A new study has found that women are more likely than men to have their open-source software contributions accepted – but only when their gender is hidden from project leaders.
The study from North Carolina State University and Cal Poly examined code committed by more than 1.4 million GitHub users and their contributions to various open-source projects on the source-code repository service.
The researchers found that women have their pull requests (or suggested changes to code) accepted by project owners more often than men overall across all programming languages, with one important caveat: acceptance rates for women drop lower than those of men when their gender is made known.
The non-peer-reviewed study separated profiles into two categories: gender neutral profiles where the profile name and generic image hid the user gender, and gendered profiles where a user photo and a gendered user name (such as "JustinA") would identify the gender.
The results showed that when gender-neutral profiles were analysed, accounts belonging to women achieved higher acceptance rates than those owned by men. When the gendered profiles were analysed, however, the results showed that project owners were more apt to accept pull requests from men.
The researchers noted that familiarity plays a major role in showing the bias. When contributions from "insiders" who were known and trusted within a project were analysed, the gender differences disappeared.
Rather, it is when analyzing "outsider" pull requests from people not well known to the project that the disparity between the genders became truly apparent, and men were shown to be more likely to have requests accepted.
In short, as a whole women contribute more successful submissions to GitHub than men do, but when faced with the choice between the submissions of a man and a woman, a project leader is more apt to use code from a man.
As a result, the researchers suggest that overall, the women contributing code to GitHub are more competent than their male counterparts, with the theory being that higher attrition rates for women in the lower levels of STEM careers lead to higher levels of average training and experience.
Despite this, the researchers also conclude that women are still seen as less competent than men and, when competing on projects, are at a disadvantage.
"While our big data study does not definitely prove that differences between gendered interactions are caused by bias among individuals, the trends observed in this paper are troubling," the paper concludes.
"The frequent refrain that open source is a pure meritocracy must be reexamined."
The study, Gender bias in open source: Pull request acceptance of women versus men was written by Josh Terrell, Andrew Kofink, Justin Middleton, Clarissa Rainear, Emerson Murphy-Hill, and Chris Parnin. ®