STUDY: Your Social Media Habit Can Predict Your Drug Use

Using Facebook, researchers found a way to figure out what drugs people use.

Social media is a drug. Well, kind of. Certainly, the amount of time we spend endlessly scrolling through our own curated feeds, feeding our addiction to artfully composed images that offer insight into the daily lives of our friends and other people we don’t even know--can make the myriad platforms seem as such. And according to a recent study, our social media use, and the words we use when we text, are actually indicative of our drug use.

The study analyzed data gathered from 11 million Facebook users, and 22 million status updates posted to the profiles of 150,000 people as a means of identifying people who suffer from, or may be prone to substance use disorder; and was a joint effort by researchers from the University of Maryland, Baltimore County, and the Virginia Tech Carilion Research Institute.

By developing machine learning software, the researchers were able to draw correlations between instances of substance use disorder, and various keywords used on Facebook. More than that, different keywords could be linked to different types of substance use, according to the study.

Screenshot via

“A preference for movies such as ‘V For Vendetta’ and ‘Boondock Saints’ is positively correlated with alcohol use, while having a hobby, liking cartoons and shows favored by kids or liking movies and brands favored by girls are negatively correlated with drug, alcohol and tobacco use respectively,” the researchers note.

Still, as uncanny as it seems, the study’s authors see social media as a “promising platform” in regard to identifying substance-use habits, and subsequently preventing future drug abuse.

“We believe social media is a promising platform for both studying [substance use disorders] as well as engaging the public for substance abuse prevention and screening,” the study’s authors conclude. “for all three types of [substance use disorder] our models achieved over 80 percent prediction accuracy.”

How any said prevention based on these findings, and in this context, would, or should go down, is open to interpretation.