Study: Facebook Can Predict When a Couple Will Break Up


Study: Facebook Can Predict When a Couple Will Break Up

C. Price C. Price • 9/25/14

Can an algorithm accurately guess who your partner is and whether you’ll break up by comparing your Facebook friends?

According to Facebook, it’s already happened. Their scientists claim to have developed a means of analyzing the friends list of both partners in a relationship, looking for the number of mutual and nonmutual names.

Facebook senior engineer Lars Backstrom, alongside Cornell University scientist Jon Kleinberg, published a study on the findings in October.

The social media giant has discovered this type of data is far more reliable than other markers, including status updates or high levels of interaction among users.

With a 60 percent accuracy rate, the program was able to correctly determine which subjects were currently involved in a romantic relationship.

Furthermore, actual couples who were not correctly identified as partners by the algorithm were 50 percent more likely to not be together in two months.

“Couples who weren’t identified as partners were

more likely to not be together in two months.”

Researchers looked at more than one million adult Facebook users who self-identified as either “Married” or “In a Relationship.”

The method they used, which is referred to as the “theory of dispersion,” found a stronger footing among those couples who essentially shared fewer friends between them.

The theory is those couples who are less dependent on one another for shared friendships are considerably less likely to break up. This is attributed to both partners typically having more of a rounded and balanced life outside of the relationship.

The algorithm was able to successfully pinpoint trouble in couples that share a high number of mutual friends.

The method of analyzing the dispersion rates proved even more reliable at predicting a breakup than comparing mutual schedules or looking at a couple’s personal messaging history.

Source: Cornell University. Photo source: