First Monday

Going rogue: Reconceptualizing government employees contentious politics on Twitter by Fatima Espinoza Vasquez, Nicholas Proferes, Troy B. Cooper, and Shannon M. Oltmann



Abstract
In 2016, following the election of President Donald Trump, dozens of Twitter accounts emerged, purporting to represent a network of resistance within the U.S. government. These alt- and rogue- Twitter accounts, known as Rogue Twitter, shared tweets aiming to rebuke the administrations new information restrictions on federal agencies.

Using established social movement theories, we investigated if Rogue Twitter is an online social movement. We qualitatively analyzed 43,569 original tweets from 102 Rogue accounts. We evaluated the tweets on three dimensions: Their attempts to challenge state institutions (macro-level), their organizing and mobilizing strategies (meso-level), and their shared understandings (micro-level).

We found that the Rogue Twitter Movement exemplifies how online social movements engage in coordinated contentious activity via an online platform. Members of this network collectively framed as their main grievance the State’s control of information. Accordingly, their mobilization repertoire focused on calling the attention of the State and the public by openly criticizing the new information control policies. They strategically released controlled scientific information and demonstrated dissent by satirizing Trump. Moreover, they supported off-line political activity by promoting protests like the Science March. This study shows how incorporating multidisciplinary approaches yields nuanced understandings of protest in Internet platforms like Twitter.

Contents

Introduction
Review of relevant literature
Methods
Data analysis and results
Discussion
Conclusion

 


 

Introduction

Shortly after President Trump won the 2016 election, there was concern that public access to federal information might be curtailed (e.g., Gidda and Schonfeld, 2016). These concerns only intensified when, on Inauguration Day, the U.S. Department of the Interior was ordered to cease tweeting with the general public (Kircher, 2017). Social media and communicative restrictions soon followed for several other agencies.

Sparked by equal measures of concern and outrage, ‘alternative’ Twitter accounts were quickly started [1], with names like @AltBadlandsPark (referencing the Badlands National Park) and @RogueFEMA (referring to the Federal Emergency Management Agency). While clarifying that accounts were not managed by federal employees during working hours (due to legal constraints under the Hatch Act; see U.S. Office of Special Counsel, n.d.), the accounts also claimed authority and respectability through linkage to federal agencies. Binkowski (2017) attempted to ascertain whether account owners had demonstrable relationships to the agencies named in their Twitter handles; as of April 2018, she was able to verify 35 out of 100+ accounts.

These accounts stated several purposes and goals, though they often focused on disseminating information (Oltmann, et al. 2020). While the tweets make evident that the account holders opposed the Trump administration, it is unclear how seriously this contention should be viewed. Considering these unexplored gaps in our understanding of the alt and rogue Twitter accounts (hereinafter Rogue Twitter) led us to these research questions:

The Contentious Politics Model (CPM) was originally developed for off-line physically co-located social movements (McAdam, et al., 1996). Our research is significant because we consider whether this model is relevant and applicable to the Rogue Twitter accounts, which may or may not constitute an online social movement. Doing so can provide us with a theoretical framework to better understand assemblages of individuals that develop and come together online during moments of social and political strife.

To address our research questions, we turn first to a review of the literature, where we describe the CPM as one way to understand social movements, with its emphasis on evaluating macro, meso, and micro level tactics. Next, in the methods section, we describe collecting over 43,000 tweets from Rogue Twitter accounts. In the results section, we detail the ways that the CPM model provides structure to our analysis of the data. Finally, we discuss the broader implications of this analysis and address our research questions in light of the data.

 

++++++++++

Review of relevant literature

Using the Contentious Politics Model (CPM) put forth by McAdam, et al. (1996) enables us to study the Rogue Twitter accounts holistically as a nascent online social movement. This model brings together structural and cultural perspectives, calling for an in-depth three-level analysis of social movements’ context, organizing, and collective sense-making. At the macro level, it explores the structures that facilitate or hinder possibilities for social movements’ emergence and participation. At the meso level, it looks at social movements’ structures and mobilization practices. At the micro level, the model investigates how collective ideas, identities, experiences, and attitudes mediate collective action.

At the macro level, opportunity structures are state institutions that regulate political participation (Tarrow, 2011, 1996; Kriesi, 2004; Tilly, 1978). In other words, opportunity structures include all official procedures, regulations, and structures that either help or hinder people’s ability to express themselves politically and participate in political processes. For example, laws that allow people to vote, campaign, protest, organize, and criticize the government constitute American opportunity structures. Such opportunity structures are often the object of social movements’ contention, as they can be influenced to make institutions more open to social pluralism (Kriesi, 1995).

The meso level describes a social movement’s internal organizational infrastructure as well as its practices — in other words, its structure and its (online) repertoires of contention. Movements’ organizational structures consist of formal and informal pre-existing organizations (e.g., civil society organizations, churches, labor unions, or cooperatives), which provide the scaffolding and the assets to help mobilize people, resources, funds, information, ideas, and tactics (McCarthy, 1996). The movements’ structures may take many different shapes and adapt over time (McAdam, et al., 2001; Della Porta and Diani, 2006).

Social movements’ mobilizing practices are also called repertoires of contention; they are the tactics used to call attention to their issues and influence the political system. They are ‘a stock of special skills, plays, and activities with which members of a group are already familiar and from which they select specific ones’ [2]. There are three types of repertoires: intra-movement, those aimed at authorities, and those aimed at security forces. We are concerned with the second kind, the repertoires aimed at authorities, as they target the state and will vary based upon the openness of the political system (Alimi, 2018).

Repertoires are situated and malleable; they evolve with changing technological environments and in response to authorities’ reactions. The affordances of information and communication technologies (ICTs) have transformed social movements’ repertoires of contention in several ways. For example, the reduced cost of participation has created flash activism, short-term, and ‘trivial’ social movements (sometimes derided as ‘slacktivism’) (Christensen, 2011). The reduced need for physical co-presence makes it challenging to prove collective participation, thus making these efforts sometimes appear inauthentic (Tilly, et al., 2010; Khazraee and Losey, 2016). Online repertoires should be seen as yet another set of tools that can be wielded by social movements. Therefore, adopting a holistic approach which incorporates online repertoires will yield better insight into mobilization practices and structures.

