I'm a bit mixed about this; it is clearly discriminatory to publish a job ad that says "women need not apply", but it doesn't feel inherently discriminatory to advertise a job in, say, "Men's Health" but not "Cosmopolitan" magazine. Or to advertise on a particular television channel or program (whose viewers may not match the population at large). That's just how advertising works. Is that crazy?
Would this be different if, for example, Facebook was targeting people for job ads based on "people who like action movies" or "people who like romance novels" (I'm stereotyping here, but just assume you pick a category that ends up being a decent proxy for gender)?
One thing you might be interested in reading about is Disparate Impact. Discrimination cases can be proven by showing that a practice has a disproportionately adverse impact on members of a protected class, even in the absence of intentional discrimination.
The Supreme Court noted in Griggs v. Duke Power Co that the Civil Rights Act bans "not only overt discrimination but also practices that are fair in form, but discriminatory in operation." In fact, "absence of discriminatory intent does not redeem employment procedures" that operate in a discriminatory manner. Unintentional discrimination is still discrimination.
In a piece on age-based discrimination, Facebook said, "age-based targeting for employment purposes is an accepted industry practice and for good reason: it helps employers recruit and people of all ages find work." (https://www.propublica.org/article/facebook-ads-age-discrimi...).
I think a lot of the industry has been casual about job ads, but Facebook is kinda the first platform that really lets you be exact with a lot of your targeting. Before, maybe you feign ignorance that "Men's Health" doesn't have a lot of women readers. Plus, women can read Men's Health. However, I can't get Facebook ads targeted to a different gender. Facebook lets you be reasonably exacting.
If I create a Facebook ad for a job application with no other promotion of the opening (including on my company's website) and target it just to men, is that discriminatory? IANAL, but in Griggs the court held that procedures that act as built-in headwinds for protected groups are illegal. Making it so that women could only hear about the job from a man seems like quite a headwind working against their likelihood of getting the job. Griggs is about job qualifications that were used as a proxy to keep black workers from getting jobs, but making it so that a group of potential workers can't even see the job application seems similar.
Going back to your example: if someone used a gender proxy like "people who like action movies", I think it would be hard to defend in court. 1) They don't need to prove discriminatory intent, just disparate impact. 2) It seems likely that a company adopting such targeting would have communication indicating their discriminatory intent. "Facebook won't let us target by gender anymore! What can we do?" "Let's use things like action movies to get the same result!" If that comes out in discovery, it would seem like a simple case. 3) At some point, a court is going to want to know what the basis of that targeting was and it's going to sound really hollow. But it doesn't even matter if they come up with a good cover story, because even if it was unintentional, it can still fall afoul of the law.
On the other hand, if I run a tech company and wanted to target "people you are into the Python programming language", that's a group that will likely skew male, but it kinda passes a sniff-test. It just doesn't smell funny. "Yeah, um, we want programmers and I thought targeting people into Python was reasonable." That does sound reasonable. It's not meant as a proxy for gender, but for an interest/skill that is closely related to the job. Action movies aren't related to my job, but a programming language is.
Going back to the magazines: I think much of the same might apply, but I'm guessing that we've been relatively lax about it as a society. It might not feel inherently discriminatory, but disparate impact doesn't rely on intentional discrimination. You say, "that's just how advertising works," but you could easily place ads in multiple places to reach a wide variety of people. One program might not reflect the population at large, but I don't think it's rocket science to advertise in multiple places to reach a more representative sample.
As I noted above, Facebook says that job targeting (at least on age) is an accepted industry practice. There are people that don't like the disparate impact theory (though it is established Supreme Court precedent since 1971). Plus, I don't know if disparate impact has ever been used in terms of narrowing the field before applications even come in. In Griggs, the procedure disproprotionately weeded out black applicants. In the Facebook case, those potential applicants never even knew the job existed.
I have definitely thought the same things you've thought. Society is hard. Fairness is hard. But I think it's important that we work toward fairness. This isn't meant to be anti-company. I just think companies should have a bit of a justification for doing something if it runs contrary to fairness. If you're only advertising your programming jobs to people under 40, why? If the answer is, "we mostly don't want to hire older programmers, but if a good one happens to come along we'll take them", that doesn't feel like a great justification. In fact, it basically admits discrimination. If the answer is, "there are more young programmers than old programmers", I think the retort has to be, "given the loads of ways you can target, why not target based on interests rather than age or gender? Targeting to people interested in CS, various programming languages, etc. seems both better for you and for potential applicants...unless you're trying to discriminate." It's not always simple, straight-forward, or easy, but I think we're trying to build a better, more fair society and that requires hard-work and questioning.
Disparate Impact is one of the most illiberal policies in America. Companies need to be so careful that their results can't be construed as discriminatory that they need to go out of their way to make their hiring practices discriminatory. You could have a totally blind hiring process (remove names from resumes, modulate voices in phone interviews, never get any indication of race/sex/age of applicant) and still be liable for a lawsuit if you don't hit the arbitrary "80% rule" benchmark.
The Disparate Treatment and Disparate Impact doctrines are diametrically opposed, yet both can lead to lawsuits. When a company gets big enough, it's going to start running into discrimination lawsuits no matter what they do. There are things they can do to offset the damages (diversity training, diversity department, certain types of donations), but the end result can still be settlements in the tens of millions of dollars.
You see a similar thing with real estate in equal housing. I wasn't aware of how strict the standards were for this stuff until I took the license courses last year.
