Abstract
Scholars have addressed how the socially marginalized, individuals with fringe viewpoints, or the politically marginalized in authoritarian regimes use social media to organize or connect in secret. Yet intensifying partisan polarization and prejudice in the United States has made it necessary to study how mainstream partisans in liberal democracies use social media to organize in secret. This study explores why mainstream partisans in the United States—average Republicans or Democrats—organize in secret online and analyzes the unique functions of social media for political organizing amid contextual marginalization. Through interviews with group leaders and a digital ethnography of a secret Facebook group in the United States, I find that mainstream partisans use social media to form secret political groups when they are the minority in their local community and that the online secret group serves several unique functions for members. First, the group operates as a community of solidarity for those reticent to disclose their political beliefs and as a community of contention to criticize and revitalize the minority Party in their region. The group also operates as a community of practice, allowing members to learn and rehearse communication among like-minded others. These findings hold implications for how scholars study communication in a digital and polarized era and how practitioners gauge public opinion.
Digital technology has long afforded individuals the opportunity to obscure their identity and their communication from others (Baym & boyd, 2012). The recent creation of concealed online groups, like secret Facebook groups, now allows individuals to form online communities that are intentionally hidden from others in their online social network (e.g., Pruchniewska, 2019). Secret groups are not searchable or accessible without an invitation from an existing member, and confidentiality agreements within the group keep members from sharing content outside of the group’s boundaries. These types of online communities are principally different from other forms in that their identity, membership, and activity are all hidden from those outside.
Perhaps because of this opacity, the value of secret groups for political communication and organizing is currently undertheorized and underexplored in research on communication and social media. Previous scholarship has focused on the conceptual affordances of digital media for socially mediated publicness (Baym & boyd, 2012), for the formation of digital communities among marginalized social identities (Clark-Parsons, 2018; Jackson et al., 2018, 2020; Steele, 2016), fringe political groups and identities (Deem, 2019), or political organizing in authoritarian contexts (Lei, 2018; Pearce et al., 2018; Pearce & Vitak, 2016). In the wake of intensifying partisan polarization and prejudice in the United States (Iyengar et al., 2012), it becomes important to attend to how mainstream partisans in a liberal democracy—Republicans and Democrats in the United States—use online secrecy for political communication and organizing.
In response, I ask why mainstream partisans form and join secret groups online, what happens within the group’s boundaries, and how these groups affect members. I consider marginalization as context-dependent, focusing on progressives in majority conservative states in the southern United States. With this focus, I conduct in-depth interviews with five different leaders of secret political groups online and a digital ethnography of a secret group of progressives across a year of content. I find that mainstream secret communities may arise in response to two of the same conditions faced by those outside of the social mainstream (Clark-Parsons, 2018; Jackson et al., 2018, 2020) or those formed in authoritarian contexts (Pearce et al., 2018; Pearce & Vitak, 2016). I also find that secret groups on social media afford certain features of community: offering a community of solidarity for those reticent to disclose their political beliefs to others in their social network, a community of contention, both criticizing and reinvigorating minority efforts in their local communities, and a community of practice, in which individuals practice politics before taking their communication and action offline.
Because discourse in secret groups cannot be scraped or immediately accessed by scholars looking to study online communication (Freelon, 2019) or journalists looking to use social media as a form of vox pop (McGregor, 2019), getting a full picture of the public in a digital and polarized era will require a nuanced approach, one that relies on embedded observation (e.g., Deem, 2019) rather than acontextual text data (Rojas & Valenzuela, 2019). This study offers one such approach and a look at how social media has transformed where and how people display their politics and who feels it necessary to keep their politics a secret in the first place.