At the micro level, social movements define their basis of contention, identify opportunity structures, and determine their mobilizing strategies. This process is called framing; it is a situated purposeful effort by social movement actors which emerges from the need to make sense of the world and to legitimize motivations and repertoires (Snow and Benford, 1988). The outcome of this process is ‘frames,’ which are also strategic and malleable (like repertoires). They evolve to fit the movement’s needs.

Social movements use a coherent blend of existing discourses that resonate with intended supporters. Frames are continuously being redefined to mobilize participants, garner support, and demobilize rivals (Snow and Benford, 1988; Benford and Snow, 2000). Framing dissemination is crucial for a cohesive mobilizing strategy; therefore, social movements employ a variety of diffusion tactics through language, narratives, culture, and media (Snow and Benford, 1988; Johnston and Noakes, 2005).

Those who dissent against the status quo and those who are underserved or marginalized usually look for alternative ways to achieve their goals. Social movements harness grassroots power to create their own mechanisms for participation, organizations, and media.

Moreover, the adoption of alternative ICTs by social movements has been extensively documented by media studies (Garret, 2006; Earl and Kimport, 2011; Carty and Reynoso Barron, 2019). Similar to alternative structures, alternative ICTs are meant to address what social movements see as a power imbalance in the flow and control of information (Atton, 2015). The adoption of alternative ICTs and media is usually a grassroots effort to create stronger information networks by and for those outside of the mainstream. Given social movements’ ‘outsider’ status, they use untapped resources that allow them to challenge traditional media practices, structure, and ideology. For example, during the Egyptian revolution in 2011, social media such as Twitter and YouTube served as alternatives for civil society groups to ‘self-manage’ their information to organize and mobilize, because mainstream media was controlled by the regime (Bhuiyan, 2011).

 

++++++++++

Methods

Social movements engage in dialogue about their issues and mobilize accordingly. To understand Rogue Twitter’s strategies, it is necessary to explore how participants attempt to interact with their institutional context (macro level), how they organize (meso level), and how they describe their internal processes of creating and sharing understandings (micro level).

To build this dataset, we first identified accounts with ‘alt’ or ‘rogue’ in either their username or profile and which referenced a U.S. government agency in some way, for example, ‘RogueNASA.’ We felt our list of 102 accounts was saturated and representative of most major accounts as of early February 2017.

Next, we collected tweets generated by the 102 Rogue Twitter accounts during the first 100 days of the Trump presidency, a convenient time frame of the administration’s initial actions. Tweets were collected using the ‘cron’ Python script for R. This data collection tool relies on the use of Twitter’s public APIs. As the number of tweets generated by these accounts was relatively small, we do not believe that collection via the API was throttled in any way. We believe the tweets we captured to be comprehensive of the tweets generated by the accounts during the first 100 days of the new administration, that had not been deleted by 1 May 2017 (when our data collection took place). We collected originally authored tweets and replies from the 102 accounts, but did not collect retweets where the authors did not add new content. We collected a total of 43,569 unique tweets from the 102 Rogue Twitter accounts.

Coding

The research team developed an initial tweet coding scheme based on the extant literature on social movements. We decided to treat each tweet as a codable unit. We applied the coding scheme to each tweet. We began with three ‘levels’ that would be applied to each tweet and to each question response. Each was given at least one MACRO level code, a MESO level code, and a MICRO level code, understanding that, in many circumstances ‘none’ would likely appear in one (or more) of the levels.

The MACRO level of coding is about official political mechanisms that afford opportunities for participating in politics. We coded for this when a tweet appeared to be about directly influencing institutions and decision-makers who exercise direct authority and hold immediate, structural power. For example, tweets that make an appeal directly to a senator to vote a particular way on a confirmation hearing would be coded as such. When we found language that looked like an attempt to influence an official political process, we recorded it as a sub-category based on whether the process was within a government agency, the legislative branch, a state government, or some other institution. Each tweet received one MACRO level code (including the possibility of ‘none’).

The MESO level of coding deals with the more unofficial mechanisms that make collective action possible (mobilizing mechanisms). These fell into two categories: 1) tactics or ‘repertoires’ and 2) organizational structures. Tactics are the ‘how’ of getting people to ‘join’ the cause and mobilizing particular resources or action. For example, tactics can include calls for donations to affiliated groups or mobilizing protest. Organizational structures are about who is part of the movement and the organization of the movement itself. For example, some tweets discuss whether influential individuals are part of the resistance movement. Each tweet received one MESO level code (including ‘none’).

The MICRO level of coding determines whether a tweet is stating a particular ideological frame or providing content that functions as a framing process. For example, a frame might be ‘The Trump administration is anti-science.’ In this case the frame is about ‘the Trump administration’ and the frame itself is that the administration is anti-science. For all frames, we coded who the frame was about and what the particular position was. However, frames must also be supported by evidence. Tweets that provide information to support a particular frame were coded as ‘framing processes.’ For example, a tweet might link to a news article about the removal of scientific information from a government agency Web site. Thus, we would code that there was information sharing (as a framing process) in the form of news links. Each tweet received one MICRO level code (including ‘none’).

As the research team proceeded through the application of the coding schema to the corpus, we noted a number of a second, third, and in some cases, fourth-level codes were necessary but not present in the initial coding structure. As a result, the final coding structure was derived both before and during data analysis. In this sense, this content analysis should be considered a directed content analysis (Hsieh and Shannon, 2005).

 

++++++++++

Data analysis and results

Overview of tweet activity

The corpus we collected includes 43,569 tweets, produced by 102 accounts. Within the corpus there are 22,542 (51.7 percent) tweets and 21,027 (48.3 percent) replies. Accounts produced an average of 427 tweets during this time frame with a standard deviation of 756. The most active account, @ALT_USCIS, generated 5,584 tweets during the first 100 days of the Trump administration, where the least active, @AltDofEnergy, only generated a single tweet.

To better understand the interactions occurring within the corpus, we parsed the use of mentions and use of hashtags across the entire corpus. As shown in Table 1, perhaps not surprisingly, @realdonaldtrump appears as the most mentioned account in the entire corpus and @POTUS as the third. As the Rogue Twitter accounts organized their efforts around direct opposition to the administration’s efforts, it is perhaps not surprising to see the messaging directed at the president. We also see that the network made frequent mention of @GOP, the Twitter account of the Republican party.

