Words of Others: A Mixed-Effects Analysis of Forwarding in Social Media

Abstract

We study the decision process that guides forwarding behavior on electronic media. Forwarding messages, which amounts to porting the words of others to new contexts not envisaged by the writers of the words, carries both benefit and risk for the users of social media. Using data collected from 250 university students, we build a linear model of forwarding behavior with respect to different types of content over three social media: e-mail, Facebook and Twitter. Our model includes fixed as well as (crossed) random effects to capture variability in the propensity to forward messages that can be attributed to the medium, the content and the individual user. We find that forwarding is strongly influenced by both the medium and the content of messages. Forwarding is most common on Twitter and least so on Facebook. Messages that are thought to be useful to the recipient, and those containing jokes or humor represent the top two types of content forwarded. The interaction of medium and content, is sparse: only one combination – current news on Twitter – leads to a significant increase in forwarding. The forwarding behavior of individual subjects, with respect to their media and content preferences for forwarding, shows significant variation. We discuss the implications of our findings for viral marketing and the emergence of norms about how social media can be used for forwarding.