On the internet, a famous New Yorker cartoon suggested, no one knows you’re a dog, but if we believe that may help to create a purely meritocratic environment, we may need to rethink our assumptions.  Research from the University of Waterloo examined the contributions to the GitHub repository of open-source software projects and found that the supposed race of contributors affected how people viewed their contributions.

On GitHub, the only identifying information about an individual is the username they choose to represent them on the platform.  The contributions of users are discussed through “pull requests”, which are the means by which changes to the software are proposed and collaborated on.  The study finds that the perceived race and ethnicity of users, based on their username, has an impact on how others respond to them.

“A developer’s contributions to an open-source software project are accepted or rejected for a variety of technical reasons, but our analysis of tens of thousands of projects on GitHub shows that contributions can be accepted or rejected because of other factors,” the researchers say. “We found that one of them is the perceived race and ethnicity of a developer based on the person’s name on the platform.”

Project contributions

The researchers analyzed over two million pull requests spread across around 37,000 different projects, which had around 366,000 developers contributing to them in total.

They utilized a tool, called NamePrism, to estimate the race and ethnicity of the developers based on their username.  The tool estimated that 70% or so of the contributions were from white developers, with less than 10% from those who were identified by the tool as Asian, Hispanic, or Black.

“This low percentage is concerning because it does not reflect the percentage of developers among these groups in the larger tech community,” the authors say.

What’s more, the analysis also found that the chances of any particular contribution being accepted by the project integrators on GitHub were lower from developers who were perceived as being non-white.

“Perceptible Hispanic and Asian developers had six to 10 per cent lower odds of getting their pull requests accepted compared with perceptible white submitters,” the researchers conclude. “We need to identify the problems, understand why the problems exist, and determine what interventions can help reduce and eliminate bias.”