> this is a terrible and naive way to think of DEI
No, it isn’t. OP is describing the intended use of DEI. A proper merit based approach would have 0 knowledge of race, gender, etc. DSA as well requires prior knowledge of irrelevant attributes of a candidate because its goal is to increase the representation of people with those attributes, not hire the best candidates.
Any process trying to select for the best developer does not need to know about race, gender, etc. Any process that asks for those is not looking for the best developers.
Hmm. Without actively taking steps to widen the net and account for those irrelevant attributes, those attributes instead act as the first filter in your hiring process. IMO the process that ignores this is the one that’s not looking for the best developers. In order to make those attributes truly irrelevant we have to correct for the massive, obvious influence they have in human decision making (on both sides of the hiring equation). Accepting the status quo is actually letting those attributes rob you of good hires.
" In order to make those attributes truly irrelevant we have to correct for the massive, obvious influence they have in human decision making"
Which is exactly what blind-hiring / blind-auditions are meant to do. The problem, from my perspective having tried to implement it, is that it requires a real cognitive leap from people who are most successful for their social acumen and don't regularly make decisions based on hard evidence. Often those kinds of people have a title with Manager or Officer in it.
What’s the correct breakdown of software developers by race / gender / etc? I think if we can answer that question we can then accurately assess whether or not those attributes are acting as filters. However, without knowing those breakdowns, taking any action to correct those breakdowns is wrong. We can’t accurately say there is a problem.
The best we can do is remove any bias we find at an individual level (ie a blind hiring process). Looking at aggregate level metrics is nearly useless because we cannot accurately account for all the factors contributing to them.
> However, without knowing those breakdowns, taking any action to correct those breakdowns is wrong. We can’t accurately say there is a problem.
You’re arguing a different point. I’m talking about the filter on candidates, not eventual hires. Though of course, people who don’t become candidates will never become hires so there’s a relationship there.
It would be easy to see whether doing X to increase candidate diversity (opening up the filter) led to more diverse hires. Especially combined with blind hiring process or whatever at the individual level. But the individual work alone is not worth much if the candidate pool itself is too limited.
If changing the candidate pool leads to more diverse candidates, and then employees as a whole end up more diverse, it would follow that some of those people brought in at the candidate level out-competed their peers & that the overall standard of hires has increased, or at least is no different.
This is not about a target ratio of employed people. Though I think observations about such ratios compared to the general population and to other similar companies can be informative.
To get at the point you thought I was making though… How I feel about quotas is: they might be an ok temporary measure, especially if managed gently, not “must be a woman” for X role, cause yikes. Hiring is capricious a lot of the time anyway, especially at the margin when there is more than one acceptable candidate and all have different backgrounds and experience. All sorts of stuff will come into play. Good people will still be in demand regardless of their race. But if you lose a couple things because you aren’t from an underrepresented group and another acceptable candidate was, who actually cares. If it wasn’t that, it would be some other nonsense like the guy was the same race as you but supports the same football team as the interviewer, or your interview took place before lunch and the other person’s was after.
Since it’s already capricious and full of coin flips, I don’t care at all that sometimes it’s capricious for a relatively good reason like giving somebody else a shot. Fine.
I disagree completely that we have to be 100% certain about a problem before we take action. We just have to know that, more likely than not, there is a problem, and take measured, cautious steps to address it. To me it would be an extraordinary, mind-boggling, coincidence if it turned out that the status quo was working perfectly to get the best candidates hired.
No, it isn’t. OP is describing the intended use of DEI. A proper merit based approach would have 0 knowledge of race, gender, etc. DSA as well requires prior knowledge of irrelevant attributes of a candidate because its goal is to increase the representation of people with those attributes, not hire the best candidates.
Any process trying to select for the best developer does not need to know about race, gender, etc. Any process that asks for those is not looking for the best developers.