A Spatial Analysis of Rape Culture News Events on Twitter
Megan Stubbs-Richardson, Colleen Stouffer, and Sierra Nelson
Extensive research has documented evidence of a rape culture across mediums and contexts. This study contributes to this body of research by examining how contextual aspects of place and mainstream versus Twitter event-driven news vary in information diffusion, both topically and spatially. Research on this topic is timely given that several rape cases (e.g., Steubenville and Torrington) first came to light via use of social media outlets (Moody-Ramirez, Lewis, and Murray; Woods 2014). Research suggests that new media outlets, like Twitter, can serve as a public sphere wherein citizens can gather and discuss ideas (Papacharissi 2002; Murthy, 2012), such as to advocate against rape culture (Keller, Mendes, and Ringrose 2016; Mendes 2015; Sills et al. 2016). Using data from the Human Sensor Project, we collected 5,000 U.S. geo-located tweets with the keyword requirement that rape and/or sexual assault be present in tweets collected from March 6, 2014 to March 19, 2015. The current study employs Critical GIScience and Feminist Theory to analyze and spatially pinpoint tweets on rape culture (i.e., tweets that condone or normalize sexual violence), as well as tweets that oppose rape culture (i.e., tweets that combat rape culture ideology) across the United States. First, we qualitatively and quantitatively analyze 5,000 geocoded tweets to examine how discussions of rape culture and news events vary spatially in the United States. Second, we incorporate contextual-level variables to estimate characteristics of the population at the county-level that may be predictive of how sexual assault is depicted on Twitter. Additional analyses were used to determine whether news events exhibit a localized clustering pattern or one characterized by greater dispersion.