We’ve long heard about the need for more female representation in STEM, and thanks to a new study on gender bias in online programming communities, the argument just got a little stronger.
As it turns out, gender bias against females working in tech isn’t just rampant in the physical workplace — it rears its ugly head on the internet, too. A recent study in Science Daily found that female programmers got better marks in programming than men — but only when their identity was gender-neutral.
The study aimed to explore whether or not a gender bias exists when pull requests are judged on GitHub, a popular programming site that invites users to collaborate on open-source projects. Projects are posted with an open call for feedback, and when a user feels they’ve identified an improvement to the code, they can submit what’s called a “pull request” to make the change official. The requests are then judged by designated project managers called “insiders,” who also have access to a user’s profile and identity.
After analyzing over 3 million pull requests submitted by 330,000 users (21,000 of whom were female) the researchers found that, of those accepted, 78.7 percent of acceptable pull requests came from women, while only 74.6 percent came from men. It’s a narrow margin, unsurprisingly, but things get a little hairier when you compare the gender identification associated with the acceptances.
Programmers who were obviously identifiable as female (either by profile picture or username) produced lower pull request acceptance rates than users obviously identifiable as male, with women bringing in an acceptance rate of just 58 percent, compared to the male rate of 61 percent.
I don’t know about you, but that leaves me feeling a little confused.
…The majority of accepted pull requests came from women, but when compared side by side, female-identifying users’ ideas were only accepted 58 percent of the time, while male ideas were accepted 61 percent of the time. Am I missing something?
Acceptance Rates Were Higher for Women in Gender-Blind Requests
I am indeed missing something, as were the insiders on the project. A further drill-down into the data revealed that pull requests submitted by females who chose to keep their gender anonymous had better overall marks than men who chose to keep their gender anonymous, with women’s ideas being accepted 70 percent of the time. The ideas of gender-anonymous men, by contrast, yielded an acceptance rate of just 65 percent.
Female programmers who kept their gender anonymous has better ratings than men who did the same.
So what’s it all mean? The researchers concluded that gender bias does in fact exist on the site, and that female programmers on GitHub are capable programmers — not necessarily because of their gender, but because of a strong tendency to self-select based on competency when considering whether or not to submit a pull request. But the implications may suggest something deeper at play than that.
The sad truth is that there really isn’t anything too surprising or ground-breaking about the results of the study. Of course women are capable programmers, and of course they’re being discriminated against based on their gender. Decades of being passed up for tech-heavy jobs and leadership roles, ignored in early science, math, and engineering courses, and pushed into clerical careers with a heavy emotional labor component have both reinforced a gender bias against women in STEM and excluded them from full consideration in the field.
Further, it almost makes sense that women would score higher on a site like GitHub, as years of underestimation have left many female programmers particularly aware of (if not sensitive to) their own abilities, along with the unfair onus to prove their worth. Because women have for so long been conditioned to second-guess their work and therefor double and triple-check it for accuracy, it makes sense that their solutions would be largely error free — possibly even more so than their male counterparts.
Thanks to the open source nature of programming in general, women have been able to break into a tech industry that didn’t readily welcome them in the first place. Thus, it’s not surprising that female programmers — a disruptive force by necessity — are pulling ahead in an industry built around disruption, hacking, and new developments. In fact, it makes perfect sense; they all but hacked their way into the industry in the first place.
Is there a gender bias in programming? Without question. And are we getting closer to closing that gap? Absolutely.
Tell Us What You Think
What do you make of these stats? We want to hear from you. Tell us your thoughts in the comments or join the conversation on Twitter.