 

Table 1: The 20 most frequently mentioned accounts.
AccountNumber of mentions in corpus
@realdonaldtrump1,221
@alt_labor734
@POTUS478
@ALT_USCIS431
@AltYelloNatPark357
@AltUSPressSec343
@alt_fda341
@AltHomelandSec279
@alt_lawyer224
@altUSEPA210
@AltNatParkSer202
@AltWHKitchen197
@jasoninthehouse191
@ActualEPAFacts188
@EPA183
@Alt_Interior169
@alt_jabroni169
@GOP146
@RogueNASA143
@BadHombreNPS142

 

However, we also note that other Rogue-related accounts are frequently mentioned in the corpus. This frequent mentioning of other alt accounts took a couple of different forms. First, the accounts often converse with each other. For example, Figure 1 shows a set of sequential remarks in which Rogue accounts referenced other Rogue accounts. Second, Rogue accounts also often mentioned each other as individuals to follow. As shown in Table 1, when we look across all mentions in the corpus, we see that these accounts mention each other in 4,360 tweets. Or to put it another way, of the 43,569 tweets and replies, roughly 10 percent are tweets or replies to other rogue related accounts, suggesting some degree of network and visibility building among them.

 

Rogue accounts referencing each other
 
Figure 1: Rogue accounts referencing each other.

 

Hashtags are another important way that this network of contention builds its identity and connections across Twitter. Using hashtags individuals can both discover others who are tweeting about similar subject matter and also potentially recruit for one’s cause or raise awareness of a particular event. As part of our analysis, we parsed the most commonly used hashtags from the corpus. Table 2 provides a frequency count of the most commonly occurring hashtags within the corpus. What emerges from this list is a prominent use of hashtags around the concepts of resistance, climate change, defense of science and science marches, immigration policies undertaken by the Trump organization, and potential connections between Trump and Russia.

 

Table 2: Most frequently appearing hashtags.
HashtagNumber of appearances in corpus
#resist2,206
#Resistance521
#climatechange277
#climatefacts224
#ScienceMarch171
#RogueRangers157
#science144
#climatechangeisreal112
#sciencematters107
#CAH106
#TheResistance106
#MarchForScience99
#MuslimBan96
#NoUSDAblackout95
#ScienceNotSilence95
#AMJoy85
#Trumprussia83
#altgovt76
#WeAreAltGov63
#altGov59

 

Contentious politics

To analyze the data in relation to CPM, we applied the codebook, as described in our Methods section, to two samples of the tweets: a random sample of 500 tweets taken from the corpus and the top 500 most retweeted tweets. We chose to use samples because of the in-depth nature of our codebook and because we believed there may be value in exploring the kinds of content that appear in a random selection of tweets versus the kinds of content that are most popular. We chose 500 for each selection as it is roughly one percent of the overall corpus of tweets, and it provides a sampling margin of error of five percent at a 95 percent confidence level. Below, we detail the kinds of content that emerged at different coding levels, as well as differences between the kinds of content that appear in a random sample versus popular retweeted content.

Macro level content. A tweet contains macro level content when it is about directly influencing institutions and decision-makers who exercise direct authority and hold immediate, structural power. For example, tweets that make an appeal directed to a specific senator to vote a particular way on a confirmation hearing, or in regards to the passage of a particular law, would be coded as an appeal to Congress. Table 3 provides a breakdown of the frequency rate of direct appeals and attempts to directly influence formal institutions and decision-makers in both the 500 most retweeted tweets and a random sample of 500 tweets form the corpus.

 

Table 3: Macro codes and frequencies.
Macro codeTop 500 RTsRandom 500
Government agency73
Local politics01
Media22
None431458
Other13
Politicians5027
Press04
State governments42
Trump50

 

The most prominent trend is that there are very few (comparatively) direct appeals at the macro level. In both samples, more than 80 percent of the tweets contain no macro level appeal. This suggests that this network is not using Twitter to try to make direct appeals to decision-makers. We do see a few direct appeals to politicians during moments of critical votes. For example, in the wake of a federal judge ordering then-EPA nominee Scott Pruitt to release thousands of e-mail messages, which were thought to potentially reveal details on his ties to the fossil fuel industry, @RogueNASA tweeted: ‘The Senate should be responsible and postpone his confirmation until the e-mails are released to the public.’

While the network frequently mentioned Trump, it very rarely made direct demands to him. This is seen in both the popularly retweeted tweets produced by the Rogue accounts as well as the random sample of tweets from the corpus. However, one such tweet from @AltStateDpt garnered about 4,000 retweets: ‘RT to send him our request. We the people request @realDonaldTrump submit your resignation, effective immediately.’

Meso level content. A tweet contains meso level content when it provides details about how to get people to join the cause or mobilizing particular resources or actions. For example, tactics can include calls for donations to particular groups or mobilizing citizens for a protest. Organizational structures are about who is part of the movement and the organization of the movement itself. Table 4 provides a breakdown of the frequencies of different meso level codes in the top 500 most retweeted items and the random sample of 500 tweets from the corpus.

 

Table 4: Meso codes and frequencies.
Meso codeTop 500 RTsRandom 500
None231409
Organizational structures — Government workers01
Organizational structure — Influential allies012
Organizational structures — Other network building105
Organizational structures — Politicians12
Tactics — Call to action6729
Tactics — Dunking on Trump4036
Tactics — Insider information1506
Tactics — Other10

 

Meso level content was more likely to appear in the top 500 retweeted tweets than in randomly selected tweets. We will briefly discuss four forms of content working at the meso level: organizational structures to promote recruitment, calls to action, ‘dunking on Trump’ [3], and the sharing of insider information.

Though they did not appear as frequently as other meso level codes, there were several attempts to recruit different actors to Rogue-related causes. For example, @BadlandsGonWild solicited other government employees to join the broad coalition of resistance with the following tweet:

 

Rogue accounts recruiting government employees
 
Figure 2: Rogue accounts recruiting government employees.

 

In addition to recruiting government employees and appealing to politicians, we also saw tweets that appealed to celebrities or other potential influencers within the network. For example, @AltMtRainier appealed to the band Pearl Jam to get support in relation to the Seattle March for Science, ‘@PearlJam We would love to get your support, @SciMarchSeattle.’