If you have a real estate office and somebody walks in, you might offer them a cup of coffee. If another person walks in later and you forget to offer it to them, and that person is part of a protected class...you've just violated equal housing.
This is why a lot of real estate offices will put out drinks, snacks and coffee with a "help yourself" sign now.
Well over half of the course was basically "how not to get sued."
While I agree, I was just shocked by the amount of legal stuff around realtors that basically keeps them from being able to be helpful. It has a huge effect on the perception of the profession because people are forbidden from really distinguishing themselves out of fear that a wrong word can get them sued (not just for equal housing but tons of other stuff).
Speaking of smells, my initial reaction is that Disparate Impact smells broken. I can see pretty clearly how it would be possible to argue legally that advertising a job internally and not externally (to a company) is discriminatory. And following that further along, in this way, I can see a pretty straightforward method of declaring any job advertising as discriminatory under that decision. I think law like this is broken because it's like many (American) laws where it ends up selectively enforced, because the base of people violating the law is effectively infinite. So in that way it's not used to fight discrimination, but instead becomes yet one more way to fight the oldest battle in history: The battle to consolidate and preserve the power of existing institutions of control and weaken the influence of competition and punish outsiders.
I think law like this is broken because it's like many (American) laws where it ends up selectively enforced
Laws like these are not criminal laws, and are not meant to be enforced if by that you mean policed. It's up to someone to file a lawsuit, if they think the law is on their side. Then a judge will have to rule whether or not the law applies in that particular case, and in doing so they may set a precedent which will influence how later cases will be judged.
And following that further along, in this way, I can see a pretty straightforward method of declaring any job advertising as discriminatory under that decision
You may think so, but this is why we have judges. Just because you can make and argument, it doesn't mean that argument will be persuasive to a judge. You may further think that this means that the outcome of a trial is arbitrary, and you would be right in the sense that law is not an exact science, and the result of a verdict can be difficult to predict if the evidence is weak and/or there is no precedent. But you would also be wrong in the sense that each verdict rendered helps to set precedent, and judges (and lawyers) always look to see what precedent has been set, both when it comes to how to interpret the law, but also when it comes to establishing what counts and doesn't count as evidence.
The fact that it's not actively enforced by the state doesn't matter.
It's basically the civil equivalent of an actively enforced law nobody can possibly comply with at all times that gives the cops the power to harass whoever they want whenever they want (e.g. the 55mph speed limit on I95 in the Boston area) except instead of empowering cops to harass people on flimsy pretext they're empowering plantiffs to harass companies on flimsy pretext.
Laws and precedents (in the case of civil litigation) that well meaning parties can't possibly comply with are bad.
> Laws like these are not criminal laws, and are not meant to be enforced if by that you mean policed. It's up to someone to file a lawsuit, if they think the law is on their side. Then a judge will have to rule whether or not the law applies in that particular case, and in doing so they may set a precedent which will influence how later cases will be judged.
This is one of the reasons that the "loser pays" approach to the cost of litigation used in much of the rest of world would not work well in the US. We use private civil lawsuits to enforce things that would be enforced by government agencies in much of the rest of the world, so it is important that individuals can bring lawsuits against much bigger, much better funded entities without having to worry about getting wiped out financially if they lose.
>Laws like these are not criminal laws, and are not meant to be enforced if by that you mean policed. It's up to someone to file a lawsuit, if they think the law is on their side.
That's worse, because then it shifts the burden of en^W^W policing onto those who actually want to go through the time and expense of a lawsuit, so only the most egregious abuses against the most powerful people will ever see judicial oversight, providing an extremely noisy signal that never provides reliable guidance, esp when (as parent notes) huge classes of activity are violating the ostensible "disparate impact" criteria but never prosecuted.
> But you would also be wrong in the sense that each verdict rendered helps to set precedent,
But those verdicts don't help when the set of cases going to trial is skewed as above. The only real signal it sends is "hey, don't do it in a way that might get noticed by someone who actually matters".
> I can see pretty clearly how it would be possible to argue legally that advertising a job internally and not externally (to a company) is discriminatory.
I think the precedent is the opposite. Companies and governments in the US do this all of the time without legal risk.
> And following that further along, in this way, I can see a pretty straightforward method of declaring any job advertising as discriminatory under that decision.
Perhaps a bit of a stretch here. Companies and governments will continue to advertise jobs
> So in that way it's not used to fight discrimination, but instead becomes yet one more way to fight the oldest battle in history: The battle to consolidate and preserve the power of existing institutions of control and weaken the influence of competition and punish outsiders.
This sounds like good old competition to me. Why would an established player not weaken competition with pricing or product value?
Regarding age based discrimination, discriminating against the "young" is allowed as they are not a protected class. Allowing the refinement and targeting of ads towards older people would be okay. But restricting the ads to only younger people would not be. At least by law in the U.S.
Yup. The young are the one class in america where discrimination is not just allowed but expanding. Businesses like car and hotel rental have discriminated against them for ages. But now big retail chains are beginning to discriminate as well as social pushes are made to violate the 18-21 year old rights to purchase firearms and the like.
It's pretty crazy. They are legal adults. They vote. They serve on juries. These are the people that fight america's wars and actually risk their lives but at home large companies are discriminating against them in an rapidly expanding sphere of things.
I'm only friends with people on Facebook who know my gender, so it doesn't really matter. Plus it only actually shows in a few places such as "xxx replied to his/her comment" notifications.
Any idea how Facebook handles people with rare gender pronouns? Doesn't Canadian law require use of chosen gender pronouns and Facebook operates in Canada.