Organizing in Secret
People have used secrecy or strategically hidden their beliefs or actions (Simmel, 1906) for centuries. In the context of social and political organizing, there is a long, layered history of research on secret groups, organizations, and movements across a wide variety of disciplines. For instance, sociology and political science have explored secret societies (McVeigh et al., 2014; Simmel, 1906) and covert social movement networks (Crossley et al., 2012), while communication research has considered “hidden organizations” (Scott, 2013) and organizational “backchannels” (Herbst, 1994; Karpf, 2012). Despite their terminological differences, this work is guided by two important questions: Why do groups of people engage in secrecy? And, what do they gain or lose when they do?
The answers to both of these questions depend, in large part, on who engages. Marginalized social identities along lines of race/ethnicity or gender/sexuality have, and in many cases still are, required to organize in secret to avoid persecution or retaliation (Crossley et al., 2012). Fringe political groups have also historically relied on secret organizing and communication to carryout subversive agendas and engage in hate speech (McVeigh et al., 2014; Scott, 2013). There are also more contemporary examples in which the socially marginalized or those in the political extreme privately meet or communicate about shared experiences or beliefs (Clark-Parsons, 2018; Deem, 2019). This is especially true in authoritarian regimes where individuals with marginal or dissenting political beliefs may be monitored and/or persecuted for publicly speaking against the state (Lei, 2018; Pearce et al., 2018; Pearce & Vitak, 2016).
Research on secret organizing is closely tied to the idea of the “public sphere”—arguably the space where ideas are dubbed mainstream or marginal—and how individuals at the margins create “alternative” or “counter” spheres for discussion and action (Fraser, 1990; Herbst, 1994). Yet secret movements, networks, societies, and organizations are not entirely the “subaltern” groups of counterpublic theory (Fraser, 1990, p. 67; Squires, 2002), in which a group creates a space separate from and to challenge the mainstream. Secret groups are often organized and intentionally concealed within the mainstream itself; comprised of individuals who, in other contexts, would not seem extreme or marginal at all (e.g., Clark-Parsons, 2018; Pruchniewska, 2019). In other words, it is not that secret groups are created to subvert the mainstream, but to hide from it. Still, counterpublics reflect many of the same motivations to organizing as groups organized in secret. Like those who join or form counterpublics separate from the traditional public sphere, individuals in secret groups may form or join these groups because they feel stigmatized from those around them (Hudson, 2008) or because they fear persecution for their political identity or beliefs (Van Duyn, 2018).
Just as it is important to know why individuals organize in secret, it is also important to know what happens when they do. There are a number of ways secret groups may affect their members. Like other marginalized groups or networks, they allow members who are reticent to disclose their identity or beliefs a place to find like-minded others (Clark-Parsons, 2018; Pruchniewska, 2019), to build networks for support or advocacy (Jackson et al., 2018, 2020), and a place to share knowledge and practices (Shklovski & Valtysson, 2012). On one hand, this role may be especially important for members of secret groups because they are protected from outside criticism and therefore more likely to ask for support in the first place. On the other hand, there is evidence that insulating members from criticism may actually intensify existing attitudes and fear of those from whom they are hiding (Deem, 2019).
Secret groups also act upon the context in which they are formed. For instance, secrecy presents a unique tension between effective influence or advocacy and effective concealment, what sociologists call a “secrecy-efficiency trade-off” (Crossley et al., 2012). Secret groups, whose recruitment and activity are categorized by low-density and decentralized behavior, are challenged with bringing people into the group and engaging in activity without exposing the group or members’ identities. As a result, while secrecy can give localized power to members, it also limits what the group is able to exert in what Simmel (1906) would call the “obvious world” outside. It is also possible the support gained inside the secret group transfers to members own behavior outside of the group, giving them a sense of security that makes them more willing to unveil their secret to others (McKenna & Bargh, 1998; Van Duyn, 2018).
Secret Groups Online
Although secret organizing is an old practice, digital and social media have transformed both the ease of and extent to which secret groups are able to form. Scholars of digital media would call these “affordances,” or, very loosely, the features or tools the technology affords the user (Nagy & Neff, 2015). One of the affordances of digital media is exposure. At the most basic level, digital and social media have made it increasingly possible to find and organize others with whom one agrees (e.g., McKenna & Bargh, 1998; Pearce et al., 2018). Social media platforms have also made it easier to manage the visibility of one’s group membership altogether. For example, individuals can now control the visibility of their online group membership to their larger social networks by joining private, closed, or secret groups online (e.g., Clark-Parsons, 2018; Pruchniewska, 2019).
