Women Deemed Better Coders— if Not IDed as Women

When a group of student computer scientists decided to test their hypothesis that software coding done by women doesn't get as much cred as that done by males, they came face to face with a couple of surprises. In the not-yet-peer-reviewed study published in the open-access PeerJ journal, researchers reviewed data culled from the GitHub, a giant open-source software hive that's faced its own accusations of sexism. The scientists found that out of 3 million "pull requests (when a developer writes code for someone else's project), developers more readily accepted code written by women (78.6%) than that written by men (74.6%), per the Guardian. The researchers eliminated a number of factors that could be clouding the data, such as females perhaps making tinier code adjustments that project owners could more quickly evaluate.

But what the scientists found next somewhat dulled the equality euphoria. They wanted to see if women coders were reaping the benefits of reverse gender bias, perhaps getting more approvals in efforts to promote diversity—instead, they found the opposite. In separating women developers into those whose profiles clearly IDed them as women and those whose didn't (and looking only at "outsiders," which excludes project owners and collaborators), the acceptance rate for gender-neutral profiles was 71.8% to female profiles' 62.5%. But some developers say finding gender bias within GitHub is more involved than simply adding up the numbers: A pull request could also be passed over because the potential recipient was simply swamped or opted to accept the work of a developer he or she knew. (More gender bias stories.)