The first is a recruitment/sourcing problem. There are many demonstrable factors, such as educational availability or social norms/pressure, that serve primarily to influence entrance rates into tech among different groups.
The second problem is a retention problem. Other social and cultural traits, like the aforementioned willingness to "sacrifice for the cause", place work above family, and so on, are also not evenly distributed among different [age|race|gender|socioeconomic] demographics; I think these traits (or lack thereof) are more strongly correlated with leaving (or being forced out) of tech. To some extent, these might also serve as cooling effects for entrance as well, but I'd hesitate to make any claims about the strength of that effect.
Perhaps by explicitly addressing these two issues independently, rather than by evaluating a "where are we at this current moment statistically" snapshot, this issue can be tackled more effectively?
I think many current approaches fall into the trap of evaluating the snapshot, effectively saying "How has it come to this?!", and then trying to treat the symptom, rather than the cause.
Source: I lived through it and started with almost no experience and was hired and trained on the job. They hired a ton of people to be trained, and many didn't succeed and were transferred to non dev jobs.
> The willingness to slog though abusive working conditions is one of the most highly-selected-for trait in tech.
The comment claims that "abuse tolerance" is one of the most highly-selected-for traits, not the only thing you need to become a programmer. Further, programming skill isn't even necessarily a "trait" - the colloquial meaning of "trait" is often a personality trait, not any possible characteristic of a person.
I don't see how the comment is substantively different from the many others on this topic.
It always shocks me when people silently do that to undermine someone's reply; it seems so blatantly dishonest. Do we need to make comments non-editable once they have replies?
Even then, I don't think your reply is correct in dismissing the claim. A charitable interpretation would be that the poster meant personality trait, not including programming skill. Even a very strict interpretation still leaves room for the poster's assertion, via this hypothetical:
- totality of traits used when determining programmer quality is 100%
- abuse tolerance is the most highly selected, at 10%
- programming skill is quantified by 90 different traits at 1% each
This would mean that people without programming skill would not become programmers, but the top trait would still be abuse tolerance.
Anyway, I feel this is getting highly nitpicky at this point, I don't feel the poster was trying to make a statistical claim, but rather just emphasize that they believe "abuse tolerance" is very important for programmers, which I don't find to be facially unsubstantive or "flamebait".
(disclaimer: I made several edits to this post as I was writing it out)
The sibling essentially sees what was going on here, I was thinking of a closely-weighted linear classfier where a small change to the weights could alter their ordering but do little to change the results.
Although it's definitely my responsibility to write clearly, in the future I'll think more carefully about what my replies are actually disagreeing with before I make edits that I think are just phrasing!
(And yes, I also had a few edits here before I found a way to write it clearly.)