A recent study conducted by researchers from the computer science departments at Cal Poly, San Luis, Obispo and North Carolina State University reports that women write better code than men.
The U.S. researchers analyzed data from Github, an open source program-sharing surface. Looking at about 1.4 million users who logged onto the site on April 1 2015, the researchers found that “pull requests,” or suggested code changes, made on Github by women were more likely to be accepted by peers than those suggested by men.
Though Github does not request gender information from its enormous developer community consisting of 12 million users, the research team was able to identify whether roughly 1.4 million were male or female. The study says this was concluded due to information on the users’ profile page, or because their email address could be matched with another social network to confirm gender.
In the analysis, the research team found that 78.6% of pull requests made by women were accepted compared to 74.6% being accepted by those made by men. They did not find a correlation among a few factors they looked into, for example, whether women were more likely to be responding to known issues, whether their contributions were shorter in length and easier to appraise, and which programming language they were using.
Here’s where it gets interesting: the study found a clear gender bias, as women’s acceptance rates drop down to 62.5% when their gender is identifiable. Though the study notes that there is a similar drop for men, it is not as significant.
The study is currently awaiting peer review, meaning other experts have not yet critically appraised it. But, the study concludes that their results suggest women have a higher acceptable rate of pull request overall, but when their gender is known, they have a lower acceptance rate than men. Therefore, despite women on Github being more competent overall, a bias against them does exist.