If you thought "liking" a Facebook page simply meant that you have an affinity for Modern Family or Mumford and Sons, think again.
Researchers from the University of Cambridge said in a recent study [PDF] that Facebook Likes can be used to "automatically and accurately predict" sensitive personal attributes, like sexual orientation, ethnicity, religious and political views, intelligence, happiness, drug use, parental separation, age, and gender.
Based on data from 58,000 volunteers - Facebook Likes, detailed demographic profiles, and results of psychometric tests - the research model ran a fairly good race, correctly distinguishing between homosexual and heterosexual men in 88 per cent of cases, African Americans and Caucasian Americans in 95 per cent, and between Democrats and Republicans in 85 per cent.
"People may choose not to reveal certain pieces of information about their lives, such as their sexual orientation or age, and yet this information might be predicted in a statistical sense from other aspects of their lives that they do reveal," according to the study, which was published in the Proceedings of the National Academy of Sciences.
Social networking isn't always about telling the truth, though, and the researchers provided some tips for gaming the system. The best way to look smart online, for example, is to "like" topics like thunderstorms, The Colbert Report, and science. Strong predictors of male heterosexuality, meanwhile, include Wu-Tang Clan, Shaq, and "Being Confused After Waking Up From Naps."
Being able to predict things about Facebook users might be helpful for online marketers, but it might also be unsafe for those who want to keep certain things about themselves private, the report said.
"One can imagine situations in which such predictions, even if incorrect, could pose a threat to an individual's well-being, freedom, or even life," researchers said. "Importantly, given the ever-increasing amount of digital traces people leave behind, it becomes difﬁcult for individuals to control which of their attributes are being revealed. For example, merely avoiding explicitly homosexual content may be insufﬁcient to prevent others from discovering one's sexual orientation."
"It is our hope, however, that the trust and goodwill among parties interacting in the digital environment can be maintained by providing users with transparency and control over their information," the study said, "leading to an individually controlled balance between the promises and perils of the Digital Age."
While some aspects of the study's results are accurate (differentiating between race and gender, for example), others are not, according to Facebook. The University of Cambridge's research uses a metric called Area Under the ROC Curve (AUC) - a measure of the ratio between true and false positives, not a measure of accuracy. So, as Facebook pointed out, stats on drug and alcohol use and relationship status are only slightly more accurate than, say, flipping a coin.
All Likes are default public information as outlines by Facebook's policies, but users can modify privacy via the Timeline settings.
"The prediction of personal attributes based on publicly accessible information, such as ZIP codes, choice of profession, or even preferred music, has been explored in the past and is hardly surprising," a Facebook spokesman said in a statement. "No matter the vehicle for information — a bumper sticker, yard sign, logos on clothing, or other data found online — it has already been proven that it is possible for social scientists to draw conclusions about personal attributes based on these characters."
It should come as no real surprise that people's Facebook liking activity can predict personal details. But as the rollout of the social network's Graph Search (and the sometimes bizarre results) showed, some people simply "like" things willy nilly. Not to worry - a Tumblr blog has been set up to catalogue all your strange and perhaps embarrassing "likes."