Calls to action appeared frequently in the corpus and were often retweeted. Often, these tweets were rallying individuals to contact Congressional members during particular votes. For example, during the confirmation hearings for Betsy DeVos for U.S. Secretary of Education, @NatParkUndrgrnd tweeted, ‘@GOP is trying to sneak Betsy DeVos vote through tomorrow at 6:30am. Need to tweet, email, call and show up! Spread the word! #DumpDeVos.’ Many shared information about upcoming or currently occurring protests. For example, @ALT_DOJ tweeted, ‘80,000-people strong #MoralMarch protest in Raleigh #resist #moralresistance.’

Interestingly, one account used the tactic of sharing a supposed inside view of the Trump White House. The account @RoguePOTUSStaff claimed to be ‘White House staffers, working at various levels, operating in secrecy to reveal hidden truths of the Trump administration to the American people’ (see Heffernan, 2017). It routinely shared information represented as coming from inside the White House itself. For example, during the protest immediately following the inauguration, the account tweeted, ‘Suspicion that Bannon urging POTUS to egg on protests, then call in National Guard to disperse, as demonstration of power. #resistpeacefully.’ This message was retweeted over 13,000 times. The content from this Twitter account portrayed the internal workings and conflicts within the White House, typically highlighting corruption or conflicts of interest. For example, the account tweeted, ‘We can confirm that POTUS is encouraging people to buy his brands in order to have the “right look” to work for his administration’ and ‘POTUS irritated that sev. members of Patriots team will not attend WH visit. Saying he will lobby owner to cut them from the team.’ Though they only represent a small portion of the overall corpus and rarely appear in the random sample, this type of tweet content appears frequently when looking at the most popularly retweeted content.

Finally, another common theme at the meso level was an emergent theme that we describe as ‘dunking on Trump.’ These tweets commonly made fun of or belittled the Trump administration. For example, in the wake of the administration’s gaffe in referring to the ‘Bowling Green Massacre’ (an event that did not occur; see Schmidt and Bever, 2017), @ALT_DOJ wrote, ‘If you died in the Bowling Green Massacre RT’.

Micro level content. Finally, a tweet contains a micro level code when it states an ideological frame or when it provides content that functions as a framing process. Table 5 provides a frequency count of the different micro level codes that we identified in the tweets.

 

Table 5: Micro codes and frequencies.
Micro codeTop 500 RTsRandom 500
Direct response1180
Frames — About Alt: Correcting misinformation03
Frames — About Alt: Defending democracy14
Frames — About Alt: Opposing anti-science16
Frames — About Alt: Other112
Frames — About Alt: Protecting environment10
Frames — About Senate: Weak10
Frames — Democrats: Weak10
Frames — Media: Incomplete information76
Frames — Media: Other21
Frames — Trump administration: Threat to democracy100
Frames — Trump administration: Corruption6517
Frames — Trump administration: Evil10
Frames — Trump administration: Explaining13612
Frames — Trump administration: Hypocrisy197
Frames — Trump administration: Incompetent106
Frames — Trump administration: Lazy50
Frames — Trump administration: Other01
Frames — Trump administration: Policy1624
Frames — Trump administration: Racism30
Framing — Information sharing: Event3118
Framing — Information sharing: Link813
Framing — Information sharing: News4548
Framing — Information sharing: Other74
Framing — Information sharing: Photography25
Framing — Information sharing: Policy2523
Framing — Information sharing: Science732
Framing — Other1626
Framing — Humor5339
Framing — Solidarity617
None96

 

The first element of note about the micro-level codes is that we see many of the tweets we collected are in fact responses to other tweets. Direct replies comprised 180 of the 500 randomly selected tweets. However, we also see that this form of content was not particularly popular, as only one direct reply made it into the top 500 most retweeted tweets.

The most commonly occurring political frames in the tweet corpus were those about the Trump administration. Frames encourage a particular interpretation of events. They are rhetorical lenses that offer us a vision of the world. Within the frames about the Trump administration, we see explanations of the administration’s actions as a prominent mechanism. For example, @RoguePOTUSStaff tweeted, ‘What many aren’t understanding is that power is being consolidated behind POTUS. Investigations won’t matter if he can get ahead of them.’ Not surprisingly, given the impetus for the Rogue accounts’ formation, the frames about the Trump administration were never positive and ranged from attempting to explain particular actions, to normatively positioning the administration as corrupt, evil, incompetent, lazy, racist, and/or a threat to democracy.

Less prominent were tweets that framed the alt movement itself, media, Senate, and Democratic party. When discussing the alt movement, tweets often described its motivations and identities quite broadly. For example, @RgouePOTUSStaff tweeted, ‘Just so we’re clear, it’s not Obama behind the leaks. Approx 50 percent of staffers are actively participating one way or another. #Iamtheleak.’ This framing suggests a broad group of individuals within the administration actively working as part of resistance. Frames about the media typically were critical, suggesting the media provided incomplete information about events occurring during the Trump presidency. For example, @RoguePOTUSStaff tweeted, ‘The media has been beaten into submission. They are too afraid to lose access again, so they aren’t pressing when they know it would hurt.’ Interestingly, potential political allies do not escape frames that critique; both the Senate and the Democractic party were discussed as being weak in opposing Trump. For example, @NotAltWorld stated, ‘Silence from @TheDemocrats this past week is deafening. What’s the point of a two party system if they can’t even be vocal in opposition.’

Frames are differentiated from framing processes, which are the means by which a frame gets reinforced through the sharing of evidence. The sharing of information is a key facet of the Rogue accounts’ activities on Twitter. We find a few major categories of information that this network shared. Of note, however, this network drew attention to the repression of scientific information via tweets as a process of information sharing. For example, @NotAltWorld tweeted, ‘The EPA has started the process of wiping all information collected during Obama’s presidency #sciencematters’ and, later, ‘The US Department of Agriculture agency has wiped all inspection reports & enforcement records from its website.’ Interestingly, we see a higher number of tweets sharing scientific information in our random sample than in the most popularly retweeted tweets group, suggesting that while scientific information was being shared by this network, it was not as popular as other kinds of content.

s

 

++++++++++

Discussion

A group of people with varying relationships to federal government agencies and the White House assembled on Twitter in early 2017 to oppose the Trump agenda. This movement brought together individuals who are not typically thought of in the context of social movements: those who were (or at least purported to be) already inside the government. This novel network [4] engaged in collectively framing and defining themselves, their grievances, and their motivations, as well as mobilizing their networks and new repertoires of contention. This Rogue Twitter Movement is an exemplar of how online social movements engage in framing processes and organizing via an online platform, highlighting the significance of incorporating multidisciplinary approaches to studying activism.

