romantic relationships and happiness. The company's Data Team sliced and diced the language used in millions of peoples' status messages, then looked at how they varied depending on the relationship status the people listed themselves with.Facebook this weekend published a special Valentine's day study about
Conclusions? Married people are the happiest, people in Open Relationships are the least happy. Men are less happy than women in an open relationship (believe it or not) and more happy in marriage. These findings are interesting, but what they really indicate is that there may be a modern-day Farmers' Almanac for understanding our lives hidden behind the company's doors. Facebook needs to set that data free or at least do more with it.
The Data Team's conclusions were written up by intern Lisa Zhang, whose relationship status isn't public on the site. Zhang explained the methodology like this:
We already have methodology for measuring the happiness of Facebook users: by considering how many positive words people use in their status updates (see the USA Gross National Happiness Index). This method allows us to see whether a person's Facebook relationship status affects how positive and negative they are. We examined the use of positive and negative words in the status updates of all English speakers over the course of one week in January. To protect your privacy, no one at Facebook actually reads the status updates in the process of doing this research; instead, our computers do the word counting after all personally identifiable information has been removed.
One of the most significant findings reported is about people who listed themselves as in an Open Relationship, meaning typically that they are committed to one person but have sex with multiple people. Those people tend to be less positive than anyone else, less positive even than widowed people!
This is truly remarkable, if you think about it. For what tiny percentage of human history has it been a common practice to publicly declare your relationship Open in such a way? It's a pretty new thing! How's it working out for people? Apparently not so well!
That could be good information to know before making certain decisions ("if I knew then what I know now..."). Of course, how does having children affect happiness in married people? How happy do people in heterosexual vs same-sex Open Relationships tend to be? Do women who are engaged to men with feminist or liberal-sympathetic interests (Fan pages) express less negativity than other engaged women or is that just a facade that in reality means nothing?
There are lots of questions raised by this data, and it's just one of an infinite combination of data points that Facebook could analyze.
It's probable that some of this data includes important observations about the human condition. Information that could help people make better-informed decisions than ever before in human history; informed not only by our own experience, and the experiences of the people we observe in our immediate lives, but by the experiences of hundreds of millions of people around the world.
Other Examples of Important Advice
Remember the Farmers' Almanac? It's a book that's been published annually since 1818, filled with advice about cooking, gardening, humor and weather predictions. The mysterious Farmers' Almanac team (there's only been 7 editors in the publication's history) refuses to disclose its methods for weather prediction, but claims to have a 80 to 85% accuracy rate over its history.
Similarly, Facebook data might not be able to tell us with 100% accuracy whether people who move from Michigan to California, or who marry young, or who stop playing football and start playing basketball tend to be more happy or less happy than before - but that data could come a whole lot closer to telling us than anything we've had before has.
Of course we're all special snowflakes, with infinite complications, and free will is important - but doesn't it seem that there's an incredible opportunity for world-wide self awareness hiding inside this social network where we're typing our relationship status, our location and our interests into fields in a form?
Of course Facebook uses that data to target advertising. Why will the company tell an advertiser that I'm part of a group of college educated, white, married men, over 30 years old and living in the state of Oregon but it won't tell me that among people in those circumstances I have a particularly geographically limited set of friends and should probably get out more if I want to really understand the world? Now that would be valuable information!
Hunch, the startup lead by Caterina Fake and Chris Dixon, makes that kind of data a big part of its social decision making service. The site provides a way for you to walk through various things you should consider in making decisions like what kind of car you should buy next or where you might like to go on vacation. It also asks you questions about yourself along the way.
Among 10k people Hunch asked "do you like Cilantro?" the ones that said they did were far more likely to prefer dark chocolate over milk chocolate. People who said they didn't like cilantro are much more likely to prefer milk chocolate. (Probably because people who don't like cilantro don't know what's good in life.) A Hunch study last week of people who own different breeds of dogs found that German Shepherd fans tend to rely more on intuition than common sense and Pug fans particularly enjoyed the movie The Shawshank Redemption.
How about Facebook coughs up the goods on the far greater supply of user data than any other site in the world has?
Unfortunately, Facebook is going in just the opposite direction. The company's Data Team puts out some lightweight analysis like these Valentine's Day conclusions about once a month. When software engineer Pete Warden tried last week to offer up user data he'd collected from 250 million Facebook users to the Academic community for study (see The Man Who Looked Into Facebook's Soul) the company contacted him and told him to put a hold on the release while privacy concerns were evaluated. Last week the company took down Lexicon, its public tool for comparing how often different words were being used across the streams of Facebook users.
Come on, Facebook! Set the data free. It's not about cilantro and chocolate, that's just the fun stuff. There are important observations about humanity hiding in that data. Check out, for example, Hunch's observations of hundreds of people who said they don't believe Barack Obama was born in the US and what else they have in common.
If you don't feel like you can set it free, then at least do something more serious with it. The Farmers' Almanac is a mysterious organization, maybe some more mystery would be ok if it came along with a whole lot of Facebook data.
This is a historically unique opportunity and one that I hope Facebook will take ahold of soon. Think of all the heartbreak the company could help prevent if only people knew their odds of being happy in an Open Relationship, among the countless other decisions we make that would be well informed by analysis of aggregate user data.