Analyzing Attitude of Second Screen Social Media Mes-sages

Partha Mukherjee, Bernard Jansen

Research output: Contribution to journalArticle


We analyze more than 3,000,000 Twitter, 800,000 Instagram, and 50,000 Tumblr posts concern-ing a single major in-real-life event, Super Bowl XLIX, to determine attitude. We consider three event phases (Pre, During, and Post). Findings show link-based recommendations and undirected broadcast patterns positively correlate with attitude in the Pre and Post phases, respectively. The usage of these specific features highlights the differing information needs of viewers during these phases, specifically the sharing of information in the Pre phase and sharing of opinions in the Post phase. The volume of postings indicates a negative attitude for all social media platforms, demon-strating that adverse information is more likely to be shared than positive information; this finding contradicts prior findings. This second screens phenomenon research is important in identifying the of sharing information on multiple social media platforms during an in real life event.

Original languageEnglish
JournalIEEE Intelligent Systems
Publication statusAccepted/In press - 1 Jan 2018


  • attitude
  • dual screens
  • panel data regression
  • Second screens
  • second screens
  • social soundtrack

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Artificial Intelligence

Cite this