“Drunk girl”: A text mining analysis of #MeToo on Twitter
Click on image to download PDF version.
Lindstadt, C., Boyer, B., Dongwoo, A., Pounders, K., Wilcox, G., (May 2019), ’Drunk girl’: A text-mining analysis of #MeToo. Presented at the annual meeting of the International Communication Association Annual Conference, Washington, DC. Top poster award, ICA Health Communication Division
Background: Upon going viral in October 2017, the #MeToo movement launched the conversation about sexual assault into the public arena (Mendes, Ringrose, & Keller, 2018). Since then, the hashtag has been used by supporters and detractors alike to take a stand on the issue and comment on sexual violence and harassment cases in the media. The intoxication of one or both parties is frequently at the center of sexual assault incidents, yet the characterization of alcohol in the #MeToo movement remains unclear.
Sexual assault has been recognized as a serious public health issue (CDC, 2018), but addressing the role of alcohol can be murky, risking increased blame for the victim (Abbey, 2002; Perkins, 2002). Research finds women who have consumed alcohol are attributed a greater amount of culpability than those who have not (Grubb & Turner, 2012). What’s more, reports of intoxication are not only harmful to victims, they can even be beneficial for offenders; for example, while female victims are assigned greater responsibility when drunk, male offenders tend to receive less blame when they are intoxicated (Richardson & Campbell, 1982). Research indicates that ‘hashtag feminist activism’, like #MeToo, can bring out complicated and contradictory public discussion on social media (Mendes et al., 2018). Therefore, this research seeks to explore alcohol’s role in the public discourse of sexual assault through the lens of #MeToo.
Methods: We collected and analyzed all tweets using the hashtag ‘MeToo’ and the word ‘drunk’ using Nuvi software (a social media monitoring tool, www.nuvi.com), in conjunction with Twitter’s Search API. Data were collected between October 24, 2017, at 12 a.m. and October 24, 2018, at 11 p.m. (CST). The keywords returned 25,014 messages, of which 5,566 were original mentions. We analyzed each tweet’s content using SAS Text Miner 12.1, a text-mining and social media monitoring software (Chakraborty, Pagolu, & Garla, 2014). The tweets were then categorized into topics using a method of textual analysis that combines algorithmic analysis with manual review (Glowacki, Glowacki, & Wilcox, 2017). The use of publicly available data in this study did not require approval from our institutional review board.
Results: Peak conversation occurred in October 2018 and trending keywords included drunk female, police records, former lawyer, medical documents, and shocking letter. The top trending co-occurring hashtags included: #stopkavanaugh, #womenriseup, #maga, and #walkaway. Nearly all of the tweets (99.9%) were in English and the top five countries were the United States (11,054), the United Kingdom (825), Canada (536), India (265), and Australia (177). From the 5,566 original tweets, six distinct topics emerged [see Table 1 for complete list of topics and their descriptions]. Preliminary analysis of the six topics indicates a range of themes related to #MeToo [Table 2 lists sample tweets for each topic], including support for a victim of a divisive rape- accusation case taking place in the media (topic 1); victims’ personal stories (topic 2); expressions of concern for men in the era of #MeToo (topic 3); accusation of an offender (topic 4); a song written by Chris Janson, ‘Drunk Girl’, describing how drunk women should be treated (topic 5); and a nuanced discussion of the role of rideshare drivers (topic 6).
Discussion: Preliminary analysis of the public discussion of alcohol within the #MeToo movement indicates that in general, tweets mentioning drunk victims (male and female) appear to express support for victims and place responsibility on offenders (topics 1-5). This finding suggests that, broadly speaking, Twitter conversations using the #MeToo tend to perceive victims as sympathetic, even when intoxication is involved in the discussion. However, our research also found a noteworthy theme of ‘concern for men’ (topics 3 & 6). It is possible that including intoxication as a factor in the conversation about sexual assault heightens public perception of risk for men — both for wrongful accusation and threat of being assaulted. Finally, we uncovered a clear indication of politically partisan use of the hashtag (topics 1 & 4). This finding supports prior research demonstrating that digital feminist activism can indeed provoke political participation in the online pubic sphere (Mendes et al., 2018). For more details, see Drunk Girl: A brief thematic analysis of Twitter posts about alcohol use and #MeToo published in Qualitative Research Reports in Communication in 2021.
