A new study that used AI to analyze 2M+ contributions by ~365K developers on GitHub finds users with white-sounding names may have more success on the platform
Software developers with names with a “white” sound could have more popularity on GitHub than those whose names are interpreted by users as “black,” “hispanic,” or “Asian-Pacific Islander,” as per the results of a study recently published.
The results, released earlier this month in IEEE Transactions on Software Engineering, are a serious concern regarding the implications of the absence of diversity on GitHub and within the open-source community generally.
Researchers from the University of Waterloo The University of Waterloo analysed more than 2 million contributions, also known as “pull requests,” made by the 365,607 developers who use GitHub. Making use of an AI tool known as NamePrism, which analyses the names of people according to their perceived ethnicity and race, Researchers found that being seen by GitHub users as “white GitHub” generally improves a developer’s chances of being accepted. This increases the chances by 6 to 10% when compared to developers perceived to be Hispanic or Asian Pacific Islander.
“Theoretically, it’s the only area where there is the possibility of having a complete meritocracy.” It’s rare to see someone working with open-source software. You’re unlikely to have seen them or had an opinion about them. “At most, you know their name.” Mei Nagappan, an assistant professor of computer science at the University of Waterloo who co-authored the study, explained
“A racial prejudice is still present, even in this context,” Nagappan said, considering the power that open-source communities like GitHub have on the development of products.”If we don’t listen to varied views, software is produced by and for a pretty homogeneous community,” Nagappan said.
It’s not just that; GitHub has evolved into a type of portfolio site for developers. This means this particular bias could negatively impact career opportunities for developers. “If you have contributions accepted to even one of the big projects, then as a newcomer, you could translate that to a successful career in a company,” Nagappan explained.
GitHub hasn’t responded to Protocol’s request for comments. The company also did not provide a response to the protocol’s request for comment. Nagappan explained that the aim of the research isn’t just to focus on GitHub specifically but to address the issues within the open-source community in general. Nagappan explained that these findings build on previous research that found female developers on GitHub have lower rates of acceptance.Acceptance rates have also been observed to differ based on the country of origin of the developers.
He explains that it’s true that the NamePrism tool that his team utilised does not have the best accuracy in making predictions about people’s race or ethnicity. Researchers did not assign a race or ethnicity to developers in cases where the tool had an extremely high degree of confidence. For the rest, they classified the developer’s race as “unknown.”
Although the Waterloo researchers did not shy away from the possibility of attributing this racism on GitHub to a specific reason, they did discover that an overwhelming majority of the developers who submit ideas to GitHub and the majority of people who respond to their contributions are identified by names that researchers believed were white. Additionally, they discovered those developers viewed as black, Hispanic, or Asian-Pacific Islanders are more likely to receive pull requests when the people who respond to them belong to the same ethnic or racial group.
To eliminate this potential bias, researchers suggest that GitHub use a double-blind or single-blind model that is similar to the way research is evaluated in academia. Another suggestion is to have multiple people evaluate a particular researcher’s contribution so that bias from one person does not interfere.
The issue of how the perceptions of race influence the online interactions of people isn’t just a matter for GitHub. In the year 2000, Airbnb launched a research project known as Project Lighthouse that also aimed to examine how racial prejudice occurs on the site and the role that the names of users can influence others’ opinions.
new 2m 365k githublapowskyprotocol
new ai githublapowskyprotocol
