Looking for a job can be like having a really bad full-time job that only pays you in disappointment. You have to put together your resume and cover letter then scour through job listings and hope for the best. Resume search engines and recruiting services offer a way to make the job hunt less painstaking. All job seekers have to do is upload their resume apply to the jobs that the service presents you with. Recruiters can also see a list of candidates ranked by relevance and choose who they like best.
But what happens when your resume is ranked lower regardless of how qualified you are? Christo Wilson, professor at Northeastern University, conducted a study that analyzed how hiring websites Careerbuilder, Indeed and Monster rank resumes. As a part of the study Wilson and his team were conducted an algorithmic audit to test specifically for any gender bias in the algorithms these services operate on.
They tested for two kind of fairness: group and individual. Group fairness means that the hiring engine would present recruiters with equal amounts of men and women because there is no difference between the two groups. Individual fairness means that the ranking should be based on an individual’s qualifications making someone more qualified appear before someone with less qualified.
After the group examined data on about 855,000 job candidates they found that there was some individual unfairness that would place men over women despite their qualifications. The group found even more group unfairness, especially for positions in tech. “Male candidates are getting higher ranks across the board,” explained Wilson.
Wilson links this group unfairness in tech jobs to the culture of the tech industry that can be inhospitable to women.
Rankings are important because people trust hiring technologies to show us the best of the best on top. How often do you click beyond the second page on a Google search? So if you aren’t at the top of the list then you pobably won’t get the job.
Wilson says that if things like algorithmic biased aren’t fixed, the issue of gender disparity in the workplace will only get worse, especially because the psychological effects algorithms can have.
A visibility issue also arises for people who may not have resumes or cover letters to upload on a website. Applications to find domestic workers like nannies are changing the employment landscape for low-income women of color.
Researcher Alexandra Mateescu found that Care.com, a hiring service that connects families to nannies and babysitters is biased against low-income care workers in its design. “They are built with a particular archetype of what an ‘entrepreneurial’ domestic worker looks like one who feels at home in the world of apps, social media, and online self branding and ultimately replicates and can even exacerbate many of the divisions that came with our pre-digital workplaces“ explained Mateescu.
Care.com asks nannies to upload photographs and videos, write a bio, detail past work experience, list education and social media accounts. None of this is required, but nannies get ranked higher based on the most information they provide. Clients can also search for nannies within a specified zip code, leaving nannies that live outside of that area invisible. This is a problem for nannies who might not speak English, aren’t internet savvy, don’t have access to the internet or don’t feel comfortable sharing that much personal information online.
As our society gets more dependent on technologies to find employees, those who cannot participate in these technologies get left behind. Today the unemployment rate for black people in the U.S. is 6.6%, and 4.4% for Latinos.
Bias in online hiring services has the potential to worsen the employment rates for these communities if it isn’t addressed now.