I think I've seen "their" used instead of "her" or "his" where people declined to specify any gender. This works in English but not in German for instance, where I think they write "ihren/seinen" which translates to "her/his".
I suppose that's as far as Facebook's support for gender choices goes.
if I run a tech company and wanted to target "people you are into the Python programming language", that's a group that will likely skew male...
"Yeah, um, we want programmers and I thought targeting people into Python was reasonable." That does sound reasonable. It's not meant as a proxy for gender, but for an interest/skill that is closely related to the job
Ok, I'm not deeply informed on Disparate Impact, but as you just described it, as intent doesn't matter, if your "job listing targeting Python developers" skews male and doesn't pass the 80% test, you're still discriminating & liable right?
Job requirements are generally immune from discriminatory impact cases. That is, only someone who otherwise was a fit for the role but was discriminated against has standing to sue.
> If I create a Facebook ad for a job application with no other promotion of the opening (including on my company's website) and target it just to men, is that discriminatory?
You make a lot of interesting points and its a difficult problem, for sure, but I think this is a bit of a straw man. Even for company that explicitly intended to discriminate by hiring only men for a position could avoid the vast majority of what you wrote here by posting the position on their company jobs website and just only seriously considering applicants with male names. Trivial, even.
Its also possible that a company could be using gender targeting in advertising specifically to _mitigate_ the tendency for certain roles to have gender-skewed candidates, by targeting women on Facebook for python roles because their applicants through the passive/non-targeted listing sites skew male as an example.
I can't say one way or the other if that's what Uber did, but its impossible to look at the activity on one platform and somehow extrapolate that to the entire company's hiring strategy.
Facebook ad campaigns heavily favor advertisers slicing up demographics as much as possible. I am going to make the assumption that at the very beginning of the campaign an advertiser created many segments, split by age and gender. After they do that, certain segments are going to respond better and the distribution for those ads is going to increase while the others fall. The alternative is that the advertiser is going to have to jack up the bidding for the non-responsive demographics.
(And for anyone who doesn't know much about digital advertising, despite all of the crazy complexities, all of these systems are basically trying to maximize displays toward people who are going to respond to the ad and eliminate displays to people who won't. The definition of a response at the simplest is a click, but at the most complex is a lead, app install, purchase over a certain order size, etc. In many cases even the deepest pocketed brand advertisers will not be able to out-bid a highly optimized ad campaign for specific inventory.)
On the other side of the internet advertising pie you have Google (excluding Indeed and LinkedIn because I know only a little about their platforms.) An advertiser is going to bid on "driver jobs" or whatever, and Google is going to do the rest. Way back when, before Google collected user demographics, the ad is going to show just for the people who searched it. I'm going to take a guess that some job keywords split 50/50 while others skew heavily toward certain genders or age groups (e.g. 0 females over 65 are searching for manual labor jobs.) With Google's demographic data and machine learning based targeting, what is really going on today, in 2018, is a mystery to me. I would assume that these algorithms are not demographically neutral. They probably learn that males respond more to driving ads. Almost certainly on the display side of things (which Google's less sophisticated ad buyers might not realize they are purchasing) the ads are going to start skewing toward certain demographics.
Solely based on the way Facebook's platform works, I would question if it is ok for any advertiser to place ads on Facebook's platform as Facebook can place the blame on the advertiser even though it is ultimately a consequence of how their system is designed. On Google's side, the same discrimination is likely happening but the advertiser can probably blame Google.
At the end of the day, larger questions persist. How much money should employers be spending on letting employees know that they have a paying job available to them? And, is it ok that job opportunities, which may be superior, and being buried by those that have deeper pocketed employers?
Because of Google's semi-recent launch and expansion of Google Jobs results, which almost certainly will end up heavily monetized, I think these are very important questions.
While I see both sides of the argument, as long as the advertiser pays per view, it seems wrong to require the listing to target demographics that are almost certainly uninterested.
Your example of listing a job in Men's Health but not Cosmopolitan seems good. Technically not a perfect analogue as a woman always could buy Men's Health, but in practice very similar.
Spinning off into a different angle, what if an advertiser gets the most qualified female applicants advertising through LinkedIn, and the most qualified male applicants through Facebook? (Again assuming pay per view) Is it wrong for them to target women & men on separate platforms with separate ads? Utterly hypothetical of course.
Or yet a different angle, what if they have ad "A" and ad "B", each subtly designed to excite one of the sexes. Perhaps an ad for the Marines, one emphasizing toughness & grit, the other emphasizing teamwork & "stronger together". Target one at men, one at women. Legal?
To me it seems to come down to the overall behavior of the employer, not Facebook or any one ad.
I’m very skeptical that their motive is hiring discrimination as opposed to cutting costs on ads by targeting the demographic that is most likely to respond to the ad campaign. In other words, if the ad campaign isn’t successful for recruiting women, why would they spend twice as much for no increase in ROI? Maybe they could run a different campaign that targets women, and perhaps they are (if there is something in the article that claimed these employers are only running campaigns that target men, I missed it), or perhaps no one has found a campaign that is comparably cost-effective for hiring women as is the case in STEM fields. I just don’t see what these firms would stand to gain from hiring discrimination and I have a hard time believing they would take a significant financial hit just to be grinches.
Yes, I was thinking about that as I wrote, one imperfect but possible fix would be "no pay per view for job listings". Which, for all I know, could already be the case.
Nobody would even think to complain if the implication behind this post was that the company was self-sabotaging itself.
To briefly speak of ethical and legal issues, it is highly doubtful that they would have a policy of not hiring women if women actually apply for the position (and that would most certainly be illegal).
