Steve Jennings/ Getty Images for eBay/AFP
We get it. Things are bad. Sexism is institutionalised. It trickles down from the upper echelons and affects everyone.
But what about when there no institutions at play? When an individual connects with another individual to shake hands on a deal on the internet? How could gender inequality come into play there?
We aren't sure why. But this is how they do:
If a man and woman were to sell an identical new product on eBay, the woman would receive fewer offers from buyers.
eBay has a treasure-trove of data about this. It's largely been kept under wraps, but two guys - Tamar Kricheli-Katz, a sociologist and legal scholar from Tel Aviv University, and Tali Regev, an economist from the Interdisciplinary Center Herzliya - gained access.
Their discovery - sexism is every bit as bad as it sounds. And it's also in the places we simply didn't imagine it could exist. Such as the the value of two identical products, depending on the gender of person selling it.
Kricheli-Katz and Regev first shortlisted 420 of the most popular products sold on eBay. Between 2009 and 2012, that amounted to 1.1 million transactions. That's enough data for most experiments.
So they started running the data to see if they found patterns based on the gender of the seller. And women consistently received 20% less than men for the same product.
The result was the same even for - wait for it - gift cards. That's right.
Think about that for a minute.
There's a gender bias when purchasing gift cards.
The researchers controlled for initial prices, reputation of sellers and even auction options like Buy it Now. None of those factors played any role.
But here's the really tricky bit: most eBay sellers don't mention their gender. So how did that gender bias even begin to take place?
This is where Kricheli-Katz and Regev went one step further. They tried to see if we can guess the gender of the seller without the seller mentioning it. And it turns out we can.
They had 400 people guess the gender of 100 randomly chosen sellers. They allowed clues like the names of the sellers and what other items they tended to sell. And participants correctly guessed the gender of 56% of the time, declared 35% unguessable, and got less than 9% wrong.
How? Katz and Regev had a hypothesis. That men, perhaps, described their wares better - and that's what contributes to the bias. So they tested that out as well. They performed a computer analysis on all the words used in the titles and subtitles of the advertisements. More positive sentiments received more points.
And here they finally did find a difference.
"The sentiment analysis showed that women and men sellers do, indeed, resort to different sentiments".
Great. Case solved. Women make less $$ on eBay than men do because they don't market their sales well. Right?
Wrong. When the researchers controlled for sentiments in titles and subtitles, they still found a massive difference between offers received by men and women. The difference had now fallen from 20% to 19%. But 19% still remains stubbornly larger than 0.
What's most frustrating about this research is that we now know a bias exists, but don't yet know why. What stereotypes about female sellers are causing them to make less money? That a buyer will have to deal with more exclamation marks if he buys from a woman? Pink stickers on his/her purchases?
We can't fight a bias when we don't know what it is. The good news, however, is that there's now one more bias we are going to one day set right.
Edited by Payal Puri