Beyond secret organizing, digital media has changed the structure that broader political groups may take. Where membership in political groups and organizations was once more formal and hierarchical (Skocpol, 2003), membership is now less defined. Members of online political groups are able to join or leave the group at their own whim and to construct and reinterpret the group’s mission and content on their own (Bennett & Segerberg, 2012). Without a reliance on in person meetings for coordination, online groups can include members that are not physically proximate or synchronous (Karpf, 2012), making coordination across various geographic communities and across time much easier.
The digital space has also transformed the function and activity of political groups. People who join or engage in politics online are more likely to participate in interpersonal forms of political action (e.g., sharing a political message online) than other acts like attending a meeting or volunteering for a campaign (Karpf, 2012; Wells, 2015). Participation in online political groups may also lead to more offline political action (Wojcieszak, 2009) or in person meet-ups (Karpf, 2012; Shklovski & Valtysson, 2012) by connecting people in person who would otherwise not be connected. In addition, the ease of participation and the temporal flexibility of online groups can encourage previously reticent or disengaged members of the public to participate in politics. For example, Gose and Skocpol (2019) document grassroots groups facilitated by social media, like closed groups on Facebook, which mobilized largely middle-class White women, many of whom were previously politically disengaged. In turn, these online groups were able to help revitalize in person party infrastructure within their local communities. Without their digital membership on Facebook, they argue, there would have been no participation offline.
This research shows that digital technology has allowed individuals to more easily find like-minded others, engaged individuals once reticent to participate in or discuss politics, and affected in person political action in addition to online political activity. Yet it is less clear how secret groups fit within this paradigm shift for political organizing. In other words, how do secret groups organize individuals around politics? What do secret groups online uniquely afford partisans and their political behavior?
Moreover, while some empirical work has addressed the role of political marginalization and secret organizing online in authoritarian contexts (e.g., Lei, 2018; Shklovski & Valtysson, 2012), it is not clear how partisans use online secret groups in liberal democracies where free speech is touted as a foundational principle and where state monitoring is, at the very least, less of a threat. This becomes especially important amid a growing body of evidence that mainstream partisans in the United States—Republicans and Democrats—may also have reason to manage the visibility of their communication (Van Duyn, 2018) and the visibility of their online activity. For instance, partisans in the United States are increasingly likely to favor and select individuals from their own party over individuals from the opposing party (Iyengar et al., 2012) and especially likely to both avoid and intentionally harm individuals from the opposite party (Lelkes & Westwood, 2017). As a result, mainstream partisans may have equal reason to manage their political beliefs and group membership online, despite the freedom they hold in expressing their beliefs, relative to authoritarian contexts. In the face of this intensifying partisan hostility in the United States, this study aims to account for how mainstream partisans use secret groups on social media to manage their political identity—why these groups form, what they afford members, and how they in turn affect members and their communities.
Methods
I use two qualitative methods that received institutional review board (IRB) approval in April 2017. Both methods used snowball sampling through previous work with secret political groups to locate and gain access to the groups included in this study. Sample selection in qualitative research is dependent on opportunity and availability (Flyvbjerg, 2006), which is further informed by the researcher’s own positionality. This is especially true for secret groups whose entire purpose is to conceal political identity from those outside of the group and especially those with different political beliefs. This makes access to and observation of these groups especially challenging for researchers looking to explore members’ experiences and the group’s function. I am a White woman who has previously studied Democratic organizations. As a result, the data to which I had access in this study are not representative of a range of partisanship in the United States and reflect only a slice of available cases as it has in previous work on protected or hidden communities (see Clark-Parsons, 2018; Wolfe & Blithe, 2015). Like other qualitative studies of social media and politics, however, my approach is intended to be “logically generalizable” (Luker, 2009) rather than statistically generalizable in that I aim to provide details about several cases rather than brief information about many. Because the groups I explore in this study are secret in nature and my IRB agreement reflects this sensitivity, I adhered to a strict protocol for maintaining the confidentiality of the groups and their members, which I explain in further detail below.