Much more likely is that whoever ran the Facebook campaign checked off that option in an dumb attempt at controlling ad campaign costs? Or maybe they ran a separate campaign to recruit women? Or maybe some other non-nefarious reason.
Fine tuning a message to an audience segment is something every marketer should do. When I first read the headline, this was the use case that seemed obvious.
Different audience segments may want different things from a job. For example, an increasing number of workplaces are offering paid paternity leave. Maybe that's something you want to specifically mention to men.
"While I see both sides of the argument, as long as the advertiser pays per view, it seems wrong to require the listing to target demographics that are almost certainly uninterested."
1) Is making a little economy really more important than giving equal opportunity ?
2) What makes you think you, or the person that designed the ad, knows best what might women (or any other group that might be discriminated against with similar practices) are interested in?
And that, especially when advertising for a job! Women may not be particularly "excited" by the message that conveys the ad, but they might be looking for a job all the same.
All your points sound the same to me, it comes down to : "women like cooking, teamwork and are not interested in being trucker, policemen or technologists, and men like sports and being stong and don't want to be a hairdresser or a nurse" now let's apply these hard facts to cut the costs when advertising for jobs. That way we make sure our stereotypes stay true, and it doesn't matter if it's unfair, it's just how are things.
And I don't think I can change your strongly stereotypical view of gender in this comment alone, but I'm telling you : that can change, and women want that to change, because they want equal pay, and equal pay also means equal opportunity to high paying jobs.
Employment opportunity is not the ideal topic to split hair, let's just not target ads at one gender.
If every person was the same on the average, then we would not have targeted advertising. Google/Facebook et al would not have invested millions (or billions?) of dollars in building a targeted ad network. The evidence is against you in terms of where people (both the seller and buyer) place their money.
Knowing what a type of person wants on the average is not discrimination.
No one is excluded though. A woman can buy Men's Health; a man can buy Cosmopolitan. What Facebook is doing is enabling straight up exclusion of genders from advertising, which is completely different in the eyes of the law.
That's not really how digital advertising targeting works though. In fact, they go to great lengths to make sure you illegally exclude by a class (Google is especially strict about this) and don't transmit anything that is considered PII or "potentially PII". Gender, like IP address, is in the potentially PII realm.
What this actually is going to do is target "people interested in typically male interests". That is going to end up including men, women and trans men. More importantly, the hit rate of gender-targeted ad spend trends towards 50% regardless of gender. It is not an effective tactic.
This is totally overblown.
/signed someone who used to be a digital advertiser in the legal space with a yearly budgeted spend north of $100 million.
I'm trying to explain that gender targeting is imprecise.
When I say the hit rate of ads targeting women is 50%, I mean that of the number of people who see ads that were targeted at women, only 50% of them were women. Targeting men works a little bit more consistently, by usually about 10%, but I don't know why.
That's industry-wide. I don't know if Facebook does better or not, but I have no reason to think so. Even if they did, it probably doesn't matter that much.
50% of viewers were women on ads targeted to women? How much better than chance is that when you take into account there are more men than women on facebook? It seems low given how well they can guess less tangible things like sexual orientation...
well, in this case it isn't perfectly fine because it shouldn't be described as "Gender. Target ads to women, men or people of all genders." if there is only 50/50 chance to display ad to woman. Isn't it just ridiculous and straight lie to potential advertisement buyer?
Ad buyers are very sophisticated spenders and know the score.
If they don't, they're going to get their money taken anyway and a lot more savagely than a mistargeted campaign hinged on gender.
I really have to drive this point home. Selecting gender as a target on Facebook isn't actually selecting gender and everybody involved knows it. It's a much simpler way to describe a much more complex set of decisions.
I was speaking more about the industry as a whole originally.
Facebook knows what you report as your gender but that's not a very good attribute to target off of. If I were doing e-commerce, for example, I would never use Facebook's implementation of that target (if they're going based off of what's reported to Facebook) because it would completely exclude gift purchases across genders. It would leave a ton of money on the table and be a very stupid thing to do.
My entire argument is that splitting hairs over this is irrelevant because any advertisers who rely on the accuracy of that attribute are stupid and pissing money away to begin with.
We are very specifically on a topic about what facebook does, and what I was originally replying to you about was that I think your general findings are not particularly relevant when looking at facebook. You said this story is totally overblown because gender targeting isn't normally very accurate, but here it absolutely will be because a large number of people will accurately set this value on their own profile.
> I would never use Facebook's implementation of that target (if they're going based off of what's reported to Facebook) because it would completely exclude gift purchases across genders. It would leave a ton of money on the table and be a very stupid thing to do.
But we're not talking about gifts, we're talking about jobs.
> because any advertisers who rely on the accuracy of that attribute are stupid
But why would it be inaccurate for adverts on facebook for logged in facebook users?
So let's think this thing through for a second. This will be best read in a Louis Rossmann-esque manner.
Hypothetical me: An HR staffer at a company with a headcount large enough to need an HR department. I get to post job listings and spend advertising dollars for those listings on Facebook.
Hypothetical my manager: Female. As ~75% of HR managers _globally_ are.
Oh yeah, hypothetical me again: Also female; depending on country and department rank, your HR staffer is 70-90% likely to be female. (It's the outlier, but in your country, new HR hires are 96% women)
The argument: Hypothetical me is going to buy ads, that my manager and probably my mostly-female coworkers can see, where I checked a box that says "only show this job listing to men". In an overwhelming social climate pushing for inclusivity and equal pay. In a work culture where these efforts are being _led_ by HR departments.
