During this week, we will discuss how to use a keyword expansion technique to find social media posts about a specific topic.

Presentations

Mathias is presenting:
You Can’t Stay Here: The Efficacy of Reddit’s 2015 Ban Examined Through Hate Speech
Eshwar Chandrasekharan, Umashanthi Pavalanathan, Anirudh Srinivasan, Adam Glynn, Jacob Eisenstein, and Eric Gilbert (2018)

Hifsa and Katrine are presenting:
The Spread of True and False News Online
Soroush Vosoughi, Deb Roy, and Sinan Aral (2018)

Readings

A lighter reading list this week: only two readings. However, please note that the lab this week requires that you understand the first article relatively well. The first article describes a method that uses a supervised learning model to aid in the selection of topic keywords. You will need a strong intuition for how the method works if you are to going to be able to follow along in the lab. To reiterate, I suspect that it will be very difficult to follow the lab without carefully reading through and understanding the article. The second article (as discussed in its appendix) uses that method in a large-scale field experiment.

  1. Computer-Assisted Keyword and Document Set Discovery from Unstructured Text
    American Journal of Political Science, 2017, 61 (4): 971-988
    Gary King, Patrick Lam, and Margaret E. Roberts

  2. How the News Media Activate Public Expression and Influence National Agendas
    Science, 2017, 358: 776-780
    Gary King, Benjamin Schneer, and Ariel White
    (Read the main article itself and SI 5.4, i.e. Appendix pages 30-31)

Lab

Reference code from the video tutorial: Keyword_Expansion_Lab.R
Tweets from Members of Congress: MOC_Tweets.rds