How social networking sites may discriminate against women

April 20, 2018, Columbia University School of Engineering and Applied Science
A network effect known as homophily may reduce women's visibility on social media when recommendation algorithms are added, says a new study. Above, a selfie from study coauthor Ana-Andreea Stoica's Instagram account. Credit: Ana-Andreea Stoica

Social media and the sharing economy have created new opportunities by leveraging online networks to build trust and remove marketplace barriers. But a growing body of research suggests that old gender and racial biases persist, from men's greater popularity on Twitter to African Americans' lower acceptance rates on Airbnb.

Now, using the photo-sharing site Instagram as a test case, Columbia researchers demonstrate how two common recommendation algorithms amplify a network effect known as homophily in which similar or like-minded people cluster together. They further show how algorithms turned loose on a network with homophily effectively make women less visible; they found that the women in their dataset, whose photos were slightly less likely to be 'liked' or commented on, became even less popular once recommendation algorithms were introduced.

By working out the math of how this happens, the researchers hope that their work, to be presented April 25 at the Web Conference in Lyon, can pave the way for algorithms that correct for homophily.

"We are simply showing how certain algorithms pick up patterns in the data," said the study's lead author Ana-Andreea Stoica, a graduate student at Columbia Engineering. "This becomes a problem when information spreading through the network is a job ad or other opportunity. Algorithms may put women at an even greater disadvantage."

The researchers scraped their data from Instagram in 2014, after Facebook bought the company but before automated prompts made it easier to connect with friends-of-friends. Though women outnumbered men in their sample of 550,000 Instagram users (54 percent to 46 percent), the researchers found that men's photos tended to be better received: 52 percent of men received at least 10 'likes' or comments compared to 48 percent of women.

A majority of hyper-influencers in the researchers' sample were women, but when the Adamic-Adar recommendation algorithm was introduced, men were three times more likely than women in this exclusive group;to be suggested as a new contact to others on the network. Credit: Ana-Andreea Stoica

As expected, homophily played a role. The researchers found that men were 1.2 times more likely to 'like' or comment on other men's photos rather than women's, while women were just 1.1 times more likely to engage with other women.

When they used two widely used recommendation algorithms—Adamic-Adar and Random Walk (friends-of-friends)—the researchers found that the percentage of women connected to, or predicted to be recommended to, at least 10 other Instagram users fell from 48 percent in the original dataset, to 36 percent and 30 percent respectively. As predicted in a series of mathematical proofs in the paper, the researchers also found that the disparity was greatest among Instagram's super-influencers—people like Instagram CEO Kevin Systrom, whose popular posts and 1.5 million followers put him in the top tenth-of-one percent for engagement.

When algorithms were turned loose on this exclusive network of ultra-engaging individuals, women's visibility plunged. Though women in the top .1 percent for engagement (with at least 320 connections) outnumbered men (54 percent to 46 percent), the men were far more likely to be suggested to new users and expand their networks rapidly. Just 26 percent and 28 percent of in the top .1 percent were likely under the Adamic-Adar and Random Walk algorithms respectively to be recommended at least 23 times and 12 times, the found.

"Algorithms pick up subtle patterns and amplify them," said the study's senior author, Augustin Chaintreau, a computer scientist at Columbia Engineering and a member of Columbia's Data Science Institute. "We're not asking that algorithms be blind to the data, just that they correct their own tendency to magnify the bias already there."

The study is the latest to show that recommendation algorithms, in addition to filtering content, may influence the long-term structure of a social network. "It's remarkable that a simple assumption of homophily leads algorithms to amplify disparities in social status," said Amit Sharma, a researcher at Microsoft Research India who was not involved in the study but recently spoke at Columbia about his own work exploring recommendation engines and social influence.

Algorithmic interventions that balance convenience with ethical goals may be one way to address the problem, he added. "Through studies like this, we're learning that the practice of optimizing a single metric exclusively, for example, number of new friends added, is not the right way. Unfortunately, the alternative is unclear. We are still scratching the surface of understanding how algorithms affect long-term human behavior."

