Does Outreachy and a Code of Conduct increase diversity? Case study from Debian
The recently published and controversial FSFE diversity strategy only includes one ongoing action: updating the Code of Conduct.
Personally, I feel that diversity and the status of women is a vital issue. Therefore, I feel compelled to check if that strategy is built on solid ground. Does a Code of Conduct, in the sense that these things are written in hobbyist organizations, really increase diversity?
The Debian Project provides an interesting source of data for a study. Debian is one of the larger free software communities and a significant amount of data on project roles is available in public. There is a convenient list of female contributors.
If anybody has more significant data on this topic and would like to examine it together please feel free to contact me.
Interventions
Debian first participated in the Outreach Program for Women (now Outreachy) in summer 2013.
Debian introduced a Code of Conduct in April 2014.
There are two rounds of Outreachy each year. Approximately $6,000 from Debian funds is used to fund each intern. There are typically two interns in each round, so this uses $24,000 per year, a non-trivial sum for the project.
As both Outreachy and the Code of Conduct were introduced at almost the same time it is hard to view their impact separately. It is possible that one may have a net positive impact while the other may have a negative impact or vice-versa.
Analysis
- Only time spent as an uploading developer is considered
- Excluded: the first few years when Debian was very small
- Trans developers are counted as female throughout their history with the project
- The figures are absolute number of women, they are not weighted by the overall number of developers
Results
| Period | Women per year (mean) |
|---|---|
| 2004-2013 | 1.000 |
| 2014-2020 | 0.857 |
What this suggests is that the number of women granted upload access each year appears to have fallen by fourteen percent after Debian joined Outreachy and adopted a Code of Conduct.
This is not a rigorous study and I make no claim that there is a causal relationship between Outreachy, Code of Conduct and the willingness of women to contribute. It is just a quick summary of the available data to get discussion started about an important issue.
Comment
Outreachy encourages women to compete for positions while men come to Debian as volunteers. This may work against the interests of diversity.
Even though these figures are only informal, I feel there is good reason for this organization to use the $24,000 Outreachy money for a different diversity program.
The Code of Conduct is very generic and doesn't include the safety mechanisms we would find in a Code of Ethics. In practice, the Code of Conduct has encouraged witch-hunts, finger-pointing and conflict. The undignified and hasty manner in which a kangaroo court was assembled to attack Jacob Appelbaum with exaggerated accusations of harassment is a case in point.
Every organization and project is different: Debian, by definition, is a voluntary effort. A project with employees may see very different diversity outcomes.
Free software organizations are using the lack of women as an excuse to play with the carrot (Outreachy) and stick (CoC). If there is no data to support these interventions then the diversity problem will persist.
Resources
You can download the table here.
Please see some of my other pages about how Outreachy fell into disrepute.