Spread of information on social media
http://www.aaai.org/ocs/index.php/ICWSM/ICWSM10/paper/viewFile/1509/1839/
In recent years, social media has developed to become a primary source of interaction between people. Information is quickly spread through social networks, yet this propagation is still not thoroughly understood. In the paper, “Information Contagion: An Empirical Study of the Spread of News on Digg and Twitter Social Networks,” researches analyze data from two social news sites (Digg and Twitter) to track interest in news stories and its resulting spread over social networks. The aim of this study is to answer the question of whether the network dynamic of a social media site affects propagation of information.
The researchers began by choosing Digg and Twitter because of their high data transparency, allowing for maximum details on user information and data. After characterizing each social media site’s individual dynamics and structure, the team mapped the spread of interest through each network. Both sites have some form of a friends/followers interface with a re-broadcasting action; on Twitter, it is known as “retweeting” whereas on Digg, it is known as “promoting.”
An important distinction of the structure of the two networks is that Digg is seen to be more strongly connected than Twitter. In other words, on Digg, user A who is followed by user B is much more likely to reciprocate the follow than is a user on Twitter. Twitter, on the other hand, has a broader distribution but a more disjointed network. This find is interesting, as it shows that stories spread further on Twitter despite its less strongly connected network. The strongly connected nature of Digg does not contribute to a user promoting a story despite its highly interconnected network. Similar network findings are fairly abundant, such as Granovetter’s discovery in the 1960’s that people discover new jobs more often from acquaintances than from close friends. Overall, the result of this study proves that a network does indeed affect spread; in particular, it proves that the reach of information does not depend on similarity between users.