The year is 2016 and unfortunately, sexism is still commonplace. The latest manifestation of this has appeared in the field of coding: Although women are considered better coders than their peers, this only applies if their gender remains hidden, as a study in the journal PeerJ reveals.
Women coders are considerably outnumbered by their male colleagues: in 2013, just 11.2 percent of coders were female. However, it wasn't always this way; for decades, the number of women studying computer science was rising far faster than men, and many computer science pioneers were female. Sadly, with the rise of personal computers across homes in the West, the proportion of female computer scientists dropped dramatically. One possible reason for this is that the marketing behind these computers was aimed almost exclusively at men, causing a huge uptick in the amount of male computer users and, ultimately, coders.
Wondering if this huge imbalance is helping to foster sexism in computer science, this new study began with the assumption that coders would be overly critical of code written by women, a type of prejudice seen across all types of technology industries. They turned their attention to GitHub, one of the most prolific open-source software communities in the world.
This giant, constantly updated library of code is currently used by 12 million people, who at any time can put out a call for code they require for their project. When a coder writes some code for this project, it’s known as a “pull request,” and the authors of this new study used these to determine if male or female coders were more likely to be accepted.
After sifting through near to three million pull requests submitted on GitHub, they found – against their expectations – that code written by women was approved at a slightly higher rate than that composed by men, 78.6 percent to 74.6 percent, respectively. That’s roughly an extra 120,000 pieces of code.
Women coders are apparently better than men. welcomia/Shutterstock
Although there are numerous possible reasons why this was the case, the researchers ruled several out, including whether women were only outperforming men using certain kinds of coding language that are more likely to be accepted in general. In fact, women’s coding acceptance rates dominate men’s in the top 10 programming languages, including Java and C.
The chance that this was because these female coders were already well-known in their field was also ruled out. With “insiders” – explicitly identifiable, renowned GitHub collaborators looking for respected coders – removed from the analysis, women’s acceptance rates were still higher than men’s, at 64.4 percent to 62.7 percent.
Nevertheless, there was a depressing caveat: this rule only applied if their gender was unidentified. If female coders’ gender was known, their overall pull request acceptance rate fell from 78.6 percent to 62.5 percent. This appears to suggest that women may in fact be better coders, but are automatically discriminated against simply because of their gender.
This study is yet to be peer-reviewed, although if accepted, its findings will stand in stark contrast to the origins of computer science. After all, one of the founders of computer programming was a woman, Ada Lovelace, back in the 1840s.