Interviews With Secret Group Leaders
I conducted in-depth interviews with leaders of five secret groups in the United States: one that organized a suburban county 1 in Florida in a congressional district that is heavily Republican, one that organized a mixed suburban-rural community in a congressional district in North Carolina that leans Republican, one that organized a suburban community in a Republican-leaning county in Texas, and two from Louisiana, one that organized an urban community in a Republican-leaning congressional district and one that organized a suburban community in a heavily Republican congressional district. Each interview was with either the group’s founder or with those currently helping to coordinate the group’s efforts and/or their communication (e.g., Facebook group, email). I gained access to these groups and their leaders through interpersonal contacts and snowball sampling through another secret group to which I had previously gained access. Each interview lasted between 30 and 60 min and followed a semi-structured interview guide that draws from research related to political organizing and discussion networks (Pearce et al., 2018; Van Duyn, 2018; Wojcieszak, 2009; see Appendix). Questions focused on reasons for the group’s formation, the group’s secrecy, the general membership, the perceived impact on members, and the group’s connection with the broader community. Each interview was recorded with the participant’s permission, transcribed, and the original recording deleted. I relied on Luker’s (2009) method combining extended case method with grounded theory to help extend existing knowledge of political organizing online while also building new theory. In my analysis, I looked for themes across each transcript. For quixotic reliability, I use the participant’s own words throughout my findings, using pseudonyms to assure confidentiality.
Digital Ethnography of a Secret Political Group
Drawing from research on counterpublics online (e.g., Deem, 2019), in addition to interviews with secret group leadership, I conducted a digital ethnography of an online secret group, which I will refer to as “TXP,” comprising political progressives in Texas. At the time of my observations, the group had approximately 27,000 members. I contacted administrators and let them know the intentions of my membership and study. I observed content and interactions posted in the group, including all posts, comments, and replies, from January 2018 to January 2019. As I observed content, I recorded my general observations in a document separate from the platform, taking care not to remove any specific wording, group names, or identifiers from the group. This approach meant that my observations often recorded the tone, topic, number of interactions associated with each post or reply, without recording any full statements.
I chose to conduct a digital ethnography for several reasons. Because the group is secret rather than “closed” or “open,” there are a number of ethical concerns in accessing the content. For one, secret groups are under protection from Facebook’s web scraping policy. 2 Second, many secret groups, this group included, have confidentiality agreements listed on their page that specify no content is to be taken outside of the group.
I conduct a digital ethnography that employs Coleman’s (2010) explication of the method as an anthropological endeavor, attentive to the self-constructed culture, language, and traditions in online communities. Through this lens, I examine how the group serves, creates, and reinforces a certain type of community and certain type of communication. Following previous research on protected spaces online (Gehl, 2016), I do not remove any content from the secret group’s page. Although there are inevitable limitations to this approach, it is one way to preserve the confidentiality of secret groups while also allowing researchers the opportunity to observe behavior across the blurring boundaries of social media.
My ethnography spanned from 1 January 2018 to 1 January 2019. I engaged in participant observation that, like Gehl’s (2016) study, focused on group “governance and interaction dynamics of the site” (p. 1221). This included attention to group rules and norms, the kinds of posts within the group, and how members interacted with one another on the page. Like Baym’s (1995) seminal study of computer-mediated communities, I considered the online community as “emergent,” formed from interactions within the group and influenced by member’s offline context. In the discussion of my findings, I group these observations into loose categories (Gehl, 2016; Koopman & Matza, 2013; Luker, 2009) that help organize practices in and experiences of the group over time.