Through an analysis grounded in the CPM framework, we found that most of the network’s Twitter activity concentrated on the micro and meso levels. The micro level activity (frames and framing processes) was predominantly aimed at three content areas: first, framing Trump and his administration’s actions; second, framing their own network’s identity as the resistance and an information-sharing network; and third, framing the media as frequently failing to adequately capture the scope of the problems and threats presented by the new administration. At the meso level, we see the account holders engaged in building mobilizing structures and carrying out a variety of online repertoires of contention such as trying to get the public involved in protest, encouraging citizens to contact their Congressional leaders to vote against key political appointments, calling for other government employees to become part of the wider ‘resistance’ movement, sharing insider information, and lastly, ‘dunking on Trump.’ In short, we see that this network was highly engaged in resisting the new administration, but resisting in ways atypical to traditionally understood social movements. We now turn to our specific research questions, to unpack how our findings illustrate the limits of traditional conceptualizations of social movements.

Rogue Twitter is a social movement

We believe that the Rogue Twitter movement is a networked/online social movement. From the perspective of current literature (Earl, et al., 2015; Bennett and Segerberg, 2012), the movement lives online, and Twitter is its primary organizing agent. It does not have large off-line organizations supporting it nor identifiable leaders and has few gatekeepers (Tilly, et al., 2019; Earl, et al., 2015). Rogue Twitter’s goals seem to be predominantly about raising public awareness and information-sharing; when the network does argue for policy change, it tends to be on narrowly constructed issues (Tilly, et al., 2019).

Individuals in this network organized around the exchange of personally relevant information across fluid, evolving social networks (Earl, et al., 2015). Moreover, through examination of their public discourse on Twitter and using a more traditional theoretical approach we can support this claim further as we uncovered the ways in which they challenge authorities through the symbolic construction and maintenance of collective frames (Melucci, 1996; Benford and Snow, 2000) and the ensemble of contentious repertoires (McAdam, et al., 2001).

Twitter served as an organizing tool to create meaning for the movement

Protests today can happen across off-line and online contexts, with each movement using ICTs based on their particular necessities and context. The Rogue Twitter network is an online movement composed of people (at least some of whom work within the U.S. government) who oppose the Trump administration’s actions and policies. Its members are particularly vulnerable to repression and punishment, such as job loss; therefore, they need a space where they can anonymously and asynchronously come together to express dissent. Dissenting groups usually rely on safe and alternative mechanisms to organize and engage in discussion (Earl and Kimport, 2011; Carty and Reynoso Barron, 2019).

The Rogue Twitter movement found Twitter to be a protected space to organize. In fact, when the Trump administration tried to compel Twitter to reveal who was behind @ALT_USCIS [U.S. Customs and Immigration], which was harshly critical of administration immigration policy, Twitter sued to protect its users; eventually, the Trump administration let the matter drop (Wong, 2017). More than a ‘new flyer’ used to disseminate information, Twitter served as a ‘communicative system that enabled and framed patterns of connection’ [5].

Rogue Twitter mobilized using two main Twitter features: its networked structure and its capability to construe and convey complex meanings. Twitter’s networked character facilitated personalized public engagement, which in turn enabled an organizational dynamic typical of online movements, in which information exchange is a prominent part of the organizational structure (Bennett and Segerberg, 2012). For example, the Rogue Twitter movement prominently promoted participation in the Science March, the March for Women, and marches against restrictive immigration policies.

In the same way, hashtags are a powerful feature that have communicative utilities crucial for the collective mobilizing and sense-making processes. Zappavigna (2015) indicated that hashtags, more than facilitating conversation, have a classificatory function that helps indicate topics or subjects (by labeling content), and they also have an interpersonal function that helps construe attitudes towards those topics and enact relationships (indicating evaluative stance). The development and reinforcement of the idea of ‘resistance’ became a major rallying cry for this network.

Thus, members of the Rogue Twitter movement convened on Twitter because it is a useful social organizing tool that offers capabilities central for this movement. It provided the capacity to engage remotely and anonymously to enact social affiliations and collective meaning-making.

As the CPM suggests, building the organization and framing are still crucial for social movements

The CPM is the product of decades of social movement research that enables a holistic study of social movements by bringing together structural and cultural perspectives. Implementing it in our study of the Rogue Twitter movement allowed for an in-depth three-level analysis of this movement’s context, organizing, and collective sense-making.

Our insights would not have been possible only by counting hashtags and mentions or through a purely inductive approach. We observed how the Rogue Twitter movement framed the structures that drove their organizing, their practices, and repertoires, and the way their collective ideas and identities mediated their actions. These insights allowed us to conclude that the Rogue Twitter movement is more than an online campaign; it is a movement simultaneously in two early stages of the social movement life cycle: social ferment and popular excitement (Blumer, 1969; Della Porta and Diani, 2006).

During social ferment, there is widespread discontent about social and political conditions. Our macro and micro level analysis showed the account holders were not directly appealing to authorities, even though they were frequently mentioning Trump. Instead, most activity was directed at making sense of Trump’s government by explaining his action as corrupt, evil, incompetent, lazy, racist, and a threat to democracy. This activity indicates they were collectively defining the meaning of their grievances by framing the political environment as closed to participation and prone to repression. They were also engaged in framing the movement as the #resistance and explaining why they must mobilize.

At the same time, the Rogue Twitter movement was in the popular excitement phase; members had started recognizing each other based on their shared framing of the issues and were engaged in initial activities to grow their mobilizing structure and in demonstrating dissent. Our analysis of mentions, for instance, showed they talked to each other to recruit and expand their network. The meso level analysis expanded this assessment and demonstrated that they carried out other mobilizing activities aimed at strengthening their mobilizing structure. For example, they engaged with influential allies (i.e., celebrities and politicians) who could legitimize the movement and procure support through resources and followers. In addition, data showed they engaged in a variety of repertoires of contention; notable examples include:

In this study, by examining the discourse on Twitter and its implications for how the movement conceives social structures and internal processes, we demonstrate that online movements also engage in constructing and negotiating collective action frames.