On what planet do think that this actually occurs? It is almost guaranteed that _at least_ one woman is involved in the decision making process or can audit it. Where are these companies with entirely sexist hiring processes that have significant advertising dollars to spend and where those dollars spent are going to make a significant impact in whether a man gets that job over a women in that industry and at that company?
Every industry is different. As far as legal advertising goes, in many states it's heavily regulated (especially Florida); traditional spend still outperforms digital as far as lead quantity.
Depending on the campaign/type of spend, digital advertising tends to convert leads in a range somewhere between 0.02% and 10%. 10% is extraordinarily good. Lead quality largely depends on the type of case, but ROI on the good cases is anywhere from low 5/6-digit auto injury cases which require little effort from a lawyer to hundred-million and even billion dollar judgements/settlements.
In the same vein, redlining wasn't targeting racial minorities per se, it was just targeting things that correlated with it highly. Under current civil rights legislation that is still highly illegal; gender-ish policy is still an explicit policy considering gendered things.
> "A recipient’s express or admitted use of a classification based on race, color, or national origin establishes intent without regard to the decision-makers’ animus or ultimate objective. Such classifications demonstrate a discriminatory purpose as a matter of law. See Miller v. Johnson, 515 U.S. 900, 904–05 (1995); see also Wittmer v. Peters, 904 F. Supp. 845, 849–50 (C.D. Ill. 1995), aff’d, 87 F.3d 916 (7th Cir. 1996). “Put another way, direct evidence of intent is ‘supplied by the policy itself.’” Hassan v. City of New York, 804 F.3d. 277, 295 (3d Cir. 2015) (quoting Massarsky v. Gen. Motors Corp., 706 F.2d 111, 128 (3d Cir.1983) (Sloviter, J., dissenting)).
Where a plaintiff demonstrates, or an agency determines, that a challenged policy overtly and expressly singles out a protected group for disparate treatment, “a plaintiff need not prove the malice or discriminatory animus of a defendant ….” Bangerter v. Orem City Corp., 46 F.3d 1491, 1501 (10th Cir. 1995); see also Ferrill v. Parker Grp., Inc., 168 F.3d 468, 473 n.7 (11th Cir. 1999) (“[I]ll will, enmity, or hostility are not prerequisites of intentional discrimination.”). Rather, the focus is on the “explicit terms of the discrimination,” Int’l Union, United Auto. Aerospace & Agric. Implement Workers of Am. v. Johnson Controls, Inc., 499 U.S. 187, 199 (1991); that is, how the recipient’s actions specifically deprived or otherwise adversely affected the individual or individuals of access to a federally funded program or benefit. Even benign motivations for racial classifications are presumptively invalid and trigger strict scrutiny in Equal Protection Clause and Title VI cases. Adarand, 515 U.S. at 223–24 (1995); Grutter, 539 U.S. at 326" [1]
>Just seemed odd to go out of your way to include the other three categories on those two axes and leave this one out.
Nothing in their post suggests they "went out of their way" to exclude a particular demographic. To people who do not deal with trans issues on a daily basis, it can be easy to overlook all the nuances of being trans. Calling someone out in a way that accuses deliberate motivations is borderline toxic, no matter what social issue you're voicing.
I wasn't calling anyone out, I don't deal with trans issues on anywhere close to a daily basis (or really ever, tbh) and I in no way mentioned or assumed the parent comment's motivation, let alone assuming it's negative.
Mentioning the "going out of your way" wasn't a call-out, it was a way of justifying the existence of my comment: if he hadn't mentioned trans people at all, I wouldn't have at all bothered with my comment. Since he was explicitly considering trans people in his comment, the addition seemed appropriate, and it came more from the stereotypical engineer-y impulse of "technically this is incorrect in a non-trivial way, IMO".
This is a bizarre position for me to be in: I get that the conversation around trans issues is extraordinarily toxic, and if anything I'm usually the one dismayed by its toxicity, but please don't impute motives to me that I in no way expressed. Frankly, the way you're looking at a simple statement and assuming all sorts of ulterior motives and implicit negative feelings is a precise mirror of the perhaps more commonly-known source of this toxicity, in which attempts at understanding from those less attuned to the issues (or as you put it, "dealing with it on a daily basis") are attacked as bigotry.
I don't intend this to be insulting, but given that your comment implies a certain level of self-awareness about intellectual honesty and charity[1], I suggest that you take a look at the assumptions about me that _you_ are bringing to this conversation. The solution to reductive, toxic discourse isn't reductive toxicity in the other direction.
[1] I haven't been active on HN for a while and have been on a board with much higher standards for intelligent, intellectually honest conversation, so I suppose I was uncalibrated and should've expected the knee-jerk down vote reaction I got, regardless of the political valence of my comment. Again, going off of your comment alone, I assume you're substantially above the median in this respect.
If someone - anyone buys Men's Health then they're probably interested in the topics they cover. If someone has a browsing profile and Likes that relate to those topics, they will get categorized very similarly. As a result, they are likely to see the ads accordingly.
Right now, Google thinks I'm a 65yo+ woman with high income. When I turn off my ad blocker (rarely!), I see ads related to retirement, AARP, etc.
If you were to only recruit from single gender schools then I personally would have an issue with it, same as this issue. However, if you are recruiting from many universities and some happen to be single gender, that is not discriminating.
I can kind of see what they're getting at. Women are under-represented in the tech industry. I don't think many people would disagree with that.
