• by Erik Sass , Staff Writer, November 22, 2013

With more and more employers researching prospective hires on the Internet and social media in particular, there’s been a lot of discussion about the potential for risqué or unflattering content to damage jobseekers’ chances. And many people (including myself) are inclined to say “too bad” -- is someone chooses to put those compromising pictures or statements up in a public forum, then they just have to live with the consequences. But what about more subtle (and less rational) forms of discrimination?

Employers do indeed discriminate on the basis of things like religious affiliation, according to a new study from Carnegie Mellon University, which found significant disadvantages for jobseekers whose profiles indicated that they were Muslim, compared to jobseekers whose profiles indicated that they were Christian.

The study, titled “An Experiment in Hiring Discrimination via Online Social Networks,” tested response from more than 4,000 U.S. employers to fictitious job applicants who shared information about their faith in social media profiles and posts. To control for more reflexive discrimination based on appearance, the researchers made sure the fictitious Muslim candidate had an American-sounding name and Caucasian profile photo.

The study found that applicants whose profiles suggested they were Muslim got fewer callbacks than their Christian counterparts. Since the social media hints were the only area of difference between the applicants, it is reasonable to conclude that employers were searching for information on the applicants and judging them (consciously or unconsciously) on that basis. The disparity was most pronounced in conservative parts of the country, where Christian applicants got callbacks 17% of the time, versus just 2% for Muslim applicants.

On the other hand, a similar experiment conducted with fictitious gay and straight applicants found no significant discrimination on the basis of sexuality, even in more conservative parts of the country.