This post has been reposted in its entirety from original source clearadmit.com.
So last week we shared some tips and cautionary tales from admissions directors at leading business schools about MBA applicants and their social media presence. In a nutshell, we found that different admissions committees use social media to varying degrees as part of the MBA admissions process, but a general rule of thumb is to not be an idiot. In fact, recent research out of the Stanford Graduate School of Business (GSB) suggests that monitoring an applicant’s social media behavior over the long run could actually reveal more intimate traits and core personality than face-to-face interviews. Adcoms: take note.
Assistant Professor Michal Kosinski, who teaches organizational behavior at Stanford GSB, was surprised to learn as part of research he conducted recently just how much what an individual “likes” on Facebook can reveal about a person.
“The thing that was most surprising to me … is that our most intimate traits can be very easily predicted from a digital footprint, and a very general one as well, such as Facebook ‘likes,’” Kosinki said as part of a recent interview with the Stanford News Service.
One of findings that most surprised Kosinski—who holds a doctorate in psychology and whose research interests include big data, crowdsourcing and psychological assessment—was that he was even able to predict whether someone’s parents were divorced or not based on their Facebook likes. In fact, those results seemed so improbable that he began to doubt his own methods and reran the analyses a few times. “I couldn’t believe that what you like on Facebook could be affected by your parents’ divorce, which could have happened many years earlier—we’re talking here about people who might be 30 or 40 years old,” he told Stanford News.
Beyond the marital status of your parents, your digital footprint is also predictive of many other intimate traits, Kosinski found, from whether you smoke, drink or take drugs to your sexual orientation and political views.
“Actually, everything we tried predicting was predictable to some degree, and quite often it was pretty accurate,” he said in the interview.
Come on—you know you’re curious to see if this is for real. Well, you’re in luck: A demo of Kosinski’s models is available here: www.applymagicsauce.com.
Not only were intimate traits predictable from Facebook “likes,” Kosinski found that even broader measures related to your digital footprints can be used in predictions—how many Facebook friends you have, the number of your likes, how many times you log in to Facebook, how often you tweet.
“Each one of those measures is not a very strong predictor of anything on its own,” Kosinski says. “But if you combine many different variables of this kind, each of them slightly predictive, the computer can get a very good idea of who you are.”
NO, YOU ARE NOT AN EXPERT ON SOMEONE JUST BECAUSE YOU’RE FACEBOOK FRIENDS
Now, none of this is to suggest that you can draw deeply personal insights about your roommate’s ex-girlfriend’s sister just because you happen to still be Facebook friends with her and can see what she publicly shares on social media. “Few Facebook likes are so obviously linked with personality or other traits as to allow a human to use them in forming accurate judgments,” cautions Kosinski, adding that language used in status updates or tweets may reveal even less to mere humans.
“Computers, however, are very good at combining thousands or millions of subtle pieces of information to arrive at accurate predictions. We humans, with our limited ability to simultaneously process more than a few facts at a time, are rather bad at it.”
This is why analysis of years of an individual’s Facebook history could well reveal a more accurate sense of the person than an actual face-to-face interaction, Kosinski says. “It’s rather easy for people to misrepresent themselves in, say, a half-hour-long interview or on a first date. It’s much more difficult to monitor your appearances and opinions in years of your Facebook history.”
Which begs a few questions. First, what and how much is Facebook already doing with the information we so gleefully offer up in our newsfeeds? And second, will Stanford GSB’s Admissions Committee ultimately employ Kosinski’s research as part of the MBA admissions process? After all, if computer analysis of a candidate’s digital footprint yields more accurate results than a face-to-face interview, an argument could certainly be made in favor of making the switch.
We checked in with Stanford GSB Admissions Director Derrick Bolton, who agreed that the research was fascinating. “But no, we’re not using it and never could,” he says. “Remember the Hawthorne effect: as soon as it became high-stakes it would change the behavior…”
Stanford GSB consistently has one of the lowest acceptance rates of all leading business schools (7.1 percent last year, compared to Harvard’s 11 percent). So, yes, high-stakes indeed.
Read the interview with Professor Kosinski in the Stanford News Service.