Findings
The secret groups included in this study are all from the U.S. South, all from states with majority conservative voters, all formed by White women, and all at some point used the secret group function on Facebook to organize. Although these shared characteristics mean my observations come from a limited sample of secret political groups online, they allow me to generalize logically from my observations and provide an in-depth look at how marginalized progressives use online secret groups to organize. In my analysis, I focus on the two questions consistently relevant to work on secret organizing. First, why were these groups founded and why did people seek to engage in secrecy? And two, what happens when they do? I focus first on the creation of the groups—who started them as well as how and why they were created. I then turn to their affordances and their outcomes—what is different in their digital and interpersonal structure from other forms of online communities, leaning heavily on the ethnographic portion of the study. Finally, I turn to the emergent community within these groups in terms of their function, content, and purpose for members.
Formation and Secrecy
Each group was started by progressive White women in the U.S. South. Although they were all started by women, three were comprised solely of women while two were not. For example, Beth’s group in Texas had unexpectedly been only female. When she had started the online group with a friend, she found the women she had invited had in turn only invited other women. Andrea’s group in Louisiana, however, was mostly women but not exclusively. While the group had been created first as a secret group and included almost entirely women, as they expanded their publicness, they attracted more male members.
All of these groups were created in connection to a physical location (e.g., a county, a town, a city). Some formed as off-shoots of a larger secret Facebook group affiliated with a larger geographic area. For instance, Beth formed her group, which was associated with her local town in Texas, by finding others on a secret group for her state.
Somebody from college told me about [larger secret group on Facebook], and then as soon as I found out, I started telling people about [larger secret group on Facebook]. And then a lot of people found each other that way. And then my friend [co-founder], who I didn’t know well before the election—our daughters were on the dance team together, they weren’t particularly friends, so she and I just knew each other as dance moms, you know? And then she asked me about starting a group . . . (Beth)
Andrea’s group in Louisiana had also formed as an off-shoot of a larger secret Facebook group for her state, using the group to locate those who lived in their local area. Others, like Leanne’s group in Florida, had formed as an independent secret group of its own making without a larger group to help funnel individual members into a local chapter.
Importantly, the groups were formed on Facebook as secret groups. That is, the leaders had reason to hide their identity and membership from others in their online social networks. When I asked why they had intentionally concealed their groups, the women mostly cited fears of social and economic ostracism from their networks. For instance, Beth referenced members who felt they needed a safe place where clients or patrons of their business would not be able to see their political leaning:
It’s a secret group because there are business owners who do not want their political opinions know in our group. And so, they really need their private place to go because there are, you know—because we live in a county, or I guess a town, where people are vocal about being Trump supporters and they have businesses that they feel would be hurt, so we respected that and that’s why we made it private. (Beth)
When I asked Beth why she would not want the group to be public, she told a story of community hostility when discussing politics, even among children.
I try not to be very political because anytime anything would come up, people would get hostile . . . [her daughter] would come home and say that people were yelling at her in the playground. There was a particular child that said that Hillary killed her own baby, and said to her [daughter], “don’t you know that Hillary killed babies?.” . . . She [daughter] must’ve talked about it and said she was going to vote for Hillary . . . (Beth)
Like, Beth, Andrea mentioned that the group had been formed as secretive, and part of it remained a secret, to protect those members that were afraid of losing their jobs for being publicly engaged in politics.
A lot of people work for state government, and so one of the big things that you will find with those people is that they’ve been kind of hammered down, ‘you can’t get involved in politics,’ right? And so, there’s a lot of people who really want to speak out, who really want to be involved, but they feel like their hands are tied because if they do, they’ll lose their jobs . . . (Andrea)
There was also strategy in Andrea’s approach. She, and the other group leaders, knew there were many progressives in their area of Louisiana who were not willing to be publicly progressive, but who were “excited” about politics. Building a secret group allowed them to reach this segment.
