Expert Commentary

Predicting tie strength with social media

2009 paper by the University of Illinois at Urbana-Champaign attempted to predict the strength of real-world relationships based on social media interactions.

As social media ecosystems have developed over time, questions over the dynamics between the online and real worlds — the degree to which they overlap — have been the subject of speculation, debate and research. What sorts of correlations can be drawn between online interactions and face-to-face friendships?

A 2009 paper by the University of Illinois at Urbana-Champaign, “Predicting Tie Strength With Social Media,” used 70 different numerical indicators from more than 2,000 Facebook friendships to predict the real-world tie strength of the interpersonal relationships.

Findings of the paper include that:

  • A model based on seven factors, including the level of Facebook intimacy, communication intensity and friendship duration, was able to predict with 85% accuracy the strength of real-world interpersonal ties between Facebook friends.
  • Structural factors such as “common stated interests” and “shared online groups” were not individually predictive of strong ties; however, when they were viewed in conjunction with other factors of online intimacy, they were strongly predictive of tie strength.
  • Factors not available to the researchers — who “friended” whom, for example — are hypothesized to be strongly predictive of tie strength when analyzed alongside traits of the online relationship.

This study has implications for social-network services, including mechanisms of friend introductions, privacy controls and group creation, the researchers note.

Tags: technology, Facebook, Twitter

About The Author