Nothing good can come from sticking my neck out on this, but:
If I were paying for advertising for a software engineering position I'd get much better ROI by excluding women and older demographics simply because they're much less likely to be suitable for the job, statistically speaking, when we're talking about the wider population in general and not just software engineers. Not because they're less capable, but because there's less of them as a percentage of the population. I shouldn't have the government mandating that I can't tune my advertising campaign for ROI, which is basically the only point of running one if the first place.
However, on the other side of the coin that's pretty unfair because clearly there are good software engineers in those demographics and I'd be excluding them to save money.
I think both alternatives suck in different ways, but as a business I'd probably choose ROI over fair. It's all pretty hypothetical though, because Facebook ads are probably one of the worst ways I can think of to find candidates.
> “It’s not cost efficient” isn’t a valid reason to discriminate against people, if you ask me.
Moreover, it's not, in general, a legal reason in the US to discriminate where a protected class is involved, including, in employment, age over 40 or gender.
Well, I mean, if you'd rather a political than a mechanics-of-law explanation as to why the law treats two situations differently mo, it's legal because heterosexual couples plus women outside of such couples combine to form a powerful voting bloc, and gender rating, compared to it's absence with no other change in practice, is a benefit to that group.
Specifically, compared to perfect allocation of costs, it's a small subsidy from people outside of opposite-sex couples to opposite-sex couples, whereas not having gender rating with otherwise status quo policies is a large subsidy from women not in opposite-sex couples to men not in opposite-sex couples.
Under status quo policies, opposite sex couples are in practice fairly free to gender-optimize insurance pricing by putting the less expensive to insure driver as primary on the more expensive to insure vehicle, regardless of actual primary driver.
I worked for a car insurance company for ten years and the management would answer, “Because loss history shows that we pay more in claims for insured men.”
So that’s ok but for some reason if it were a particular race they wouldn’t think to charge people
More by race so honestly I don’t know how to answer your question in a logically consistent manner.
But no one said it is was ethical in the case where it is legal, so if that was the intended question, not only is it phrased wrong but it starts from an unsupported premise, to wit, that people in the discussion believe that one is ethical and the other not.
But of course, there could be laws that prohibit linking vehicle insurance cost to gender.
My solution, when I've been married, has been having my wife get insurance, and add me as a driver. But that doesn't matter so much now that I'm older.
Edit: I do agree that discrimination by gender is fundamentally unethical.
Because there is no specific law that bans charging men and women different prices for car insurance. There are specific laws regarding discriminatory advertising with employment and housing.
What if the bias was learned by a neural network or other algorithm (eg historical insurance costs, page rank, ...)? Does that somehow make it legal if they make an inference about a protected class and stop advertising to them?
That kind of thing is why the doctrine of disparate impact was developed. There are a lot of messy edge cases (everything has some correlation with one or another protected class), but if I end up never advertising to men the fact that I'm using a neural network won't protect me.
> That kind of thing is why the doctrine of disparate impact was developed.
It's really not; disparate impact deals with it, but the doctrine was articulated in a case with ample evidence that the policy at issue, though facially neutral, was adopted as a deliberate replacement for explicit racial discrimination with an intent to maintain it's discriminatory effect and without other business purpose, and the doctrine was clearly intended principally to address such veiled-substitute policies.
The doctrine was first articulated in a case (Griggs v. Duke Power Co.) where the Supreme Court explicitly held the employer didn't intend to discriminate.
> the Supreme Court explicitly held the employer didn't intend to discriminate.
The Supreme Court did not hold that, it merely noted that the lower courts had held that, and noted that a lack of discriminatory intent was “suggested” by the Company's efforts to help undereducated employees by financing a major portion of remedial education (part, but not all, of the policy held to be disparate impact was adoption of high school diploma requirements for positions for which such requirements has no substantial relation to job performance.)
The Supreme Court explicitly did not consider whether there was intent to discriminate, because it was not necessary to do so to resolve the case.
So then are algorithmic approaches to advertising for employment off limits? Facebook and google then really shouldn’t go near this kind of work since their tech is legally inappropriate.
> So then are algorithmic approaches to advertising for employment off limits?
