As scholars re-theorize and re-define digital rhetoric, they also actively re-theorize and re-define methodologies to account for these developing understandings. VanKooten argues that digital rhetoricians must develop “new and hybrid methodologies” to account for in-flux and emerging complexities of digital spaces, including circulation studies. Broadly speaking, I used a technofeminist approach in my research, where I used feminist methodologies to study technologies (in this case, activism on social media) (Bates, McCarthy, & Warren-Riley, 2019; Tetreault, 2019). In particular, I enacted feminist practices by choosing to protect people who could possibly be implicated in the write-up of my study, as I elaborate elsewhere, though I recognize that they could also be considered co-authors. I also align with Dadas (2016), and Porter & McKee (2009) who advocate for flexible approaches that respond to “messy” research situations and ethics. Following these scholars, I blend autoethnography, hashtag tracking, and quantitative methods for this project. In this section, I outline my methodological assumptions, explain how my methodological assumptions applied to my study, and then describe the tools and methods that I used to collect and analyze data about #womenswave.
Since 2015, scholars have used “iconographic tracking” (Gries, 2015) to study how images and hashtags circulate (Gries, 2015; Edwards and Lang, 2018). Tracking the circulation of hashtags is important because as “mediatized place[s]” (Bonilla & Rosa, 2015, 6), circulation can point to the material effects of those hashtags (Edwards & Lang, 2018; Gries, 2015) which “enliven” the things with which they come into contact (Alford, 2016). As such, hashtags are an appropriate site for circulation research.
Tracing these material effects is important because hashtags impact humans. Though some posthuman and new materialist scholars have challenged the anthrocentric orientation of previous research, and I concur with these scholars that rhetoric and composition scholars need to think about non-human forces (Johnson, 2018; Reyman, 2018), I take a technofeminist stance and try to carry feminism’s focus on humans into my research stance. Like Walsh et al. (2017), I am concerned with the effects on everyday people, many of whom may be overlooked in new materialist scholarship (448). These material effects can include perpetuating bigoted ideologies (Noble, 2018), but they can also include activist outcomes (Bonilla & Rosa, 2015; Edwards & Lang, 2018; Navar-Gill & Stanfill, 2018; Vie, 2014). If hashtags have material force (Edwards & Lang, 2018) that affect people, then it is important to understand not only why and how hashtags work and affect the material world by catalyzing both activist movements and hate groups, but also why and how they fail to have material consequences for humans. As such, this orientation aligns with feminist and queer methodologies which disrupt dualisms; it challenges the binary distinction “offline” and “online” and attends to the ways that technologies and bodies are integrally intertwined (Haas, 2012). However, I recognize that these material consequences are always in-progress. As Queen (2008) points out in her articulation of rhetorical geneaology rhetorical productions are “continually evolving rhetorical actions that are materially bound” (476).
Queen (2008) defines rhetorical genealogy as "a process of examining digital texts not as artifacts of rhetorical productions, but, rather, as continually evolving rhetorical actions that are materially bound, actions whose transformation can be traced through the links embedded within multiple fields of circulation. Rhetorical genealogy is rhetorical analysis that examines multiple processes of structuring representations, rather than seeks to identify the original intentions or final effects of structured (and thus already stabilized) representations" (475-476).
With these methodological assumptions, I chose to follow the #womenswave as it emerged and travelled in the rhetorical ecosystem of the 2019 Women’s March because I was interested in the nuanced use of the hashtag as it accrued positive and negative affect. Following digital research methods outlined by Wolff (2018), I used TAGS (version 6.1.9) (Hawksey, 2019) to track hashtags’ online circulation. Using the template automatically collected public Tweets (and the corresponding metadata) using #womenswave, #womensmarch, and #womensmarch2019. I ran the script on January 24 over the last seven days to collect Tweets from both before and after the march.
Next, I downloaded and used the free software Orange v3.16 (Demšar et al., 2013) to separate my data into parts, select specific information, and calculate co-appearing words. I used Orange to sort through the “text” column of my TAGS data Orange and then I filtered out non-semantic information. Orange generated a word cloud (see Figure 1) and a list of the most commonly co-appearing words. The word cloud tells me little on its own; I cannot know if the words were used to positively or negatively index the hashtag (Bonilla & Rosa, 2015; Wolff, 2017; Beveridge, 2017; Navar-Gill & Stanfill, 2018). However, this quantitative data augmented my qualitative data and gave me a bird’s-eye view of the hashtags’ circulations.