But this feels like a blunt instrument for a highly-nuanced issue. If diversity of experience is a core value at Duolingo, why not recruit everywhere and tailor the interviewing process to favor candidates from non-traditional backgrounds?
While they've achieved an important milestone, I'd be very worried about the line of thinking that led to this decision if I were a Duolingo investor. All the diversity in the world doesn't mean much if they can't find a way to keep the lights on.
> Duolingo has a policy of only recruiting from schools that have an above-average percent of women in their CS program.
Thats a misleading statement.
IMO, The main reason why they did that is because Carnegie Mellon University has more woman than men in its incoming CS classes, and Duolingo wanted an excuse to exclusively recruit from CMU (Which one of the founders teach at, and is strongly associated with).
So they are more interested in being politically correct than in recruiting the best talent available? My respect for them just went down. This is not a way to fight gender discrimination, but rather to perpetuate it.
It's more like: you're in a certain group, you get hired, and you know that the company that hired you has this policy of prioritizing hiring from your group. You spend your days in that company wondering whether you got are there because you're good at what you do, or merely because you are member of the prioritized group. I wouldn't want to be in such position.
Would you enjoy always wondering whether e.g. people in your team are listening to you because what you're saying is worth saying, or if they are merely humoring you and tolerate you in their team on account of orders from above?
Something like this must do ugly things to one's mind and one's perception of self-worth.
What's fair about perpetuating the systematic discrimination that women in tech face?
And in what way is this special treatment?
Companies use a variety of different metrics by which to determine which schools they should recruit from. "We want to increase our diversity" is a reasonable metric to use, and "recruit from colleges with an above-average percentage of people in a minority group we've chosen to view as important" is a reasonable way to accomplish that.
I shouldn't have to point this out, but diversity is more than just a buzzword. Increasing diversity on the development side improves the final product, for a variety of reasons.
What if the companies that were placing job ads on Facebook were also placing job ads on other platforms that weren't targeted to men only? Playing devils advocate here...
I think the previous question was about the institution. In this case, Facebook has people of all genders, not just one.
It only be an apt analogy to the Universities if Facebook were some single-gender social network. Or if an all-gender University let you only recruit one gender.
The devil probably has enough advocates at this point and doesn't need your contribution.
But in general, there are classes of ads that you just can't target based on certain classes, period. Even if you hypothetically might have the ability to target a counterbalancing ad somewhere else. Facebook repeatedly has been in hot water over this, and both they and the people using their ad platform to do this stuff need to get more than a slap on the wrist, in order to deter others from trying the same in the future.
Well if Facebook wasn't a large portion of the online advertising market then there might be an argument. When you try to become a monopoly you get extra responsibilities as you gain market share
Where in the world did you get that from? I never said anything of the sort.
I literally take issue with it, as in I don't think it is right. I said nothing about the implications and I certainly don't think someone should be jailed for poor recruiting practices.
Making something illegal means putting people in jail directly or if they refuse to cooperate. In many countries this is what happens to people with what you would consider poor recruiting practices. It's usually called something like discrimination against a protected class. Many people in power today consider your rejection of jailtime for poor recruiting practices a sexist position.
Advertising on one platform for one gender doesn't mean the job is only for one gender. And if job was open to one gender only, there're already laws against that...
Saying that "FB is not a single gender platform" is like saying that "Higher education is not a single gender platform".
Recruiting only from Barnard doesn't mean that job is only for one gender either.
You're missing their point about using limits on the market supply to discriminate for a job - i.e. even if that job is "open to all", but you've limited applicants to one gender.
HR and legal depts that don't want their company to be sued for violating Title VII will make sure that they aren't recruiting only from Barnard, unless they're some bona fide exception (e.g. hiring models for women's clothing). This idea that they need to recruit from men's only career fairs is your own.
I'm saying that one-gender-only job advertising is not a novelty that happened for the first time in Facebook.
As other commenter pointed out, how is this different from advertising in women/men magazine? Should we ban those as well? Sure, one can argue that their readers are only 98% one gender. But then some people on FB lie about their gender and some browse FB on somebody else's account. E.g. My GF never logs out off Facebook on shared devices and I see ads targeting women all the time.
All in all, we should look at bigger problem of targeted advertising rather exclusive loaded cases. I've no idea what targeted advertising line is "good enough". On the other hand, targeted advertising is definitely causing massive societal problems and political polarisation.
It's not entirely different in the eyes of law - in fact, advertising a job solely in men-targeted magazines would be considered as a discriminatory recruitment practice and be just as illegal as these targeted Facebook ads.
Man, what an unimaginable dystopia, where VS’s hand were forced to the point where they made ads that played on men’s desires and were visually appealing to them ... /s
If you mean, “what if they had to make ads that promoted men wearing their products”, I get your point, but I thought it was a funny ambiguity.
Basically, I think the better lens is one that looks at discriminatory hiring practices, rather than one that looks at advertisements to a selected demographic.
> Would this be different if, for example, Facebook was targeting people for job ads based on "people who like action movies" or "people who like romance novels" (I'm stereotyping here, but just assume you pick a category that ends up being a decent proxy for gender)?
No, this would still be illegal (in the US, anyway). The only difference is that Facebook wouldn't be aiding/abetting an illegal hiring practice. Employers can and have been penalized for hiring from homogenous applicant pools.
Exactly. This doesn't necessarily mean these companies advertised the job to one gender. It simply means the ad was targeted at a specific gender. For all we know, the companies spent twice as much on a different ad set for women.