Not if what the algorithm optimizes is a measure provably closely linked to job performance.
In fact, as the array of protected classes and the statistical means available will probably over time allow proving prohibited disparate impact for any hiring method other than an algorithmic method optimizing a true job performance measure, algorithmic methods may soon be practically mandatory.
OTOH, algorithmic methods targeting any figure of merit that isn't provably tightly linked to job performance are going to be problematic, sure.
I don't know how it works with Google, but the article says Facebook is building a new ad console specifically for the kinds of advertising where discrimination is illegal. Presumably they'll make sure ads run through it aren't producing any disparate impact. (But note that this wasn't the problem with the cases in the article; in those cases, advertisers were explicitly pressing a "don't send this to women" button.)
Women software engineers are a much smaller percentage of women than male software engineers are of men. That's pretty obvious to anyone who's ever worked in software.
However that is in part because employers have been historically hesitant to employ them (and clearly if they are not advising to women then the will not be getting employment) this creates a self supporting feedback loop.
You say that as if it is fact. Do you have any evidence to support it? I always assumed that it was mainly down to more males taking an interest in the subject. In my career in IT women have always been seen as equals where I have worked. There just aren't many of them.
The wording was unclear to me at first because it wasn’t explicit.
Anyway, I completely changed my comment to more closely go along with the point I want to be making.
Regardless, besides being unfair the idea that the policy is economically justified doesn’t seem to be immediately intuitive to me. It seems like you should advertise on subject matter for that kind of thing, e.g. people who watch EEVBlog would get an advertisement for electrical engineering jobs.
I think the parent comment's logic is sound from a financial point of view, but you do a wonderful job of reframing it to bring it back to an equal rights point of view. Women and older workers have just as much right to the same advertised opportunities as others. Sticking a job ad in the mens restroom or within an age-restricted night club when you're looking for accountants seems misguided, but it's easy to see how that wasn't considered.
Yes, that's true. It seems to be illegal, so talking about ROI is really beside the point. And it's a pretty hypothetical situation. I'm just trying to think why the companies might have taken this approach apart from outright discrimination.
What, in the stretch of American labor history, compels you to think that individual companies need or have ever needed more than "outright discrimination" to do something that is not in their economic best interests?
> “It’s not cost efficient” isn’t a valid reason to discriminate against people, if you ask me.
It's literally the perfect reason to do discrimination. It is just like discriminating for having work experience, education, demonstrably market skills, etc.
We have decided as nations and societies that some forms of economic discrimination are too damaging to permit. The details vary, but typically the things we can't easily change: race, nationality, age etc. Such discrimination is generally outlawed.
Someone lacking relevant market skills or education etc can go out and fix that.
> The details vary, but typically the things we can't easily change: race, nationality, age etc. Such discrimination is generally outlawed.
Also generally perpretrated, by the same body that bans it.
A great beauty about human beings is that they have a natural tendency to find what is best for them and their communities in spite of legalities that attempt to distort reality.
In any case, what you are ascribing as unlawful discrimination is absolutely lawful and you would be naive to believe that there were castings for white asian geriatric women for the role of Black Panther to abide a law about the positively-discriminated groups you mention.
>Someone lacking relevant market skills or education etc can go out and fix that.
Not really, if they lack economic resources, have dependents, have low IQ, low grit, or lots of other reasons.
Ultimately it comes down to what parts of a person you think they are 'responsible for' and what parts they aren't. You're saying that someone being a certain skin color isn't something they should lose opportunities for, but it's okay if they do so for being too dumb to complete and advanced education, even if they were born that way.
My point isn't that you're right or wrong about either of these, but that if you really dig into this you'll find that any such narrative supporting an egalitarian notion that 'people ought to be equal' collapses with a few simple questions about genetics, free will, parenting, intelligence, and other such topics. Nobody gets an equal shot at anything.
You should read the Equal Employment Opportunity Act. None of the things you listed are examples of discrimination in the context
of civil rights. The Equal Employment Opportunity Act.
Perhaps I should also remind you that this isn’t 4chan.
You seem to think the law is the be-all end-all definition of perfect moral reasoning.
