Data is omnipresent in our daily lives, Big Data—the massive compendium of statistics that is used to guess our political preferences, chart our buying patterns, predict our next movements and so forth—even more so. It gives an impression of order and even science, assuring us that we are rational beings whose actions are predictable and logical, guided by self-interest but with room in our hearts for altruism and justice.

The real world is far different. Just ask Christian Rudder, author of the new book Dataclysm: Who We Are (When We Think No One's Looking), which charts the inner and external lives of Americans along several lines of argument, examining how and with whom we connect, whom we stay away from, whom we aspire to be like, whom we aspire to be. 

The results of his wide-ranging surveys are often unsettling. For one thing, he observes, technically, “a woman is over the hill at twenty-one.” Wearing a tattoo, for another, is likely to have measurable results that follow a person throughout his or her life—just as mom warned. We are obsessed by outward appearances, and much of our virtual interconnection online is really a not-so-subtle beauty contest. Perhaps most unsettling of all, while we profess to be fair-minded and without awareness of difference—racial, ethnic, religious—we are haunted and halted by the fear of anyone who is not quite us.

Rudder came to such conclusions through a body of Big Data that he himself gathered. A math graduate of Harvard, he is one of the co-founders of the online dating site, which went live in 2004. Facebook existed then but with far fewer participants. Twitter hadn’t been invented. “There was no such thing, really, as social media, at least not as we think of it today,” Rudder says. “I remember sitting around at school talking about this thing called Friendster, which no one had seen or used. We pretty much predated all that, and so I never thought about data, let alone Big Data, when we started.”

Continue reading >


The data accrued all the same as Rudder worked on an algorithm that would predict likely success in what people were coming to his site to find—namely, a romantic connection, the old business of the heart subjected to the rigors of science. In assembling this data, Rudder writes in Dataclysm, he was soon working with statistics that were orders of magnitude larger than anything from a Pew or Gallup poll. “What’s being collected today,” he writes, “is so deep it verges on bottomless; it’s easily forty days and forty nights to that old handful of rain.”

Indeed, with 11 or 12 million new accounts being created on the site each year, the numbers are flooding in, and they lend credence to some old truisms. Why is it that a woman’s options, in terms of finding and keeping a male partner, should narrow so markedly as a function of the y-axis of age? Identifying prime causes are for the philosophers, if not the sociologists and ethnologists, among us, but the numbers point to an inconvenient truth: If women seem forgiving of the passing of time and its effects, then men constantly seek younger women even as they age. For every 100 men whose attention might be focused on a 20-year-old dating candidate, Rudder observes, the number narrows tenfold when she turns 30, and it becomes vanishingly small by the time she’s 40 and then 50.

The men talk a good game, of course. A 35-year-old man who comes to the site may set his filters to find candidates between 24 and 40, yet rarely will he select anyone beyond the age of 29. That 30-year-old woman, conversely, may set her filters between 28 and 35 and mean it yet find no candidates willing to spend the time of day with her. As Rudder notes, in a nicely observed passage, “You could say they’re like two ships passing in the night, but that’s not quite right. The men do seem at sea, pulled to some receding horizon. But in my mind I see the women still on solid ground, ashore, just watching them disappear.”

If men and women are talking past each other, perhaps barely catching sight of one another, Rudder_Coverthings get more complicated still when matters of ethnicity and race enter the picture. Rudder stirred up some controversy when, not long ago, he noted in a blog post that OkCupid was experimenting not just with its algorithms, but with users of them to look at matters such as how behavior differed when candidates had no visual clue as to what the other looked like. When they did—well, Rudder notes, “People are very judgmental online.”  They may profess not to care about skin color or the number of vowels in a prospective partner’s name, but they tend to choose narrowly to avoid people who are unlike them. Such signals as “outdoor lover” and “country girl” are giveaways in that regard: A country girl isn’t likely to cross the aisle to date an old-school hip-hop fan, in other words, any more than a tea party supporter is likely to cross the aisle to vote yes on an entitlement increase. 

“I can’t fault someone for not wanting to go on a date with someone else,” Rudder writes. “There’s rarely any malice in that decision.” Nonetheless, there are volumes of prejudice attendant in it, and his numbers point squarely to that fact: White American users of the site are far less likely to “like” African-American candidates than are users in the United Kingdom, Japan, Canada, and other places not cursed by our ugly past and uncomfortable present.

Yet there’s hope. “One thing that surprises me about the numbers is what happens when people actually get together,” Rudder says. “Then a lot of our predictors go out the window. For example, we see that when someone who’s relatively good-looking sits down with someone who’s relatively not good-looking, they get along just fine. It’s the online component of online dating that exacerbates people’s most judgmental tendencies.”

In other words, one key to getting along is to stop looking into a screen and to start looking into someone’s eyes. And that, courtesy of Big Data, is news we can use.

Gregory McNamee is a contributing editor at Kirkus Reviews.