We noticed that when we called a meeting, 120 people showed up. And we’re like, “there’s a progressive voice here in [town],” and it’s just that there’s people who are just not speaking out. There are people that are excited too, but there’s just this group in the shadow . . . So we started calling more meetings. (Andrea)
Donna said her group had formed secretively in North Carolina out of a fear of “social ostracism.” When I asked her to describe what it was like for progressives or Democrats in her community, she told me a story of having another parent overhear her answering machine and finding out she was a Democrat:
And I found out later on when I was out of the room that she [another mother present] overheard my answering machine say something about [local NPR station]. And so, I came back into the room and she leaned over and whispered to me, “are you a Democrat?” and I whispered back, “yes”—I kid you not, just the two of us sitting in the living room whispering—then I said, “then why are we whispering?” . . . And so she shared a couple of friends of hers that were progressive-leaning, which surprised me, and I shared a couple of my friends that were progressive-leaning . . . She ended up joining the group . . . (Donna)
In some ways, the founders’ stories reflect the “secrecy-efficiency trade-off” (Crossley et al., 2012) of earlier research on secret organizing. That Donna and her fellow Democrat felt the need to whisper points to the challenges of recruiting or engaging with like-minded others offline where being overheard is possibility. But Andrea’s story is not what the secrecy-efficiency trade-off would have predicted. The group’s founders were able to recruit unique members precisely because the group was a secret and offered members participation without publicity.
While all of the groups originated through a secret Facebook group, not all of them stayed there. For instance, Donna’s group tried a number of different tactics for keeping connected, ultimately falling on email to help coordinate a membership who was older and less technically savvy.
We tried a couple of different apps for connection. This is just not an app connection group. I don’t know if the demographic is—like I have a couple of people who are very technologically-based, most of those are on the younger side . . . So right now we do it over email, but admittedly, that is not great. Email is fine with 25 people but not so good with 120. (Donna)
Andrea’s group had moved from the secret Facebook group to a more public organization, although the secret group remained active and available to members who were not interested in posting their beliefs publicly. She felt that accommodating people’s willingness to be public was an important part of planting a seed of activism in their members.
There’s a range of people—there’s the person who’s okay with being handcuffed to a door and there’s everything in between. And there’s a certain amount of people who are willing to give everything . . . And so, we’ve kind of tried to create avenues for each person to be able to be involved. So even if we have the person who all they do is read our email or all they do is donate, that’s okay. That’s where they’re at. (Andrea)
Similarly, Beth’s group had remained secretive online, but some of their members had taken to more visible political action offline. Yet the group still accommodated those who were unwilling to go from the secret group online to their picture in the local newspaper:
I told the group, if anybody gets in a picture and they don’t want their picture in the paper, they just need to let me know and that’s not the picture that we will publish . . . there are about half of us in the group that are ready to just be like “I don’t care what anybody thinks,” you know what I mean? But there are still people who are very much quiet about what they think because they are afraid of repercussions on their business. (Beth)
Affordances and Functions
Community of Solidarity
As evident in their formation, the groups explored here helped fill a communal void for solidarity. Many of their members had few people, if any, with whom to discuss politics. Many felt their political beliefs put them at a disadvantage in their online or offline communities and had sought to hide their beliefs from others. Some, like those in Andrea’s and Sandy’s groups, felt as if everyone in their physical proximity held different political views than their own. Some, like Donna and Leanne, knew almost everyone around them was on the other side. It makes sense, then, that upon founding or finding their respective groups, members established a new form of community and sense of political solidarity that was not only like-minded, but also understood their local experience. Donna mentioned that not only did the group provide understanding and opportunity for conversation, but also “safety” in their community, that “there’s strength in numbers,” and that the group gives members “another opportunity to continue the conversation.”
Upon joining their group, some members expressed a feeling of instant camaraderie simply because the existence of others that were politically like-minded in their area had felt impossible. Leanne told me when members first joined, they would often express their shock to others in the group: “what we would hear most from members when they joined is ‘I thought I was all alone.’”
