Study: Women Less Likely To Be Shown Ads For High-paid Jobs On Google

An anonymous reader writes: A team of researchers from Carnegie Mellon University has found that women seeking jobs are less likely to be shown ads on Google for high-paying jobs than men. The researchers created more than 17,000 fake profiles, which were shown roughly 600,000 ads on career-finding websites (abstract). All of the profiles shared the same browsing behavior. "One experiment showed that Google displayed adverts for a career coaching service for '$200k+' executive jobs 1,852 times to the male group and only 318 times to the female group." The article notes, "Google allows users to opt out of behavioral advertising and provides a system to see why users were shown ads and to customize their ad settings. But the study suggests that there is a transparency and overt discrimination issue in the wider advertising landscape."

Comments

Algorithm

[bondsbw]

Perhaps women are 6 times less likely to click an ad for $200k+ executive jobs. If the algorithm prioritizes ads based on past behavior of other persons, given all identifiable traits of each person, then this is very well to be expected.

And would go to show that stereotyping is not always evil, but sometimes it comes from innocently putting together past information to be more efficient today.

Re:Algorithm

[Anonymous Coward]

But of course that won't stop someone with a spreadsheet & a mission from finding a correlation & implying a sinister causation.

This is what happens when you let sociology/psychology students think they're engaged in "science." I don't think people attracted to social fields have the capability to investigate and analyze things with dispassionate rationality.

Re:Algorithm

[Penguinisto]

Then prove him wrong. Show us *why* it's dumb... you cared enough to reply, now let's see your reasoning.

His theory has solid reasoning when one considers that the vast majority of advertising in other media is geared toward women, because women do the most purchasing (one count shows it at ~80% ) .

So what's your rebuttal?

Re:Algorithm

[dave420]

Well, to be honest, the original poster just suggested an explanation and went no further. Until they proffer their evidence, it's perfectly acceptable to call bullshit without offering your own study in rebuttal. "Put up or shut up" starts at the beginning, not at an argumentally-convenient point thereafter.

Re:Algorithm

[AmiMoJo]

The ads that Google shows you are based on your search terms most of the time.

Except when it's not. Which in this case clearly indicates there's a profile that's made up of more than just search terms.

You understand what "most of the time" means, right?

If you had read the abstract you would know that the researchers used the privacy settings and minimal profiles that didn't provide any other information, then after building up a search history set only the gender field. So, they carefully made sure that that the only information available was search history and gender.

Are you suggesting that advertisers shouldn't be allowed to target ads? Are you suggesting freedom to engage in advertising should be modified by rules?

Yes, of course. Advertisers already have lots of rules that they must follow. No lies, no misleading claims, no adverts for tobacco products, no adverts for toys in the breaks between children's programmes based on them etc. The exact rules depend on your jurisdiction of course, but there is certainly precedent for not allowing behaviour that is deemed harmful to society in general.

This might also run foul of European human rights rules. For example, insurers can't give discounts based on gender. Banks can't weight women's lower than men's when considering mortgage applications. It's possible that without a good reason (e.g. advertising for products that can only be used by one gender) the advertisers may not be allowed to discriminate in this way.

Re:Algorithm

[grahamsz]

I can't offer you anything more than an anecdote, but i work for an apparel brand and generally women buy far more than men. I'd even guess that women buy more men's clothes than men.

In targeting ads it's generally a good strategy for us to buy ads that just target female buyers because the roi is significantly better. Not sure if that factors into other decisions, but I expect that might have some impact on it.

Re:Algorithm

[penandpaper]

Or perhaps advertisers look at the revenue of those ads and act accordingly.

...it's not a good thing.

Why is it not a good thing? You mean, ads follow market forces which is made up of individuals that act on their own accord and interests? How in the hell is that not a good thing? individuals may act against their own interests... but are you going to be the good dictator and ensure that everybody does what is best for them?

Re:Algorithm

[amicusNYCL]

Women purchase the vast majority of consumer products, including half of "traditional male" products. So, it would make sense that the majority of ads would also be targeted at women. If the majority of ads are targeted at women, then the chance of any one woman seeing a particular ad is reduced compared to the chance that a man would see the same ad (assuming, of course, that the ad in question targets both men and women).

Also, I'm not sure what "LMOL" means, but your reply was one of the dumbest ever. If you don't have anything intelligent to add, just keep your mouth shut instead of letting everyone know that you have nothing intelligent to say.

Re:Algorithm

[bondsbw]

So, my career potential should be limited because people of my gender weren't sufficiently interested in high-paying jobs?

No, your career potential is limited because you are relying on Google to advertise a job to you instead of searching for it yourself.

Re:Algorithm

[Vitriol Angst]

I'm a dude, on a Mac, using plugins to control JavaScript and social networking. My advertisements are 90% for some Mac cleansing product that is 10% worse than paying for a virus.

So not only am I not being targeted by high paying jobs, I'm being profiled as an idiot. I'm tempted to burn my digital jock strap.

"Advertisement"?

[mrex]

Is this an article or an advertisement? The icon says "Ad", but it's listed as submitted by an "anonymous reader" and gives the appearance of being a news article.

Is Slashdot trying to destroy itself?

Newest Study:

[Mocko]

Women Less Likely To Be Shown Ads For Shitty Jobs On Google
In an obvious policy of sexism, female's browsers were less likely to be sent openings or training for plumbing, roofing and landscape services.

No explanation was given by press time.

Re:Newest Study:

[AmiMoJo]

Do roofers and landscapers really make $200k+ in the US? Sounds like a great job, landscaping stuff for $200k/year.

Focused advertising based on detected trends

[gurps_npc]

The article said browsing behavior was identical, but I doubt google was magically detecting women.

At some point the women told Google their gender. Why? What moron thinks Google needs to know their gender?

But once you give Google (or Facebook, or Yahoo, or basically anyone...) information like gender, then I guarantee you they will correlate it with other people.

What this means is that somewhere in Google's algorithm they have found that people that claim to be women (this is the internet after all), are less likely to click on ads for high paying jobs.

So Google wisely decides to show them less such ads.

Do not blame Google for basing their ads on what they know about you and ALSO what they know about people like you.

Do blame yourself for telling Google that much about you.