“Number is the ruler of forms and ideas, and the cause of gods and demons,” proclaimed Greek mathematician Pythagoras—likely toga’d before an adoring audience of laurel-wreathed maidens and bronze-chested gents, daydreaming about how they’d really like to square his hypotenuse. Jump ahead 2,500 years, and Western society has fallen head over heels for the quantitative method. Numbers, math and statistics drive our software, our elections, our measure of wealth and intelligence, our insurance policies, our health records and increasingly our school systems. “Sexy,” though, is not the way mathematics is generally viewed. In fact, much of the Romantic tradition was waged in opposition to the contention that qualitative abstractions like love and attraction could be quantified. “God forbid that Truth should be confined to mathematical demonstration!” proclaimed William Blake—likely before returning home to his wife Catherine’s non-Euclidean geometry.
Were he a 21st-century man, and had his marriage become compromised by the dashing Lord Byron, it’s unlikely that Blake would have turned to OkCupid. It’s not just the stigma of online dating that every user has to overcome before turning to a matchmaking site; OkCupid, the free, wildly popular service, is unabashed in its quantitative approach to the laws of attraction. Unlike, say, eHarmony, which disguises its algorythms in the conceit of matching compatible souls, OkCupid is borderline autistic in its embrace of numbers as an indication of statistical compatibility. Through a virtually endless series of user-generated questions, a potential couple is literally offered a numerical percent compatibility (and plenty of statistical evidence for this equation, should the user decide to parse it out).
Some will say (and Blake would agree) that the numbers are just the beginning, a quantitative way to winnow the field before exploring true chemistry in the warm-blooded arena of instant messaging. Yet, matchmaking effectiveness aside, OkCupid has almost accidentally come to serve another larger purpose. The vast amount of data that the site collects from its nearly 7 million users—at first, for marketing purposes—can offer all kinds of statistical insight into how users interact. Launched in 2009, OkTrends, the site’s official blog, has been poring over this data and posting its findings. The reliability of this data is upheld, the site claims, not only due to its enormous sample size (compare the 3,050 people Gallup used to project the outcome of the 2008 presidential election with the millions of interactions that happen on OkCupid everyday) but the fact that dating site activity tends to be more reliable than general blind polling.
The site’s findings range from the instructive (“What to say in a first message,” based on an analysis of grammar and content in a salutation vs. the reply rate) to the ridiculous (based on 250,000 users, vegetarian women are twice as likely to enjoy performing oral sex; vegetarian men are a quarter more likely), market-research fodder (“Sexual activity by smartphone brand”—anyone surprised that iPhone users are the most promiscuous?) to the stuff of sociology class (religion and its relationship to writing proficiency; ethnicity in relation to profile sophistication). Acknowledging a user’s inclination to game the site in order to make themselves more attractive, the site has dedicated a number of studies to the ways users lie. (On average, men list themselves as two inches taller than they really are; both genders are 20 percent poorer than they list themselves; and 80 percent of those who list themselves as bisexual are ultimately interested only in one gender.)
Not unlike WNYC’s science radio show Radiolab, the authors and analysts at OkTrends present their quantitative findings in an impressively human manner, describing their hypotheses and findings in colloquial English and colorful infographics. The goal isn’t just to help users maximize their service; there’s a geeky imperative at the heart of the blog that seems to genuinely want to understand this human problem of love and sexuality through the most concrete tools we have. And they aren’t withholding the uncomfortable stuff. Wyoming ranks atop a list of states where users are most likely to act out rape fantasies. Black women are categorically the least likely to receive a reply to an initiated first message. And some of the site’s findings have profound political implications. In response to homophobic sentiment on the part of politicians like Jim Demint, OkTrends statistically undermined the stereotype that gay people are more promiscuous than straight or interested in “converting” straight people to gay. Meanwhile, “red” states show some of the highest rates of “gay curiosity.”
While Pythagoras might view the OkCupid algorythm as humans finally finding the equation for human attraction, it seems that Blake would be equally pleased by OkTrends’ brave display of love’s diverse nuance. Love might not be fully communicable in numbers, but everyone gets off on a good pie chart, right?