This is not unlike the solidarity provided by other marginalized online communities. For instance, even less structured communities created through social media hashtags have helped organize and support trans individuals, particularly trans women of color (Jackson et al., 2018), which “prioritized in-group solidarity from day one” (p. 1,883). Other, more structured groups have also provided women who feel marginalized in their careers solidarity and information (Pruchniewska, 2019) and fringe political groups solidarity for their beliefs (Wojcieszak, 2009). What is unique is that while these groups existed online, like other digital political organizations (Karpf, 2012) they were still very much rooted in place. The members, however, relied on social media to find one another locally because doing so through other means was too hard.
TXP also served as a space for discussion and solidarity for those who felt locally outnumbered. Because I analyzed TXP’s content across a year, I can loosely categorize the solidarity I observed into two groups (Table 1). Many of the posts featured callouts to others in the group that celebrated their shared beliefs, or ingroup solidarity. Like ingroup favoritism, which posits that people select messages that favor their own social group (Hewstone et al., 2002), these posts conveyed a sense of ingroup membership but sought to connect oneself to the ingroup even further. This solidarity took two forms: members either expressed their solidarity with others through a declaration of shared beliefs (e.g., “I’m so happy to know I’m not alone”) or by displaying an action they felt shared their solidarity (e.g., showed pictures of them holding an “I Voted” sticker or wearing a local candidate t-shirt).
Secret Group Content Categories.
Note. Categories taken from content observed between 1 January 2018 and 1 January 2019.
On the other hand, many of TXP’s posts involved forms of negative solidarity, or outgroup derogation, in which members would post some kind of outrage toward conservatives or Republicans. Like outgroup derogation, which suggests that people select messages that present the opposing side in a negative light (Jackson et al., 1996), outrage on TXP was largely directed at conservatives or Republicans. This outrage took two forms. Often, members would express elite outrage, or resentment toward conservative elites, like politicians or news media. These commonly included links to articles as evidence. For example, one member posted their horror that Republicans were attempting to appeal the Affordable Care Act and posted a link to an article as evidence. Another member posted a screenshot of a local politician’s email to constituents, commenting on its hypocrisy. Members would also post outrage at their own experiences with conservatives or Republicans online or offline. What I call experiential outrage was content that expressed disgust or horror at an interaction, observation, or conversation with someone from the opposing side. These included firsthand encounters or experiences members had with conservatives (e.g., conversations with neighbors, bad experiences working the Democratic primary table or conversing with a conservative online). These posts often featured pictures or screenshots of bad conversations with conservatives or pictures of vandalized Democratic yard signs taken in their community.
That solidarity was both positive toward the ingroup and negative toward the outgroup presents the dual, and perhaps unique, qualities of a secret community. That is, a community that protects members from outgroup scrutiny and retaliation and opens the door to bonding along shared antagonism rather than just commonality.
Community of Contention
Beyond a space for understanding and mutual dislike, the groups helped members build a shared narrative of contention within the progressive space. This often took the form of criticizing the local Democratic Party, which in the areas where these groups had formed, was historically and currently under resourced, under supported, and underpowered.
Akin to Lei’s (2018) contentious public sphere that arose vis-à-vis digital media in China, these groups helped members establish a contentious private sphere—a place where they could simultaneously organize efforts to support progressive values, candidates, and movements, while also criticizing the existing infrastructure. That these groups were not directly affiliated with the Democratic Party meant they could mobilize support while also pushing back on the existing structures, hierarchies, and flaws of the Democratic Party that alone had failed to mobilize them in the first place. That the groups were secretive meant they could challenge existing infrastructure without losing their connections to the Party and its leadership.
Some of these groups included members of the Party itself, often partnering with the Party to campaign for candidates or help support events. For instance, Donna’s group in North Carolina partnered with the local Party to help promote female candidates running for office. In fact, they had invited two of the candidates to a group meeting, which had prompted one of their members to get involved in the Party, run for office, and win.