The law is not God's word. It's open to criticism, and we're criticizing it. If you want to participate in the discussion, great. But, "It's illegal because the law says you're not allowed to do it," is not a meaningful contribution.
The discussion point was already made above, in that we make certain things illegal because even though it may make economic sense, it results in massive damage to things such as the environment, vulnerable populations and so forth. The poster chose to respond by saying it was a perfectly valid reason to discriminate, which goes against the law for the already stated reasons.
There was no additional criticism offered here. No one has made the point that 'it's the law therefore it's good', and you're massively misrepresenting the arguments here.
And in fact, there are laws against pollution because it's an externality, it doesn't make economic sense to the whole of society. Markets are good at maximising the total good in the case where costs and benefits can be totally internalised. Where you have externalities you see either overprovisioning or underprovisioning and in net, everyone is worse off than the rational outcome. That's where intervention is better than a market by itself.
This is why I bemoan the current form of identity politics because it allows an out. "I don't hate women, it's just disadvantageous to me to hire them," or, "I don't hate black people, I just won't rent to them because they statistically commit more crimes." And without an iota of malice in anyone's heart, relevant actors can reproduce the same exact disparities that perpetuate inequality. This is called systemic discrimination.
I know the example you're offering is a step removed from that ("I want to tune my ads to maximize ROI, but if a woman sends me a resume and is a good fit, I won't discriminate") but the problem with disparities is no individual actor is incentivized to buck the trend, which assists in perpetuating it. That's at least the logic why the government steps in and changes the incentives and why such intervention is warranted.
> And without an iota of malice in anyone's heart, relevant actors can reproduce the same exact disparities that perpetuate inequality. This is called systemic discrimination.
That’s not systematic discrimination. It’s express discrimination and illegal.
Systemic discrimination results from policies that are facially sex or race neutral, but have a discriminatory impact because of structural discrimination in society. (E.g. not hiring people for having marijuana convictions may be an instance of systemic discrimination if black people are convicted far more often for possession despite using drugs at similar rates to white people.)
> This is why I bemoan the current form of identity politics because it allows an out. "I don't hate women, it's just disadvantageous to me to hire them," or, "I don't hate black people, I just won't rent to them because they statistically commit more crimes."
That's direct discrimination, and I don't see how “the current form of identity politics” supposedly creates an “out” for it. Please elaborate.
Since the fading of the New Left, progressivism in general has increasingly turned away from their traditional causes which were rooted in economic concerns and more towards different kinds of social causes. Identity politics isn't the only thing, but it is a primary focus of the left today. Identitarianism was prominent in the New Left, but the one thing that exemplifies the left's use of it today is a decided move away from the systemic to the personal, concerning personal attitudes and how they effect inequalities in society, vs systems of oppression. It's not that the personal doesn't matter, but it is part of a whole; both society and the personal matter[0].
I want to use the N-word (neoliberalism) but that term is a little overused. I guess the better way to express it is a kind of recuperation of identity based issues to fit into the dominant individualist ideology in society (the N word fits here but avoiding it) like the personalization of sexism, racism, turning them from being systems of oppression to being mere personal attitudes.
> Since the fading of the New Left, progressivism in general has increasingly turned away from their traditional causes which were rooted in economic concerns and more towards different kinds of social causes.
I don't think that's really true, I think progressivism just faded itself for a while (it's starting to come back again).
> Identity politics isn't the only thing, but it is a primary focus of the left today.
Identity politics has always been a primary focus of the Left; class consciousness is identity politics. OTOH, class consciousness has spread beyond the left; Neoliberal identity politics is a (deliberate) bourgeois distraction from proletarian class consciousness, and is a tool primarily of the center-right. Progressive identity politics is intersectional, integrates other identity axes with the economic, remains deeply concerned with systems of oppression (economic, patriarchical, racial, and other). Both exist, as do farther-right identity politics.
There's perhaps some confusion in American politics because both the neoliberal center-right and progressives are found as wings of the Democratic Party, the former dominant since the early 1990s but decreasingly so in the last few years.
> the personalization of sexism, racism, turning them from being systems of oppression to being mere personal attitudes.