Reconceptualizing social movements

In this study, we witnessed the process of collective framing happening publicly during the early stages of an online movement; we saw the #resistance literally being hashed out on Twitter amongst Rogue Twitter account holders. Concerning the emergence and development of social movements, the original framework proposed by Della Porta and Diani (2006) argues that social movements evolve in a linear fashion where social ferment takes place first, then popular excitement. Our data suggest that social ferment and popular excitement might happen simultaneously (or overlap) in online movements and that a movement’s collective interpretations shape the ongoing relationship between social ferment and popular excitement. However, we note that this notion requires further examination in future studies.

Our data also highlighted two forms of political repertoires online. The first one is user-generated political satire as a contentious tactic. Satire is a mechanism to ridicule and criticize those in powerful positions; it has the potential of attracting attention, persuading, offering a way into complex issues, dissolving social barriers, and encouraging message sharing (Chatoo, 2018). Satire has traditionally been performed by media professionals, notably by television shows like Saturday Night Live and magazines like The Onion. Over the last decades, however, there has been an increase in user-generated satirical commentary on social media. Limited research in this area shows counterintuitive findings. On one side, the portrayals of authorities in user-generated satire do not seem to affect individuals’ attitudes toward them. Thus, it may not be a useful tool to change attitudes (Rill and Cardiel, 2013). On the other, however, having the capacity to respond to satire directly may give authorities an advantage, as their responses promote a favorable attitude towards them (Becker, 2018).

We see user-generated satire as a new type of repertoire that speaks of social movements’ ability to create their own content, the affordances of social media like Twitter, the movement’s nuanced understanding of authorities, and the new capacity authorities have to respond to mockery. Thus, we think that it is crucial to continue to explore satire as a form of contentious activity and its potential intended and unintended consequences. Perhaps it can be seen as a way to coalesce a burgeoning movement behind commonly held humor, shedding light into online movements’ repertoires and framing processes dynamics.

The second repertoire is the emergence of whistleblowing as a contentious activity. Whistleblowing has been conceptualized as the broad citizenry disclosure of illegal, immoral, or illegitimate practices predominantly related to government corruption, wrongdoings, and misconduct from an insider; which is more likely to happen in places with weak institutional anti-corruption practices (Su and Ni, 2018). Research on whistleblowing and digital ICTs indicates the affordances of digital media provide whistleblowers the ability to go global with their message reaching a broad audience directly, thus having the potential of inspiring and stimulating movements (Carty and Reynoso Barron, 2019). While the exposure of an organization’s wrongdoing by an insider is not a new phenomenon, this study showed a new way of doing it, in which the insider circumvents the editorial control of the mainstream media or an intermediating organization (e.g., Wikileaks).

Moreover, it could reveal a new type of social movement composed of government insiders. In this case, account holders remained anonymous; thus, it is difficult to establish which of these accounts were operating from within the government. However, they certainly took on that identity, and the content they shared suggests they were indeed government insiders. Undoubtedly, this type of subversive activity within the state is interesting, as generally, social movements are external from the government. Recent research suggests rebellious activities from within the regime are not isolated in the digital age. There are documented cases where government members acted in subversive ways and refused to follow orders from the state. For example, in Egypt, the Minister of Telecommunications was “less than cooperative” when the regime ordered him to block access to social media sites like Twitter and Facebook during the 2011 revolution (Wilson, 2015).

Therefore, we recommend further research answering questions about the strategic use of whistleblowing by social movements and the role and motivations of government insiders in digital social movements. Doing so can reveal the new dynamics of online contentious activity and framing mechanisms.

Limitations

While we have attempted to gather all tweets generated by the Rogue accounts during the first 100 days of the Trump presidency, because of the time delay associated with our collection, it is possible that some tweets were deleted by the accounts. Further, while we have used the first 100 days of the Trump presidency as our time range, ultimately, the first 100 days of a presidency is an entirely artificially constructed time period by which to measure the efficacy of a presidency. Lastly, for reasons of time, our coding methodology was not applied to all tweets, but rather a random sample of 500 tweets and the top 500 retweeted tweets from the corpus. While we feel our methodology provides representative insights into the corpus and what was popularly retweeted within the corpus, it cannot be said to be entirely comprehensive of potential outliers. Future work using machine learning tools could provide deeper analysis and additional insights from this corpus. Because of these potential limitations in our approach, future work should focus on triangulating our conclusions in other online social movements.

 

++++++++++

Conclusion

As social media platforms are increasingly incorporated into the public sphere of politics, it is likely that we will see online communities and social movements form, hoping to influence government actions. Using the Contentious Politics Model, we demonstrated that the Rogue Twitter network was, in fact, an online social movement operating in opposition to the Trump administration, especially using meso and micro levels of engagement. Uniquely, members of this social movement claimed membership within the government.

We anticipate the rise of additional online social movements, with similarities to Rogue Twitter, in the future, across various social media platforms. What remains to be traced, however, is the degree to which these movements use the platforms to successfully influence the actions of the government as a whole, and the degree to which this functions as a public spectacle. We believe that the Rogue Twitter accounts constituted a social movement, though perhaps one that does not cleanly resemble previous models. Future social movements may also stray further from that conceptualization. End of article

 

About the authors

Fatima Espinoza Vasquez is assistant professor at the School of Information Science at the University of Kentucky. Her research is at the intersections of sociotechnical systems, social movements, Latin America, social justice, and political participation.
Direct comments to: fatima [dot] espinoza [at] uky [dot] edu

Nicholas Proferes is assistant professor at the New College School of Social and Behavioral Sciences at Arizona State University. His research interests include users understandings of socio-technical systems such as social media, societal discourse about technology, and issues of power and ethics in the digital landscape.
E-mail: nicholas [dot] proferes [at] asu [dot] edu

Troy B. Cooper is assistant professor at the School of Information Science at the University of Kentucky. His research focuses on rhetoric, visual culture, and consumerism.
E-mail: troy [dot] cooper [at] uky [dot] edu

Shannon M. Oltmann is associate professor in the School of Information Science at the University of Kentucky. Her research interests include censorship, intellectual freedom, information policy, public libraries, privacy, and qualitative research methods.
E-mail: shannon [dot] oltmann [at] uky [dot] edu

 

Notes

1. We note that one of the most popular accounts, then called @AltNatParkSer (now @NotAltWorld), was started in 2015 by a British manwell before Trump’s inauguration. This necessarily complicates discussion of the timeline and account ownership.

