It could out you as gay.

It might reveal how you vote.

It might even suggest that you’re an unmarried introvert with a high IQ and a weakness for nicotine.

That’s the conclusion of a study published Monday in Proceedings of the National Academy of Sciences. Researchers reported analyzing the likes of more than 58,000 American Facebook users to make guesses about their personalities and behavior, and even whether they drank, smoked, or did drugs.

Cambridge University researcher David Stillwell, one of the study’s authors, said the results may come as a surprise.

“Your likes may be saying more about you than you realize,” he said.

Facebook launched its like button in 2009, and the small thumbs-up symbol has since become ubiquitous on the social network and common across the rest of the Web as well. Facebook said last year that roughly 2.7 billion new likes pour out onto the Internet every day – endorsing everything from pop stars to soda pop. That means an ever-expanding pool of data available to marketers, managers, and just about anyone else interested in users’ inner lives, especially those who aren’t careful about their privacy settings.

Stillwell and his colleagues scooped up a bucketful of that data in the way that many advertisers do – through apps. Millions of Facebook users have surveyed their own personal traits using applications including a program called myPersonality. Stillwell, as owner of the app, has received revenue from it, but declined to say how much.

The study zeroed in on the 58,466 U.S. test takers who had also volunteered access to their likes.

When researchers crunched the “like” data and compared their results to answers given in the personality test, patterns emerged in nearly every direction. Since the study involved people who volunteered access to their data, it’s unclear if the trends would apply to all Facebook users.

The study found that Facebook likes were linked to sexual orientation, gender, age, ethnicity, IQ, religion, politics and cigarette, drug, or alcohol use. The likes also mapped to relationship status, number of Facebook friends, as well as half a dozen different personality traits.

Some likes were more revealing than others. Researchers could guess whether users identified themselves as black or white 95 percent of the time. That success rate dropped to a still impressive 88 percent when trying to guess whether a male user was homosexual, and to 85 percent when telling Democrats from Republicans. Identifying drug users was far trickier – researchers got that right only 65 percent of the time, a result scientists generally describe as poor. Predicting whether a user was respectively a child of divorce was even dicier. With a 60 percent success rate, researchers were doing just slightly better than random guesses.

The linkages ranged from the self-evident to the surreal.

Men who liked TV song-and-dance sensation “Glee” were more likely to be gay. Men who liked professional wrestling were more likely to be straight. Drinking game aficionados were generally more outgoing than, say, fans of fantasy novelist Terry Pratchett. People who preferred pop diva Jennifer Lopez usually gathered more Facebook friends than those who favored the heavy metal sound of Iron Maiden.

Among the more poignant insights was the apparent preoccupation of children of divorce with relationship issues. For example, those who expressed support for statements such as “Never Apologize For What You Feel It’s Like Saying Sorry For Being Real” or “I’m The Type Of Girl Who Can Be So Hurt But Still Look At You & Smile” were slightly more likely to have seen their parents split before their 21st birthday.

Some of the patterns were difficult to understand: The link between curly fries and high IQ scores was particularly baffling.

Jennifer Golbeck, a University of Maryland computer scientist who wasn’t involved in the study but has done similar work, endorsed its methodology, calling it smart and straightforward and describing its results as “awesome.”

But she warned of what the work showed about privacy on Facebook.

“You may not want people to know your sexual orientation or may not want people to know about your drug use,” she said. “Even if you think you’re keeping your information private, we can learn a lot about you.”

Facebook said the study fell in line with years of research and was not particularly surprising.

“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,” Facebook’s Frederic Wolens said in a written statement.

Wolens said that Facebook users could change the privacy settings on their likes to put them beyond the reach of researchers, advertisers or nearly anyone else. But he declined to say how many users did so.

For the unknown number of users whose preferences are public, Stillwell had this advice: Look before you like.

The like button is “quite a seductive thing,” he said. “It’s all around the Web, it’s all around Facebook. And it’s so easy.”