Sexism, racism, etc., have always been personal attitudes (but not “mere” personal attitudes) that were of particular concern because of their capacity (and reality) of grounding oppressive insitutions and systems. It's true that as a misguided (easy, but ultimately self-limiting in many respects, including where credibility is involved) attempt to deal with certain deflections, some in the racial civil rights community attempted to depersonalize racism and claim that the term only could be applied with a dominant social system of oppression, but that was a reaction, not the original understanding, progressive or otherwise.
My Aunt wrote computer code into her 60s. This was up until at most 10 years ago. She has a degree in math and went back to school after being a SAHM and got a degree in computer science from a community college, and banged out code for about 15 years.
She coded circles around (and out-adulted) her younger colleagues.
Good on her, and whenever I see/hear folks say that hiring younger workers is cheaper overall for a company, I very simply wonder if they have ever actually worked for a good experienced older worker.
An older worker can get things done in an hour that can take younger workers a day (or even longer!), and thats not even including having someone come in and fix issues with the younger person's code. The older person can mentor younger people, and will ensure that the younger workforce will turn into better workers.
On top of that you hit on the other thing, they are often more mature, have a better work life balance, and know what they want. This tends to mean less personnel issues, the person is much less likely to burn out, and the person will stay there longer.
I would happily pay 3-4x for that more seasoned person for the quality work alone, but there are so many reasons to have older, more experienced people in your company.
She's not a boastful person and doesn't like talking about work, and we don't have much in common. But in an effort to bridge the gap, she would tell me about work. The few conversations we had about her job made me angry on her behalf for how much of the workload, and more importantly management responsibilities, she shouldered on behalf of the company for no increase in title or pay.
It was a small, family owned company that made medical software that was a little "fly by the seat of the pants" if you get my drift.
For the removal of doubt, I have worked as a computer programmer.
If we examine Fair Credit and Reporting Act (FCRA), we see that it is not ok to discriminate against protected groups like race even if that means you're taking on more financial risk. Let's assume the data indicates that certain minorities are much more likely to default on a loan. You still cannot use their race to deny them a loan.
The reason is that it would perpetuate inequality. Given that minorities were systemically discriminated against previously by red lining by the federal government and could not get a federally backed mortgage in for decades (which resulted in Urban ghettos, white flight, etc). And given that those injustices have not been properly corrected, discriminating based upon race must be forbidden in the credit markets.
But you can discriminate using other variables. For example, you can discriminate based on education, income, credit history, credit score, etc. And some of those variables might correlate heavily with race.
In that scenario, you can use FCRA compliant variables even if they correlate with race as those variables are (a) cleared by the government, (b) derived from actual facts about the individual (as opposed to predictions). Otherwise, if you in your underwriting machine learning model accidentally find a proxy variable for race that isn't an FCRA compliant variable, your compliance department will require you to not use that proxy variable in your underwriting model.
Your argument "I'm taking on an extra financial burden" doesn't excuse you from ensuring fair access to credit. That's akin to saying "the cost of being handicap accessible is unfair to my business as it costs us money so I don't want to do it". We have made certain rules to ensure people have a fair shot in life and those inherently cause an increased cost to businesses.
Now, there is actually more nuance to this. For example, if you provide a credit product but you only make that product available in certain zipcodes, you could actually run afoul of fair credit laws if the zipcodes you include versus exclude are heavily correlated with race. However, if you are a very small business and don't have the resources to provide your services in all markets / zipcodes, then you likely would be ok simply because having to provide your services in the many places might be an undue burden on the business. However, as you grow and become more financially stable, you would/night need to fix that.
If all/most employers exclude workers based on demographics, then smart employers will realize that they can find great employees in those demographics.
A good portion of Quant trading is based off a similar principle. If I can model priors and variance better than my competitors, I can make money.
It isn’t that simple. In my local talent market, big FAANGs + Microsoft-style companies have been aggressively recruiting female SWEs, creating even starker imbalances in smaller tech companies + startups. It doesn’t seem unreasonable for those companies to completely give up on finding women to hire, because that supply is limited (at least until the educational pipeline catches up).