Theoretically? For the same reason many companies have different advertisements that target different populations - because different advertisements play to different needs/desires, so you can engage with people in the most efficient way in terms of conversion.
Under the law it would different, none of your examples are protected classes, discrimination in employment is only illegal if the targeted/excluded group is a protected class, those being: men and women on the basis of sex; any group which shares a common race, religion, color, or national origin; people over 40; and people with physical or mental handicaps.
The company I work for, for example, does not hire smokers (and makes this clear during hiring and makes the specifics part of our contracts), and this is perfectly legal. But it would be illegal to not hire anyone based on being in, or not being in, one of those specific protected classes.
Now, if they were being used as proxies for those classes, then yes it would be illegal, but it would be up to a court to determine if that was the case. If I ran a company that made action movies "Must love action movies" being a basis for my hiring decisions is obviously not intended as a proxy for a protected class and would be legal. If my company makes printer paper, then I'm on shaky legal ground.
That being said, I'm not sure if there are actually laws making it illegal to publish discriminatory job ads for someone, but Facebook assisting in illegal and discriminatory actions is an issue regardless of if their part is specifically illegal or not.
Context matters. The question is one of harm... is it perpetuating or expanding an existing injustice or privilege, and to what extent? People who dislike action movies aren't the subject of any severe underrepresentation in a job market, so it's clearly not as big of a problem.
>I'm a bit mixed about this; it is clearly discriminatory to publish a job ad that says "women need not apply",
In general that is discrimination, but as with all things in law the general rule only exists for exceptions. One exception is a “Bonafide occupational qualification”...some famous cases involve men playboy, hooters, and Abercrombie and Fitch (though A&F was a case of discrimition against blacks, which A&F attempted to defend by arguing being white was a bonafide occupation qualification because it fit their overall marketing and target demographic).
Let's say that a man has a subscription to The New York Times, but the NYT chooses to print a Men's version that has some job postings and a Women's version that has other job postings without telling you. What if the man is a cosmetologist and isn't seeing any cosmetology job postings because they're all in the Women's edition? The issue with targeting by sex is that it presupposes interest in certain areas, and while a person's interests can change, a person's sex can't [without significant medical intervention].
Yes, it would be very different. For example, the two things you listed are probably rather bad proxies for gender (even if it's only hypothetical).
All they had to do was pick the niche they're going for (which may incidentally turn out to be highly skewed to one gender) rather than making it 100% one gender. I honestly don't even understand the purpose of advertising anything to only men or only women. (The obvious answer would be "sexism", but presumably the advertiser is making some assumption of economic benefit for only targeting one or the other.)
I honestly don't even understand the purpose of advertising anything to only men or only women.
If you pay per view, it's entirely rational to target towards the demographic that produces the most spend for your ad budget.
Why would you waste money advertising, e.g., mascara or sanitary pads to men? Sure, there's probably some infinitesimal fraction of men who would be interested, but targeting them profitably is not a trivial exercise.
Why would you waste money advertising to women who don't want mascara or don't need pads? I would think you'd be better of targetting people who bought some sort of eye makeup if you wanted to sell mascara, and target people who bought pads or tampons if you were selling pads.
I'm not an ad man, but you have imperfect data about your customers, and I gather it's all about trying to predict what they might buy based on very imperfect/incomplete information about them. Plus, your ability to apply more advanced heuristics is constrained by your ability to develop them, both from a technical & marketing standpoint.
"User is a woman" would, in this hypothetical example, be one factor in your larger weighting algorithm. A weighting factor that is, might I add, quite easy to implement compared to detecting male performing arts majors with a specialty in costume design.
> Why would you waste money advertising to women who don't want mascara or don't need pads?
Because you don't know in advance. It's entirely why advertisers offer so many different bits of information on you, to optimise this.
> I would think you'd be better of targetting people who bought some sort of eye makeup if you wanted to sell mascara, and target people who bought pads or tampons if you were selling pads.
And people do this, if they have this information. Of course, there's also the side that you don't want to advertise to someone you strongly expect to buy, because it's the conversions from no-sale to sale you care about.
In classic rubber duck-debugging style, I ended up realizing that as soon as I pressed submit.
In that case I suppose it's not very ethically questionable for things that are very clearly 100% (within a minuscule margin of error) men's only (prostate health, etc.) or women's only (sanitary pads).
I guess the real quandary is for stuff that's male-dominated like hard, dangerous physical labor (coal mining etc.), but not 100% men. If a recruiter can statistically prove that the most efficient use of their advertising dollars to hire for a coal mining job is by targeting only men, 1) is it sexist and/or unethical, and 2) should it be legal?
> a recruiter can statistically prove that the most efficient use of their advertising dollars to hire for a coal mining job is by targeting only men, 1) is it sexist and/or unethical, and 2) should it be legal?
I put this at the same level as only hiring women to teach elementary school. Yes, it is sexist and uneducated. No, it shouldn't be legal.
I think there's a difference between "hiring only men/women" and "advertising to only men/women". A company could launch a recruitment campaign which is only visible to one gender, but still not discriminate against applicants from the opposite gender. Presumably they have many more ways to list job openings than targeted Facebook ads, so they're not specifically excluding one gender from that job; just from the ad.
Isn’t it discriminatory to have an intent to discriminate based on gender? “We want more <gender> at our company.” Unless gender is a bonafide qualification, that should be illegal just as “we want to hire more white people.”
Right, but let's go with the premise that you have an imbalance, and are trying to bring into balance. Then a gender or race targeted ad seems like a reasonable approach.
