Facebook Says Its Female Coders Are Not Victims Of Gender Bias. Sort Of
A former Facebook employee analyzed five years of data and concluded that code written by female engineers was less likely than code written by male engineers to make it through Facebook’s internal peer review system.
Facebook refuted the analysis, saying it was incomplete, inaccurate and used incorrect data. It conducted its own analysis and concluded that it was rank, not gender, that determined whether code got accepted. Facebook has more female engineers at lower levels than at higher levels.
The internal Facebook analysis found no statistically significant difference between female and male engineers within the same level, Facebook said.
The original analysis was posted in September, and Facebook’s own analysis was posted in October. Both were shared internally at Facebook in 2016, but have become part of a national debate after this week’s report by the Wall Street Journal. The Guardian says it has independently confirmed the report.
A longtime Facebook software engineer studied Facebook’s code review process, looking at the number of times code was rejected, commented upon, or updated; how long it took for code to be accepted; and demographic data about the coder, such as gender and length of employment, the Guardian reported.