2. Alimi, 2018, p. 410.

3. ‘Dunking’ in social media, particularly on Twitter, refers to making a joke or sarcastic comment at someone’s expense, often by quoting-tweeting them (see Schwedel, 2017).

4. In a previous paper, we discussed in more detail the ways in which these accounts collaborated and formed a loose, ad hoc network. See S.M. Oltmann, T. Cooper, and N. Proferes, 2020. “How Twitter’s affordances empower dissent and information dissemination: An exploratory study of the rogue and alt government agency Twitter accounts,” Government Information Quarterly, volume 37, number 3, https://doi.org/10.1016/j.giq.2020.101475.

5. Pond and Lewis, 2019, p. 228.

 

References

E. Alimi, 2018. “Repertoires of contention,” In: D. Della Porta and M. Diani (editors). Oxford handbook of social movements. Oxford: Oxford University Press, pp. 410–420.
doi: https://doi.org/10.1093/oxfordhb/9780199678402.013.42, accessed 18 May 2021.

C. Atton (editor), 2015. Routledge companion to alternative and community media. London: Routledge.
doi: https://doi.org/10.4324/9781315717241, accessed 18 May 2021.

A.B. Becker, 2018. “Live from New York, it’s Trump on Twitter! The effect of engaging with Saturday Night Live on perceptions of authenticity and the salience of trait ratings,” International Journal of Communication, volume 12, pp. 1,736–1,757, and at https://ijoc.org/index.php/ijoc/article/view/8626, accessed 18 May 2021.

R.D. Benford and D.A. Snow, 2000. “Framing processes and social movements: An overview and assessment,” Annual Review of Sociology, volume 26, pp. 611–639.
doi: https://doi.org/10.1146/annurev.soc.26.1.611, accessed 18 May 2021.

W.L. Bennett and A. Segerberg, 2012. “The logic of connective action: Digital media and the personalization of contentious politics,” Information, Communication & Society, volume 15, number 5, pp. 739–768.
doi: https://doi.org/10.1080/1369118X.2012.670661, accessed 18 May 2021.

S.I. Bhuiyan, 2011. “Social media and its effectiveness in the political reform movement in Egypt,” Middle East Media Educator, volume 1, number 1, pp. 14–20, and at https://ro.uow.edu.au/meme/vol1/iss1/3/, accessed 18 May 2021.

B. Binkowski, 2017. “Twitter’s alts and rogues,” Snopes.com (1 June), at https://www.snopes.com/news/2017/06/01/alts-and-rogues/, acceessed 7 July 2018.

H. Blumer, 1969. “Collective behavior,” In: A.M. Lee (editor). Principles of sociology. Third edition. New York: Barnes & Noble, pp. 67–121.

V. Carty and F.G. Reynoso Barron, 2019. “Social movements and new technology: The dynamics of cyber activism in the digital age,” In: B. Berberoglu (editor). Palgrave handbook of social movements, revolution, and social transformation. Cham, Switzerland: Palgrave Macmillan, pp. 373–397.
doi: https://doi.org/10.1007/978-3-319-92354-3_16, accessed 18 May 2021.

H.S. Christensen, 2011. “Political activities on the Internet: Slacktivism or political participation by other means?” First Monday, volume 16, number 2, at https://firstmonday.org/article/view/3336/2767, accessed 18 May 2021.
doi: https://doi.org/10.5210/fm.v16i2.3336, accessed 18 May 2021.

D. Della Porta and M. Diani, 2006. Social movements: An introduction. Second edition. Malden, Mass.: Blackwell.

J. Earl and K. Kimport, 2011. Digitally enabled social change activism in the Internet age. Cambridge, Mass.: MIT Press.
doi: https://doi.org/10.7551/mitpress/9780262015103.001.0001, accessed 18 May 2021.

J. Earl, J. Hunt, R.K. Garrett, and A. Dal, 2015. “New technologies and social movements,” In: D. Della Porta and M. Diani (editors). Oxford handbook of social movements. Oxford: Oxford University Press, pp. 355–366.
doi: https://doi.org/10.1093/oxfordhb/9780199678402.013.20, accessed 18 May 2021.

M. Gidda and Z. Schonfeld, 2016. “Donald Trump’s threat to press freedom: Why it matters,” Newsweek (12 November), at https://www.newsweek.com/2016-election-donald-trump-press-freedom-first-amendment-520389, accessed 18 May 2021.

V.S. Greene, 2019. “‘Deplorable’ satire: Alt-right memes, white genocide tweets, and redpilling normies,” Studies in American Humor, volume 5, number 1, pp. 31–69.
doi: https://doi.org/10.5325/studamerhumor.5.1.0031, accessed 18 May 2021.

V. Heffernan, 2017. “Why @RoguePOTUSStaff is the best thriller on the Internet,” Fast Company (10 March), at https://www.fastcompany.com/3068332/why-roguepotusstaff-is-the-most-important-work-of-fiction-on-the-internet, accessed 18 May 2021.

H.-F. Hsieh and S.E. Shannon, 2005. “Three approaches to qualitative content analysis,” Qualitative Health Research, volume 15, number 9, pp. 1,277–1,288.
doi: https://doi.org/10.1177/1049732305276687, accessed 18 May 2021.

H. Johnston and J.A. Noakes (editors), 2005. Frames of protest: Social movements and the framing perspective. Lanham, Md.: Rowman & Littlefield.

E. Khazraee and J. Losey, 2016. “Evolving repertoires: Digital media use in contentious politics,” Communication and the Public, volume 1, number 1, pp. 39–55.
doi: https://doi.org/10.1177/2057047315625076, accessed 18 May 2021.

M.M. Kircher, 2017. “The National Park Service’s Twitter goes rogue, gets banned from tweeting,” New York Magazine (21 January), at http://nymag.com/selectall/2017/01/national-park-service-tweets-about-trump-inauguration-crowd.html, accessed 18 May 2021.

