[rcoveson]

I think the question is whether or not the discriminating factor was age. Mass firing requires establishing some sort of discriminating criteria (unless you just fire at random), and some of those criteria might be correlated with age. That doesn't imply bias or discrimination against age. If the most heavily weighted criteria established for firing is based on salary, and salary is correlated with age, then people in some age groups may well be more likely to be fired than others.

[mirimir]

Right, some of the discriminating factors might just have correlated with age. If there are two workers with comparable skills for tasks at hand, it makes sense to fire the more expensive one. Or if there are too many workers with skills that aren't so relevant anymore, it makes sense to fire the more expensive ones.

It's worth noting that for at least some firms, there was traditionally age discrimination that favored high-seniority workers. We used to call it "loyalty". But there was also the reality that they knew important stuff, and played a role in training new workers. But with technology changing so fast, that's arguably less relevant.

[nitrogen]

Technology changes fast, but technology concepts do not. More experienced engineers can recognize when a new technology is repeating an old pattern, and (if they'll listen) keep their team from repeating old mistakes.

[mirimir]

That's an excellent point.

But the problem may be that HR just processes nominal data, without considering nuances like that. So they end up firing experienced engineers, based on superficial (and perhaps outdated) data.

I'm not a software engineer, or familiar with relevant business practices. I get that physicians, for example, are certified for various specialties. But is there a formalized system like that for software engineers?

For some stuff, I know there is, based on ads I've seen. But what about particular programming languages, toolkits, etc?

Edit: coherence

[Fnoord]

Performance data never shows the complete picture. It should not be allowed to be used for definitive decisions, but it could be used to investigate.

(IMO the same is true for mass surveillance vs targeted surveillance.)

[TheOtherHobbes]

It appears to make sense. It doesn't necessarily make sense, because it assumes you're accurately able to gauge productivity.

This implies you have some reliable and realistic metrics for code quality, and/or that all workers in a team are functionally identical.

Neither are likely to be true in practice. This doesn't mean older = better, or older = worse, or older = identical. It means you need to assess the value of individuals in a team individually.

Otherwise you're just doing scorched earth HR, with predictable consequences.

[mirimir]

I agree. I wasn't arguing for the validity of those approaches. Just that they could account for older workers being fired disproportionately, with no "intentional" age discrimination.

Which isn't to say that HR doesn't realize that their approach targets older workers. Just that they can say that it wasn't intentional.