I guess my point is an targeted ad isn't evidence of intent to discriminate, and in fact can be a key component of pro-diversity actions.
If it's sexist(Wikipedia definition, yeah... I looked it up) then it's unethical, and if it's unethical I'd assume that's because of the sexism.
Is it unethical and sexist? If it is, being overtly sexist/racist/classist and supporting it with data seems worse to me that just the former. Even it it is, it should surely not be outlawed.
I tried to avoid confusion by referring to the Wikipedia definition, which by my ethical framework pretty much always unethical, but should very rarely be outlawed. It would be a nightmare to validate.
Well, there are many gender specific products there, for example female hygiene products. Doesn't mean the opposite sex will straightly not buy them (Yes, they do, https://mashable.com/2018/01/28/amazon-go-tampons-pads-tech-...), but it is safe to assume that they are probably not interested (Look at the previous link and there is actually an outrage because of it.).
The usual scenario is that you make a bunch of variants of an ad, do some testing and find out that option A works a bit better on women, option C works a bit better on men, and option B sucks; so then you'd run ad A targeting only women, and ad C targeting only men. You don't want to show ad C to any women since you have reasons to believe that showing ad A would be more effective for the same price.
It is unlikely only and all men. They probably don't target men without college degrees. They way they probably get where they did was by successively narrowing down by selecting more attributes.
>>but it doesn't feel inherently discriminatory to advertise a job in, say, "Men's Health" but not "Cosmopolitan" magazine.
Similar but still very different argument. A woman can pick up Men's Health but FB ads just don't show up in her stream. Also Uber specifically and purposefully left women out, unless they say, "oh, we clicked the checkbox by mistake" ("or didn't click the women one..." )
What if they would advertise in a tech journal? Most readers are still probably male... But the journal is related to the job. Would that be discriminatory?
No, because the point is that "most readers" are male. The advertisement should be unisex. If a woman picks up that journal - whether just looking for want ads or because they are legitimately interested in the topic - they should be considered with complete equity to any male candidate. The point isn't where you advertise, it's if you target. You can target by geography. You can target by education. You can't target by sex.
What if the ads were targeted at "people with an interest in driving"? I'm fairly sure the percentage of men that fall under that category is higher than women. I mean there's a bunch of car magazines that are gender neutral (that is, they avoid juvenile jokes or cat suit models), if there's an Uber ad in there it'd still mostly be men that read it.
Anyone can buy Cosmopolitan and read ads, but no one can pretend for being someone other on fb/etc. If zuck and alphabet provided options to tune your profile card and live the other live for a while, an argument would be over and ads would not suffer. Default settings rule the world anyway. People could also escape few echo chambers as a side effect.
> That's just how advertising works. Is that crazy?
Why should advertising be considered the default or the correct method ? People still need jobs to feed themselves and their loved ones. It's about their survival and oppression and discrimination is still a thing.
Would there be a debate if ethnic was used instead of gender ?
This is not lifestyle advertising. Tenders, job ads and other categories that require equal access and transparency have rules to prevent exactly this kind of discrimination and skullduggery.
Facebook had made an error here, as has Uber and this is a pretty straightforward violation.
I don’t know about the US but there are rules in regards to job postings in the UK and likely the EU at large that would make that a violation of anti-discrimination laws.
I can't speak for the UK but in Belgium it's certainly illegal to discriminate based on gender, age, race, ....
There was even a case recently where a man won a lawsuit because he was excluded by a clothing store that was searching for a female employee for doing an inventory job.
"Brussels, 23 August 2018 - The legal action taken by the Institute for Equality of Women and Men reminds employers that sex is not a criterion for refusing an application. The Leuven Labor Court has sentenced a clothing store to discrimination based on sex. He refused a man's candidacy because he wanted to hire a woman."
That being said I also seen adverts asking for women but technically it's against the law and even with all those laws discriminations happen more often than not.
There are exceptions like modelling or acting where roles where there could be reasonable reasons for being gender-specific but not for the tech and public sector jobs described in the article.
Given the way issues like this overlap with the European Convention on Human Rights I would expect it to be very similar across Europe (certainly within the EU).
You're telling me, if I run a clinic or rehab facility where women with trauma are more comfortable with other female therapists, I'm somehow required to advertise to and interview both men and women for a job - even when I'm only short on demand for a female therapist from a specific demographic or with certain qualifications? The same would go for a shortage of male employees based on patient or clientele requests and demands.
I'm not saying Uber has a reason to do this. They may or may not, but to say this feature or ability is illegal is just idiotic. In fact, it's rather useful and worth paying for in my opinion.
> If I run a clinic or rehab facility where women with trauma are more comfortable with other female therapists, I'm somehow required to advertise to and interview both men and women for a job
No, you're not. This is a well-established area of the law, called a bona fide occupational qualification:
> One example of bona fide occupational qualifications are mandatory retirement ages for bus drivers and airline pilots, for safety reasons. Further, in advertising, a manufacturer of men's clothing may lawfully advertise for male models. Religious belief may also be considered a BFOQ; for example, a religious school may lawfully require that members of its faculty be members of that denomination, and may lawfully bar from employment anyone who is not a member. Fire departments can require firemen to be able to lift a given weight to demonstrate that they will be able to carry fire victims out of a burning building.
Men’s health and FB operates at very different scales. FB ads can influence at the global scale and their discriminatory nature can bring potentially much more harm.
Would this be different if, for example, Facebook was targeting people for job ads based on "people who like action movies" or "people who like romance novels" (I'm stereotyping here, but just assume you pick a category that ends up being a decent proxy for gender)?