We have seen various examples of how we can use data from Twitter, Facebook, and the web for social research. Similarly, LinkedIn is also used as a data source for social science research in recent years. LinkedIn is a social networking website for business and professional relationships. It may be one of the best networks for connecting with professionals and colleagues, however, LinkedIn’s API is different than other social networking sites. Obtaining API access to query the data on LinkedIn is a difficult task because of its restricted regulations. Google’s people management utilizes different methods such a predictive modeling and a retention algorithm to determine their employee’s performance and retention rate (Sullivan, 2013). There are many companies who use data mining methods to hire and retain employees. However, does data collected from LinkedIn accounts by using algorithms really help human resources to make informed decisions about hiring practices? Can these decisions be responsible for causing bias in hiring practices?
U.S. Equal Employment Opportunity Commission’s press release (2016) described the utility of using LinkedIn data in making employment decisions. Jenny Yang suggested that Big Data has a potential of generating new techniques that can reduce bias and promote anti-discrimination practices in hiring, performance evaluations, and promotions (EEOC, 2016). Dr. Michael Kosinksi, a professor of Organizational Behavior at Stanford Graduate School of Business also echoed Yang’s optimism and suggested that Big Data collected from professional networking website such as LinkedIn can increase employer’s ability to identify talent and increase equality in access to jobs. However, Dr. Ifeoma Ajunwa, a Fellow at the Berkman Klein Center at Harvard University and Assistant Professor of Law at University of the District of Columbia School of Law insisted that such practices make employees or potential hires vulnerable to discriminatory practices as employers may gain access to their personal information such as demographics, names, photographs etc.
In 2014, Senderowicz conducted a study to determine whether the impact of weight bias is manifested through selection process on LinkedIn for female job applicants. 194 participants who had previous hiring experience were hired via LinkedIn and asked to rate a screenshot of the LinkedIn profile picture of overweight and thin female job applicants. The findings suggested that there was no bias as overweight candidates did not receive lower ratings. However, this brings up a really important question. Does your appearance on LinkedIn profile picture really matter or contribute to the decision of hiring you or not? One may have an impeccable resume and skills that are required for accomplishing the job. However, the appearance of one’s LinkedIn profile or their personal information can be responsible for a decision against hiring them. In fact, Richter (2016) claims that academic identifying information such as the date of graduating from college is collected by algorithms and can often give away an applicant’s age, making them vulnerable to age discrimination from potential employers. Windham (2015) also notes that Google Plugins such as Age-Insight can scrape the data on LinkedIn by using the simple algorithm and calculate the age of a LinkedIn user based on the information they have provided in their profile. When LinkedIn challenged the Age-Insight developer for violating the terms of service and the Age Discrimination in Employment Act of 1967 which protects the people over 40 years old from discrimination, the developed decided to shut down the extension.
Even though these practices have practical use of hiring the perfect candidate quickly, many take this utility too far for their own agenda. It is very important to consider whether we are violating the basic rights of the potential employee or an applicant by mining their personal information.
EEOC. (2016). Use of Big Data Has Implications for Equal Employment Opportunity, Panel Tells EEOC. Eeoc.gov. Retrieved 3 April 2017, from https://www.eeoc.gov/eeoc/newsroom/release/10-13-16.cfm