And it kind of makes sense: it will take a few more years for supply in the industry to actually increase (especially for senior talent), while companies like Google are under the spotlight now and no one really cares about that 10 developer payroll company that Google poached talent from. (Not to mention employment laws work different for companies of different sizes)
To put it in trading terms, if your bigger competitors are sucking up all the supply of X, then you simply don’t worry about buying X if you can get by with Y anyways.
> If I were paying for advertising for a software engineering position I'd get much better ROI by excluding women and older demographics simply because they're much less likely to be suitable for the job, statistically speaking, when we're talking about the wider population in general and not just software engineers.
I'll go the other way. I can tell you, specifically, that if I want embedded software engineers, generally older is better.
It takes a lot of experience to wrangle software architecture, C, crappy development tools, communication stacks, PCB boards with bugs, microprocessor architectures, real-time operating systems, oscilloscope traces, Wireshark traces of CANBUS, etc. into a functional system
I have yet to meet someone with less than 15 years of experience who can pull that off without me standing over their shoulder telling them what to do.
Not everybody is building the exact same VC funded, fad driven, social malware website over and over and over.
(your assertion is that ad views in some demographics will so outweigh useful impressions that it is worthwhile to exclude those demographics. But ad views don't cost much!)
If you can target software engineer on Facebook, then clearly you'd just do that.
If you're targeting more broadly for some stupid reason, then there's an economic incentive to target unfairly by demographic, just to narrow down the sheer number of irrelevant ad views by any means possible.
I doubt that you could do that on Facebook. But on LinkedIn, you could obviously target people who claim experience in software engineering.
As you say, that might well effectively select for younger men. And as far as I know, that wouldn't be illegal. But you can't target by age, gender or disability. Because it's explicitly illegal.
> I think both alternatives suck in different ways, but as a business I'd probably choose ROI over fair.
Understandable. And probably even common. That's why society should be adding disincentives--and that the people running most businesses do not suffer clawing panic attacks at the disincentivizing retribution appropriate for this behavior is indicative that we do not make businesses fear bad behavior nearly enough in this country. Nor, really, do we nearly often enough puncture the fictive-person boundary to reach the decision-makers themselves.
I say this a lot around here and it never stops being true: society lets the corporation exist. There's a grant of power and a grant of trust inherent in a corporate charter and the advantages therein. Every member and executive of a corporation should fear the everloving daylights out of betraying that public trust. And yet at this point we almost take for granted that betraying the public trust is what corporations do.
If I were paying for advertising for a software engineering position I'd get much better ROI by excluding women and older demographics simply because they're much less likely to be suitable for the job, statistically speaking, when we're talking about the wider population in general and not just software engineers.
I think your logic is flawed. What you are not taking into account is that when you advertise, you ask people to do a certain amount of self-selection. If you advertise that you need a Go programmer, it's not like you have to work really hard to only advertise to Go programmers for fear that you will get 1,000,000 resumés from Java programmers.
And even if you did, you can throw out all those resumés cheaply, because they don't have Go experience on them. The big costs in hiring are the cost of showing ads to people, or even the cost of sorting through resumés.
The big costs of hiring are 1. Interviewing people, and 2. The opportunity cost of not finding the right person as soon as possible.
Not advertising to women because you have a theory that there are fewer qualified women out of the entire population of women than there are qualified men out of the entire population of men is premature optimization.
If, for example, you are advertising to Hacker News people, do you really think that the statistics about women in general are relevant? Or perhaps those women who read Hacker News are not representative of women in general?
Same for if you advertise to people who have shown an interest in Go, or Elixir, or Lisp, or distributed system, or DevOps. If you were advertising for people with experience programming for space flight in 1971, there might have only been a dozen women in America with relevant experience.
What kind of logic leads you to say, "Forget it, I am not interested in hearing from Margaret Hamilton, because there are so many women who actually have experience as secretaries?"
I have said this many times before, but when you are seeking exceptional people, people who are actually very rarely available and have many opportunities to choose from, the overall statistics are irrelevant. If out of 10,000 people on Hacker News there are four men you would hire, and only one woman, this is no time too think, "my ads get four times the efficiency if I only advertise to men."
