Social Media, U.S. Presidential Campaigns, and Public Opinion Polls: Disentangling Effects

  • Jennifer Stromer-Galley
  • Patricia G. C. Rossini
  • Jeff Hemsley
  • Kate Kenski
  • Feifei Zhang
  • Lauren Bryant University at Albany
  • Bryan Semaan
Keywords: campaigns, polls, Twitter, Facebook, algorithms


Political campaigns in the United States routinely use social media as part of their strategic communication. Although digital campaigns have been a topic of scholarly interest for nearly two decades (Bimber and Davis 2003; Foot and Schneider 2006; Stromer-Galley 2014), few studies examine the relationship between public opinion polls and communicative strategies online. Our study fills this gap by looking at how voting intention polls relates to types of campaign messages on social media. We analyze Twitter and Facebook messages of all Republican and Democratic candidates during the surfacing stage and primaries of the 2016 U.S. presidential election. All messages from candidates' campaign accounts are coded via automated text analysis focusing on strategic messaging, including attack and advocacy messages. Using time series analysis, we compare the types of messages candidates produce on social media and their standing in public opinion polls. Our hypothesis is that the candidate's position in the polls will drive certain types of campaign messages for strategic advantage. We predict that time matters. Candidates are sensitive to the opinion climate and adjust strategies in response to their standing in polls. Our results contribute to understanding the relationship between strategic campaign communication on social media and public opinion polling.
How to Cite
Stromer-Galley, J., G. C. Rossini, P., Hemsley, J., Kenski, K., Zhang, F., Bryant, L., & Semaan, B. (2016). Social Media, U.S. Presidential Campaigns, and Public Opinion Polls: Disentangling Effects. AoIR Selected Papers of Internet Research, 6. Retrieved from
Papers S