Sad Guns, Bad Guns: A Sentiment Analysis of Print Media Discussion of Firearms Policy

• In Event: Weapons

Abstract

Conducting a Lexis-Nexis search on the terms “gun(s)” and “firearm(s)”, eight regional and one national newspaper were sample for articles. These newspapers were the Chicago Daily Herald, the Denver Post, the Houston Chronicle, the Minneapolis Star Tribune, the New York Daily News, the New York Times, the San Jose Mercury News, USA Today, and the Washington Post. These articles were examined and articles that were solely about the commission of a crime were excluded from the dataset. The articles included had to have some discussion of policy to be retained. This filtered dataset produced 19,633 articles from the 17-year period from 1998 through 2014.
For this exploratory study a stratified random sample was generated with 100 articles from each of the nine papers for a nominal sample size of 900 articles for analysis. Each paper’s contribution was converted to a corpus for analysis with textmining tools in the programming and graphics environment of the R open source software. After cleaning theses corpora of punctuation, numbers, whitespace, and common-use words, they were coded for original source and combined into one mass corpus. The sentiment analysis is performed on this combined corpus. A discussion of the results will be presented.

Authors

• Isaac T Van Patten, Radford University

• Nicole Hendrix, Radford University