Instead, you should be thinking, "There are only .05% of the readers that I want to hire. If I advertise to both men and women, I increase the chance of finding one of those .05% by a whopping 25%!"
If hiring is easy, that may not matter. But if hiring is hard, if you might make several offers and only get one person to accept, if you might go a month before finding the ideal candidate... You want to advertise far and wide, and use other methods to filter out spammy submissions.
Advertising is cheap. Not being able to hire the right person at that magical moment when they are available is very expensive.
Let's say a lot of people in software agree with you. I think they're wrong, but that doesn't matter. Here's the thing: When we analyze hiring, we often act like it's some kind of two-person game:
You place ads, interview people, select the best person, make them an offer, and boom, you're done.
But it isn't a two-person game. It's a market, and you're actually competing with all the other hiring companies, just as the candidate is competing with all the other people you interview.
Under the circumstances, I suggest you crack open a copy of "Moneyball." The worst ROI is when you overpay for people who aren't actually great performers, but everyone is pattern-matching, and these people have the superficial attributes of a great hire.
Those people get more offers than good performers that don't have the right gender or didn't go to the right school or whatever. And it's expensive and hard to hire them.
Meanwhile, there's this woman over here, and that trans person over there, and that self-taught person over there, and so on, and so forth, and they are all cheap to hire because you have less competition for them.
Not only that, but if everyone is using hacks to save money by not advertising to them, it's easer to reach them. It costs less to advertise to them, and the qualified ones are more likely to click on your ads.
If you buy my Moneyball analogy, the highest ROI goes to those who can ferret out signals of hireability that are contrary to what the "market" uses.
You might still not want to use FB to reach qualified women, maybe it sucks for that, but you ought to ask, "How can I find those qualified women, everyone else's discrimination is my opportunity."
In which case, if you don't like the ROI on using FB to reach women, instead of thinking that advertising to women sucks, you should be brainstorming how to effectively advertise to women.
If you figure it out while everyone else ignores women, you win.
> But it isn't a two-person game. It's a market, and you're actually competing with all the other hiring companies
There's another related point here too: women and other people who form smaller groups in some context talk to each other. Have openly shitty approach to them - it will be known. Treat them with respect - you'll get recommendations in that employee market.
Your ROI will go way down when you get sued for breaking the law. There are also much better ways to target your ads (in ways that directly relate) to improve your ROI (like advertising on StackOverflow or HN).
The negligible ROI improvements you get by discriminating are not worth the costs to society.
The doctrine of disparate impact is plainly immoral and flies in the face of freedom of association. Society should punish companies which display behaviors they do not approve of using the market instead of the government.
Were we to take the second sentence seriously, we would have to conclude that the general idea of prohibition of discrimination of whatever kind in non-governmental action (whether employment or public accommodation) was immoral, not merely the specific doctrine of disparate impact.
I agree to the general idea that the government should try to do as little moral arbitration as possible. The concept of "separation of church and state" would be best modernized and secularized as the "separation of morality and state".
You can't post like this here, regardless of how wrong another comment is. In addition to contributing to destroying this place, it discredits what you're saying—which if your position is correct, means that you discredit the truth. That's a bad thing to do to the truth. So please don't comment like this on HN.
Edit: it unfortunately turns out the you've been breaking the site guidelines egregiously and repeatedly. As you know, we ban accounts for doing that. Would you mind just reviewing them and taking the spirit of the site to heart, so we don't have to ban you? It's not hard to use HN as intended if you want to.
If I were paying for advertising for a software engineering position I'd get much better ROI by excluding women and older demographics simply because they're much less likely to be suitable for the job, statistically speaking, when we're talking about the wider population in general and not just software engineers. Not because they're less capable, but because there's less of them as a percentage of the population. I shouldn't have the government mandating that I can't tune my advertising campaign for ROI, which is basically the only point of running one if the first place.
However, on the other side of the coin that's pretty unfair because clearly there are good software engineers in those demographics and I'd be excluding them to save money.
I think both alternatives suck in different ways, but as a business I'd probably choose ROI over fair. It's all pretty hypothetical though, because Facebook ads are probably one of the worst ways I can think of to find candidates.