A notable limitation to my quantitative data is that I only collected data from one mode of circulation (Twitter). As Queen (2008) notes, hashtags are “embedded within multiple fields of circulation” (476) and increasingly move across modes (Alford, 2016). However, for this study, I focused solely on Twitter because hashtags circulate widely on Twitter and TAGS collects Twitter data.
Scholars have argued that digital ethnographic methods can add depth and complexity to quantitative research (Murthy, 2013). Following McNely (2016), to augment my quantitative research, I engaged in a form of digital autoethnography and attended and documented the local downtown Women’s March in Orlando, Florida on January 19, 2019. The march embedded itself into the 35th annual Martin Luther King Jr. parade that began at 10 a.m. (Souverain, 2019) and ended in Lake Eola Park. McNely (2016) argues that autoethnographers must acknowledge their own subject positions and document their own experiences with empirical evidence, which can include photographs and field notes (146). As such, I documented the material manifestations of the march with photographs. I paid particular attention to protest signs and materials, especially those featuring #womenswave and other hashtags. I also took videos to document speeches and actions. However, following Queen (2008), I also recognize that these material consequences and meanings which I traced are not fixed (476), thereby embracing the “messiness” of the research (Dadas, 2016; Porter & McKee, 2009).
Though the group norm at the public march included taking pictures for the purposes of sharing them on social media, due to the sensitive nature of the issues addressed, I choose to omit images and direct quotes from individual people. Though my banner image in this webtext includes people to showcase the march, I chose to blur the image to further protect participants. My decision aligns with Porter and McKee's (2009) feminist approach to internet research ethics; they contend that researchers should take a case-based approach to ethical choices about including online images in publications as researchers renegotiate shifting public/private divides (Porter & McKee, 2009). My approach is important because it points to how researchers’ conceptions of publicity and ethics in offline research sites have shifted in response to changes in thinking about the public/private dichotomy which stemmed from concerns originally focused on online spaces. Additionally, following Tetreault (2019) and Bates et al. (2019) I recognize that my own embodied and lived position inflected my research; my interpretations are limited by my experiences as a non-Jewish, white, woman and all of the privileges that come with that position.
Broadly following the method of “thick description” (Geertz, 1973), I paid attention to protest signs and took field notes about roles, rules, behaviors, and the space. While I noted elements of interest to my study, in alignment with McNely (2016), I also noted my own sensations and affects (145). Though I recognize that I can never fully extricate myself from the movement and that my presence itself altered and affected the material intra-actions, I chose to not overtly align myself with the march, as I was uncomfortable with the anti-Semitic affective components that had “stuck” to the movement.
After the march, I typed my field notes and phrases on the signs in a password-protected account. I uploaded my pictures to the same account and hyperlinked these digital pictures back to my field notes to re-create the experience for myself as fully as I could in what I am calling “augmented thick description.” Since rhetors constantly re-attune to their environments (McNely, 2016), it was important for me to recreate the context as fully as possible for my analysis. Furthermore, following technofeminist researchers, I am reminded of my own intersectional and embodied position as integrally implicated in my technologically-mediated research through details in these images, such as my shadow.
Please feel free to see my examples of augmented thick descriptionby clicking on the link provided. However, please also note that this thick description is truncated and incomplete to protect the people who attended the march, as part of my technofeminist approach in since the people at the march may not have anticipated being included in my research. This example shows how I used augmented thick description to do two things; first, I hyperlinked to images to remind myself more fully of the affective experiences of my own embodied position. Second, I demonstrate how I documented and augmented my own experiences by looking up answers to questions that I had, thereby augmenting my experiences at the time of the march as well. Though the document is linear, it enabled me to walk back through my experiences and reminded me of my affective experiences.
Combined, I collected large-scale data about the global circulation of the hashtag online and local data demonstrating the material effects of the hashtag on people. I acknowledge that “success” can mean many different things. Mimetic theory would see the replication of the hashtag as success (Vie, 2014), though Dieterle, Edwards, and Martin (2019) contend that virality does not necessarily equate to success. However, I am interested in successful policy implications, rather than how many times something has been replicated. In their article on what makes a hashtag successful, Bonilla & Rosa (2015) claim that hashtags are successful because they a) create interpretive frames/say what something is “really” about, b) create intertextual chains by linking to other hashtags, c) create complex publics and reduce noise, d) create dialogues, e) are eventful and allow people to participate, and f) can become condensed symbols of larger arguments (5-8). I analyzed my quantitative and quantitative data with these definitions of successful hashtags in mind.