Facebook is still doing damage control in the wake of public outrage after the social media website admitted to toying with some of its users' emotions as part of a science experiment. But seemingly unafraid of the same bad publicity, the popular dating website OKCupid admitted to playing with its users' hearts when it published the results from a series of its own user experiments this week.
On Monday (July 28), in an unapologetic blog post titled "We Experiment On Human Beings!" co-founder and president of OKCupid Christian Rudder seemed to back up Facebook's stance on turning its users into guinea pigs, writing, "If you use the Internet, you're the subject of hundreds of experiments at any given time, on every site. That's how websites work."
Rudder then detailed the results of three studies that yielded some surprising insights into how potential partners interact with one another in the digital age. [The Top 10 Golden Rules of Facebook]
Love is blind?
In the first experiment, OKCupid hid all the photos from its website for seven hours, Rudder explained. Company researchers compared activity on the site without photos with activity on a typical day with photos. The company found that without pictures, users responded to first messages 44 percent more often, conversations were deeper, and contact details were exchanged faster. However, when the photos were put back up, most of the 2,200 conversations that started during the photo-free time frame stopped dead in their tracks.
"It was like we'd turned on the bright lights at the bar at midnight," Rudder wrote in the blog.
In another experiment to find out how much pictures mattered, OKCupid analysts examined how crucial the actual text of someone's profile was compared with their physical attractiveness.
In the past, the dating site allowed users to rate the profiles of others based on both looks and personality. The company expected to find at least some people with great personality ratings who were lacking in the looks department. But that wasn't exactly the case. After plotting the personality and attractiveness ratings of users from past data, OKCupid found that most people's scores for looks and personality were within half a point of each other. There was a positive correlation between "looks" score and "personality" score, they found.
Rudder pointed to one female user who was ranked in the 99th percentile for personality; she didn't fill in a word of text on her profile, but she did use a revealing bikini shot for her user picture. "So, your picture is worth that fabled thousand words, but your actual words are worth … almost nothing," Rudder wrote.
Do people like each other just because OKCupid says they should?
OKCupid assigns a match percentage for every pairing based on an algorithm developed by the company. Unsurprisingly, people are more likely to message and reply to messages from people who share a high match percentage. But in another experiment, when OKCupid told people who were actually a bad match that they were a good match, the odds of a single message sparking a conversation was almost the same as that of a genuine 90 percent match.
But that doesn't mean the match algorithm is bogus, Rudder said. OKCupid then took actual pairs of people with 30 percent, 60 percent and 90 percent compatibility and randomly displayed one of those three match percentages, regardless of the users' actual compatibility. The 30 percent matches who saw the correct match percentage were the least likely to strike up a conversation, with just a 10 percent chance of exchanging four messages. Even when pairs with 30 percent compatibility were told they were a 90 percent match, their odds of talking were still quite low, at 17 percent. Those with 90 percent compatibility who were told they only had a 30 percent match had a 16 percent chance of starting a conversation, while the 90 percent matches who saw the correct compatibility percentage were the most likely to start a conversation, at 20 percent, OKCupid found.
Privacy concerns
In the Facebook emotion study, researchers removed all the positive posts from the news feeds of some users and watched how that influenced the unwitting subjects' moods. They found that emotional states are indeed contagious across social networks. But when the findings were published in the Proceedings of the National Academy of Sciences last month, the tests sparked controversy and concern over users' personal privacy.
Facebook may be facing investigations after European privacy companies filed complaints to the Federal Trade Commission, suggesting the experiments may have violated Facebook's terms of service agreement with its users, according to news reports.
OKCupid's user agreement says personal data may be used in research and analysis. Rudder noted in the OKCupid blog that the only way to improve the website is to run experiments like these.
Your move, Match.com and eHarmony.
Follow Kelly Dickerson on Twitter. Follow us @livescience, Facebook & Google+. Original article on Live Science.
