contacts suggested in the People You May Know section. That means some people were connected to more close friends, while other users saw more loose acquaintances recommended.

The problem, however, is that it turns out that second category increases your chances of getting a job. According to the study's finding, weaker social ties were "twice as effective in securing employment" compared to stronger ties.

This naturally leads to questions about users unknowingly being put at a disadvantage while the algorithm experiments were in effect. After all, some people were randomly assigned a higher proportion of stronger social ties, which the research suggests could have affected their opportunities to find a position.

Questions about data privacy have also been brought up.

Karthik Rajkumar, a LinkedIn researcher who co-authored the study, insisted that "[no] one was put at a disadvantage to find a job."

LinkedIn maintains that the research conducted using the data is in accordance with the site's user agreement terms, and the privacy policy states that the site may use personal data for research purposes.