Nick and I met at Drupalcon New Orleans to discuss doing a survey to get some data about the diversity within the community of people who make the internet. Many such surveys have been done (here two: Slack and New Relic) but we wanted approach everything differently. Both of us have a long history of open source involvement and deeply believe in open-source philosophies. We wanted to do a survey completely in the open, and completely transparently. We have been guided by the philosophy of “Nothing about us, without us.”
Following that logic, we wanted to draft the survey with the community, making all decisions about its structure and content in public, and then open source the data we obtain and any insights we can extract. We started a repo and began work on the project in June.
The problem(s) with other diversity surveys
It’s frustrating to complete a survey and then wait, and have the numbers crunched by people who I don’t know, delivering results that may or may not be, really, accurate to the data. Their interpretations may or may not be ‘correct’ and may be biased in any number of ways, but there’s no way to confirm that because the raw data was never shared.
This project gets around that by saying - everything is public from the start. We’re writing the survey questions in public, we’re going to be making the raw data public, and when conclusions are drawn, those will be done in public as well.
I jumped in and expected (naively, I realize in retrospect) that everyone I spoke to about this would be just as excited.
Challenge #1: Involvement
I gave a talk at DCNOLA about diversity in Drupal, and followed that up with a well-attended BOF to brainstorm strategies we can implement surrounding diversity in our communities (both the Big Drupal community, and our smaller individual communities). The main take away, the majority of feedback I received was “let’s just please continue the conversation. This is an important conversation”.
So, since DCNOLA, I’ve hosted a diversity and inclusion chat in IRC and Slack (meeting notes and schedule). These have been adequately attended, but not to the level of interest that existed at Drupalcon. It’s possible, and likely, that attending these is not a value-add for a lot of people, and it’s difficult to take time away from paying work to have these conversations (part of the problem with recruitment and retention in open source communities to begin with). Getting the voices of a diverse group of people talking about the survey is an additional constraint on the project.
But the question remains: If people think this is so important, why are we having such problems getting people involved? How can we make participation meaningful and worth people taking time away from their jobs, children, partners, hobbies, etc.
Part of the idea of “nothing about us without us” is that if there were a survey addressing dimensions of identity in technology workers, technology workers would be active participants in creating that survey. I want to make sure that people’s voices are heard when the determination about what diversity indicators/measures/metrics are being used. Left to my/our own devices, surely something is going to be overlooked.
Challenge #2: The MVP of Identity
If you asked 100 people what parts of them make them unique, or different from others, you’d certainly get 100 different answers, but there would be overlaps in those sets. The big 8 is an attempt to codify the main categories/buckets of identities that people use to self-identify.
However, they are incomplete. For some people, religion is a huge part of their identity; for others, being nuero a-typical might be very important; for others, it might be their family’s working class socio-economic status.
So, we’re using the Big 8 as a starting place, but it’s here again that we need help from a wide variety of voices to make sure that we’re capturing as inclusive as possible a set of characteristics.
Another factor to consider is whether or not seeing questions about dimensions of identity with which people do not resonate would be a barrier to completing a quiz. How can we capture the minimum viable identity dimensions (and this would include the smallest data set that would also be useful for data analysis) without excluding people?
Challenge #3: Measuring for Inclusion
However, while we can measure for diversity (perfectly or imperfectly), it’s much more difficult to quantitatively measure for equality and inclusivity. A workplace can be diverse by the numbers, but still feel exclusionary and hostile to those not in the minority.
How can we get data about whether or not people outside of the majority identity dimensions feel welcomed in their workplaces? How can we measure whether or not they feel as though they’re being equitably compensated for their work and effort?
One could conclude that greater diversity will naturally begin to create a culture of inclusivity and equality, but studies have shown that this is not necessarily the case.
Going through this, we’ve learned (no surprise) that working on diversity and inclusion issues involves working through these and other challenges. We’re not sure what outcomes we’ll drive or change we’ll effect through this survey, but we each believe this topic is important to us, the teams we work on, and the communities we are a part of. Our aim is to increase awareness via both the survey itself, and the open source results. Stay tuned, and keep the discussion going with us at @nstielau and @drnikki.