Explore further: Snapchat challenging Facebook among US youth: survey

More information: Ana-Andreea Stoica et al, Algorithmic Glass Ceiling in Social Networks, Proceedings of the 2018 World Wide Web Conference on World Wide Web - WWW '18 (2018). DOI: 10.1145/3178876.3186140

Related Stories

Researchers study nuances of social media 'likes'

June 24, 2016

The simple act of pressing a "like" button on a social media site can speak volumes about a user's preferences, relationships and behavioral patterns, according to Penn State researchers, and such findings about a user could ...

Recommended for you

Researchers make coldest quantum gas of molecules

February 21, 2019

JILA researchers have made a long-lived, record-cold gas of molecules that follow the wave patterns of quantum mechanics instead of the strictly particle nature of ordinary classical physics. The creation of this gas boosts ...

11 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

JamesG
4.3 / 5 (6) Apr 20, 2018
Now do a study on how it discriminates against men. Be fair or you're discriminating.
psychotherapist
3.7 / 5 (6) Apr 20, 2018
Hear that boys, it's now discrimination to not engage equally with women's content.

"This becomes a problem when information spreading through the network is a job ad or other opportunity. Algorithms may put women at an even greater disadvantage."

Researchers fail to explain how the visibility of women's posts to people has anything to do with the visibility of job ads to women. The only way algorithms could reinforce women not seeing job ads is if they were already ignoring them.

"We're not asking that algorithms be blind to the data, just that they correct their own tendency to magnify the bias already there."

Such a moderate position.
Lasz
4.1 / 5 (9) Apr 21, 2018
Um, sorry but unequal outcomes DO NOT prove bias. By then "correcting" for this imaginary bias, you are actually introducing bias into the system. You haven't taken ANY other factors into account as to what attributes to the differences in outcome, other than "bias against women." This is not scientific research, this is propaganda for equality of outcome.
tallenglish
3.6 / 5 (5) Apr 21, 2018
Women dont like other women as they usually see them as the enemy, men don't care in the same way (and I bet most of the comments were insults). Women tend to like other mens posts as a form of flirting or networking, men don't like things as much (I know I never did) as they likely would get nagged from the wife/gf as to why they liked x persons picture, especially if it is another woman.

Not bias from the algoritm, but human behaviour differences between men and women.
Lopital
3.7 / 5 (6) Apr 21, 2018
This man hunt needs to stop. You are destroying your own future.
jjesterj
2 / 5 (4) Apr 21, 2018
I bet her parents are crying $75,000 tears over this Insta-narcissism, Russian like-bots would've been cheaper.
julianpenrod
2.3 / 5 (3) Apr 21, 2018
So, now, people won't be allowed to like the people they like? The Democratic Racket promises groups various things like popularity, political control through influence, even if undeserved, and they're determined to arrange it! Forget about that some people aren't likable, the Democratic Racketeers will work to arrange a fraudulent "popularity" for them! The algorithms work by employing rules accepted as being legitimate and at the base of popularity, but that isn't serving the political conniving of the Democratic Racket. So they're going to change things to make them what will make the Democratic Racketeers richer. The Democratic Racketeers act as if they feel what choice do the people on "social media" have?
mrburns
4.2 / 5 (5) Apr 21, 2018
Why is an algorithm to enforce equality of outcome worth any attention, except perhaps as a manifestation of psychopathology ? Social Justice doesn't belong in science in any shape, manner or form whether it is algorithmic or just the author's opinion. Besides this kind of thing isn't new, Social Justice Gangsters have been altering data, ignoring reality they don't like and silencing presenters of genuine science for years now as they try to create a reality in which they control all.
Geni-us
2.3 / 5 (3) Apr 23, 2018
Social Justice, Virtue Signaling, and Forced Equality/Diversity is actively "regressing" our ability as a species to develop and maintain complex "BENEFICIAL" relationships. We are literally WITNESSING the dumb-ing down of the human species by FORCED EQUALITY OF OUTCOME.
tblakely1357
1 / 5 (1) Apr 23, 2018
Many feminists claim that women are smarter and 'stronger' than men yet also claim that women are always victimized by men. Never made sense to me.
PTTG
3 / 5 (2) Apr 24, 2018
Holy fuck. This is it; I'm leaving Phys.org. This place is infested with sexless misogynists. It's not that the articles are bad, It's just the disgusting community has finally pushed me out.

You all disgust me.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.