H. Kriesi, 2004. “Political context and opportunity,” In: D.A Snow, S.A. Soule, and H. Kriesi (editors). Blackwell companion to social movements. Oxford: Blackwell, pp. 67–90.
doi: https://doi.org/10.1002/9780470999103.ch4, accessed 18 May 2021.

H. Kriesi, 1995. “The political opportunity structure of new social movements: Its impact on their mobilization,” In: J.C. Jenkins, and B. Klanndermans (editors). The politics of social protest: Comparative perspectives on states and social movement. Minneapolis: University of Minnesota Press, pp. 167–198.

D. McAdam, S. Tarrow, and C. Tilly, 2001. Dynamics of contention. New York: Cambridge University Press.
doi: https://doi.org/10.1017/CBO9780511805431, accessed 18 May 2021.

D. McAdam, J.D. McCarthy, and M.N. Zald (editors), 1996. Comparative perspectives on social movements: Political opportunities, mobilizing structures, and cultural framings. New York: Cambridge University Press.
doi: https://doi.org/10.1017/CBO9780511803987, accessed 18 May 2021.

J.D. McCarthy, 1996. “Constraints and opportunities in adopting, adapting, and inventing,” In: D. McAdam, J.D. McCarthy, and M.N. Zald (editors). Comparative perspectives on social movements: Political opportunities, mobilizing structures, and cultural framings. New York: Cambridge University Press, pp. 141–151.
doi: https://doi.org/10.1017/CBO9780511803987.008, accessed 18 May 2021.

A. Melucci, 1996. Challenging codes: Collective action in the information age. New York: Cambridge University Press.
doi: https://doi.org/10.1017/CBO9780511520891, accessed 18 May 2021.

S.M. Oltmann, T. Cooper, and N. Proferes, 2020. “How Twitter’s affordances empower dissent and information dissemination: An exploratory study of the rogue and alt government agency Twitter accounts,” Government Information Quarterly, volume 37, number 3, 101475.
doi: https://doi.org/10.1016/j.giq.2020.101475, accessed 18 May 2021.

P. Pond and J. Lewis, 2019. “Riots and Twitter: Connective politics, social media and framing discourses in the digital public sphere,” Information, Communication & Society, volume 22, number 2, pp. 213–231.
doi: https://doi.org/10.1080/1369118X.2017.1366539, accessed 18 May 2021.

L.A. Rill and C.L.B. Cardiel, 2013. “Funny, ha-ha: The impact of user-generated political satire on political attitudes,” American Behavioral Scientist, volume 57, number 12, pp. 1,738–1,756.
doi: https://doi.org/10.1177/0002764213489016, accessed 18 May 2021.

S. Schmidt and L. Bever, 2017. “Kellyanne Conway cites ‘Bowling Green’ massacre that never happened to defend travel ban,” Washington Post (3 February), at https://www.washingtonpost.com/news/morning-mix/wp/2017/02/03/kellyanne-conway-cites-bowling-green-massacre-that-never-happened-to-defend-travel-ban/, accessed 18 May 2021.

H. Schwedel, 2017. “‘Dunking’ is delicious sport,” Slate (4 December), at https://slate.com/technology/2017/12/dunking-is-delicious-and-also-probably-making-twitter-terrible.html, accessed 18 May 2021.

D.A. Snow and R.D. Benford, 1988. “Ideology, frame resonance, and participant mobilization,” International Social Movement Research, volume 1, pp. 197–217.

X. Su and X. Ni, 2018. “Citizens on patrol: Understanding public whistleblowing against government corruption,” Journal of Public Administration Research and Theory, volume 28, number 3, pp. 406–422.
doi: https://doi.org/10.1093/jopart/muy011, accessed 18 May 2021.

S. Tarrow, 2011. Power in movement: Social movements, collective action and politics. Revised & updated third edition. New York: Cambridge University Press.
doi: https://doi.org/10.1017/CBO9780511973529, accessed 18 May 2021.

S. Tarrow, 1996. “States and opportunities: The political structuring of social movements,” In: D. McAdam, J.D. McCarthy, and M.N. Zald (editors). Comparative perspectives on social movements: Political opportunities, mobilizing structures, and cultural framings. New York: Cambridge University Press, pp. 41–61.
doi: https://doi.org/10.1017/CBO9780511803987.004, accessed 18 May 2021.

C. Tilly, 1978. From mobilization to revolution. Reading, Mass.: Addison-Wesley.

C. Tilly, E. Castañeda, and L.J. Wood, 2019. Social movements, 1768–2018. Fourth edition. New York: Routledge.
doi: https://doi.org/10.4324/9780429297632, accessed 18 May 2021.

U.S. Office of Special Counsel, n.d. “Hatch Act overview,” at https://osc.gov/Services/Pages/HatchAct.aspx, accessed 18 May 2021.

S.L. Wilson, 2015. “How to control the Internet: Comparative political implications of the Internet’s engineering,” First Monday, volume 20, number 2, at https://firstmonday.org/article/view/5228/4204, accessed 18 May 2021.
doi: https://doi.org/10.5210/fm.v20i2.5228, accessed 18 May 2021.

J.C. Wong, 2017. “Government seeks to unmask Trump dissident on Twitter, lawsuit reveals,” Guardian (6 April), at https://www.theguardian.com/technology/2017/apr/06/twitter-lawsuit-anonymous-account-trump-alt-uscis, accessed 18 May 2021.

M. Zappavigna, 2015. “Searchable talk: The linguistic functions of hashtags,” Social Semiotics, volume 25, number 3, pp. 274–291.
doi: https://doi.org/10.1080/10350330.2014.996948, accessed 18 May 2021.

 


Editorial history

Received 17 December 2020; revised 5 March 2021; accepted 10 May 2021.


Creative Commons License
This paper is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

Going rogue: Reconceptualizing government employees’ contentious politics on Twitter
by Fatima Espinoza Vasquez, Nicholas Proferes, Troy B. Cooper, and Shannon M. Oltmann.
First Monday, Volume 26, Number 6 - 7 June 2021
https://firstmonday.org/ojs/index.php/fm/article/download/11631/10147
doi: http://dx.doi.org/10.5210/fm.v26i6.11631