Comment by brnt | Hacker News Reader

Comment by brnt | Hacker News Reader

Parent Full thread

> I strongly disagree with equality of outcome,

There's more to hiring than numerics, there's always the prerogative and perhaps even need to take a chance. People have not always had similar chances or abilities to produce the results you'd like to see in all candidates, but it just may be you can give a person a place to shine. The candidate that already shone may be safer, but in practice it may as well mean more of the same (that you've hired). Since you're using the same metrics (paper output etc. in your case). Sometimes more of the same is not what you need.

There's a whole universe of reasons for (not) hiring someone. Pretending anyone is hiring (or even able to) based on merit is not seeing clearly. You rarely know, apart from superstar hires for identical roles they've performed elsewhere already, is all about taking a chance. You may as well freshen up the org with atypical hires, in my experience worth their shots far more than 'default' hires, which, in my view, a nearly always just hires most similar to you.

I don't like to push metrics like paper output or number of recitations, because I believe many understand the obvious shortcommings of that. But the alternative you are suggesting is pure favoritism, which is even worse.

It's not favoritism. It's making an educated guess, educated by numerics, personal fit, but also stuff like how many chances did this person have and am I (as the hirer) able/inclined to provide one, how did the person handle adversity (may be one of the most valuable things you get into your org), and so on.

Using your estimate of what a person might do/become given a chance is not at all the same as favoritism. Favoritism is hiring people just like yourself and calling it meritocracy.

This strikes me as switching out one type of bias for another, with no proof of improvement as to the end result. As long as you're not breaking labor law, not a problem, but without stats showing outcome, you can't say it's objectively better.

Proof is hard because hiring is not replicable. However, doing nothing provides proof for the status quo, and if you are happy with that, by all means do not change.

I am not, and will advocate for the contrary.

Honestly I'm having a hard time following your position. It sounds like you're making some kind of circular argument; that we should hire more women because they face discrimination. But the whole debate is about whether or not they face discrimination in the first place. Or maybe you're advocating for discrimination in general on the basis that a person's immutable characteristics informs about his/her character (their work ethic, resilience, etc) or their ideas (e.g., the black candidate will bring in new ideas on account of his race compared to the white candidate) with some degree of reliability?

I think clarity would improve your posts.

Of course metrics are not everything, as I stated in my original comment, but it should have nothing to do with gender.

Maybe you want a better cultural fit, and that's 100% acceptable by my book. But leave gender out of it.

The quickest way to get a cultural shift in a technical office is to go from 98% male to 70% male. Things get real different real quick. There's no reason to leave gender out of culture shifts, except that you don't like it.

> There's no reason to leave gender out of culture shifts, except that you don't like it.

Or that it's illegal in many jurisdictions to select a candidate because of their gender. Can't speak for others, but if I see evidence of this happening in writing, I have no qualms about forwarding the docs/emails/whatever off to the Dept of Labor and a labor attorney to pursue this discriminatory behavior (discrimination is discrimination, whether "anti" a specific gender or "pro") by an employer.

Culture shifts should be a byproduct of effective, non-discriminatory hiring practices (which includes blinding to protected classes).

> Culture shifts should be a byproduct of effective, non-discriminatory hiring practices (which includes blinding to protected classes).

Good idea, but nearly nobody does it (removing socioeconomic indicators such as name/gender/address from CVs). It is also difficult to insulate yourself from your network when hiring, as well as controlling for other items on a CV (you want Ivy League? Hello bias!).

Which is why I think some active thought on the subject of diversity is essential when hiring. You will never remove bias completely, so we might as well acknowledge the 800 pound gorilla and deal with it directly.

I would really like any sort of supporting evidence for this claim. I imagine churning 28% of your employee base regardless of gender would produce a culture shift--probably a very negative one. Even more negative if you're targeting people based on their immutable characteristics--and that's before the (rightful) discrimination lawsuits start pouring in.

Maybe start with free catered lunches on Fridays or something.

It is you who obsesses over gender I'm afraid, not me. What I am saying is that as the hirer, you are able to give someone a chance. Therefore, you can look at the potential of your candidates, rather than accomplishments. You can also look at how much you can afford a wrong hire, because it is a risk, sure.

Knowing that people from group X (be they women, men, minorities, people from poor background without access to networks) on average are going to need bit more help than people from the right background, the right education, the right everything, that is something you can use in order to a) be the change you want to see b) run the risk of fresh ideas in your org. Even if you do not feel inclined towards a), b) is needed in all dynamic companies, and certainly institutes such as CERN.

> It is you who obsesses over gender I'm afraid, not me. What I am saying is that as the hirer, you are able to give someone a chance. Therefore, you can look at the potential of your candidates, rather than accomplishments. You can also look at how much you can afford a wrong hire, because it is a risk, sure.

It was ambiguous whether or not you were advocating for using gender as a proxy for "better culture fit". The GP was just being explicit that that's a bad idea.

> Knowing that people from group X (be they women, men, minorities, people from poor background without access to networks) on average are going to need bit more help than people from the right background, the right education, the right everything, that is something you can use in order to a) be the change you want to see b) run the risk of fresh ideas in your org.

It seems like you're actually arguing for using gender, race, etc as a proxy for "better culture fit" or "fresh ideas". This is illegal in most civilized places as it's immoral and based on dubious assumptions (notably that there is more inter-group intellectual diversity than intra-group diversity, i.e., all men (or women, or whites, or asians, or etc) are the roughly same).

Most countries here in Europe have not made positive discrimination illegal.

I'm arguing for understanding that most hirers simply hire people with similar backgrounds as their own. That creates the barrier I think we are discussing here. Using indicators such as gender (if your type of org has an extreme disbalance) may help you see that, and correct for it, if you so chose.

'Culture fit' is actually one of those criteria I hate: it usually means hiring similar people.

> 'Culture fit' is actually one of those criteria I hate: it usually means hiring similar people.

Yeah, I agree that it seems to imply something like that; I meant it in the sense of "hiring people who will move the culture in the direction you'd like it to move", which I think is what we're discussing, but I'm happy to use any such term.

> Most countries here in Europe have not made positive discrimination illegal.

"positive discrimination" is one of the terms I hate. It seems to imply that it is correcting for something when part of the debate is whether or not there is anything to correct for.

> I'm arguing for understanding that most hirers simply hire people with similar backgrounds as their own

I'm happy to accept this understanding, but not blindly. I'd like to know that it's based on robust evidence and at least the obvious counterpoints have been addressed. If the understanding is correct, it should survive rigorous debate and analysis; however, many proponents of this conclusion/theory seem to fervently oppose any such debate, analysis, etc. Hopefully we can at least agree that it would be reasonable to be skeptical of this theory/conclusion if indeed its proponents are trying to stifle debate/analysis even if we don't agree that the stifling is real.

And if you don't agree that the stifling is real, do you feel like there is rigorous debate and analysis on the topic? Do you see skeptics of the theory raising interesting or legitimate questions/issues? If so, do you believe those issues are being adequately addressed (and how so / give examples?)? If not, what issues do you perceive them as raising and why are those issues not legitimate? Can you provide a steelman argument?

Reply on news.ycombinator.com