But one of the things they said was “women need to run.” They were really pushing it. And one of our members said, “Okay, guys. I talked to the House caucus at 5:00 today. I’m running.” And I was like, “Oh, my God.” And she won her seat. She beat one of the [Republican] members of our legislature. She started the race at [low number of] points behind and won. She won by [small number of] votes. (Donna)
Sandy’s group in Louisiana also had a member run for office after joining the group. Unlike in Donna’s region, the local Democratic Party had not come to her aid during campaigning. Sandy reflected on their absence, mentioning that the state Democratic Party had not given “anything” to their area and had failed to help them “solidify” the progressive movement in the state and locally.
[Author]: So, it sounds like from my read of things that the infrastructure of the Party in this area isn’t or wasn’t strong? Is that correct? S: Oh, absolutely not strong—pathetic. Democrats are certainly not the high number of registered voters in the state . . . They just aren’t even here. They’re not invested. They’re not giving time, money, anything to these areas. And so, it’s becoming even harder. (Sandy)
Others were strictly unaffiliated with the Party. Andrea, whose group is also in Louisiana, considered much of their membership to be distinct from the membership of the local Democrats. Like Sandy, she felt that the state and local Democratic Party didn’t “care” about them. The Party had failed to provide resources or invest in their efforts, so the group felt they had to fight for progressive values on their own.
I feel like the people who are involved in our group don’t necessarily subscribe to the Democratic Party nationally, and then also don’t necessarily subscribe locally . . . I do think that there’s a sense of “what does the Democratic Party give me?” A lot of people feel like there’s not a lot of investment, national investment, into Louisiana’s Democratic Party. It’s kind of like, “oh, it’s a red state—who cares?” It’s like the red states get forgotten about. (Andrea)
The tension between Party support and Party criticism among mainstream progressives was also evident in TXP. The time frame in which I observed the group included the lead-up to the 2018 midterm election in the United States. As a result, I saw a large number of posts referencing or alluding to the Democratic Party in Texas, a Party that had not seen much success outside of the municipalities in urban areas and, like Louisiana, had failed to invest in local efforts or regions that seemed unviable. On occasion, members of the group would post Party-positive content. Sometimes these posts referenced “turning Texas blue” or the “blue wave,” a slogan promoted by the state’s Democratic Party during the 2018 midterms. Sometimes these posts included references to Democratic candidates that supported their values or celebrations of Democratic legislative victories.
On the other hand, content in the group was also Party-critical in that it criticized or condemned the existing actions of Democratic politicians or Party efforts. Members sometimes asked each other to reach out to politicians or Party leaders to change their behavior or position. Many of these posts referenced frustrations like Andrea’s and Sandy’s. Members felt not only isolated in their communities as one of few Democrats but also abandoned by the Party altogether. Some members were quick to distinguish themselves as “progressives” rather than Democrats, much like Andrea’s distinction as a “liberal” rather than a Democrat. These tensions between Party-supportive and Party-critical are also reflected in Gose and Skocpol’s (2019) research on women’s grassroots organizations in Pennsylvania. Like their grassroots groups, the secret groups in this study had formed outside of the Party. Yet the secret nature of the groups in this study shows how private spaces online help contentious narratives form within the same political side rather than outside of it.
Community of Practice
Although these groups started as a secret and continued the private side of their organizing, they also gave members a pathway to public action. In many ways, the groups served as “communities of practice,” or organized spaces for sharing what Duguid (2005) would call “form and contexts” for information “as well as content” (p. 113). This was largely by offering a place to practice being political, particularly for those who were reticent to disclose their beliefs, unskilled in political action, or both. Sometimes this was a transition from unwillingness to be publicly political to taking a public political stance, or become what Donna called “more confident in their progressive values.”
Sometimes the groups transitioned members from private discussion online to public action offline. For example, Beth’s group helped mobilize members to join the Women’s March and to start block-walking, something Beth herself had never done before, as well as putting a political bumper sticker on her car.
It will be my first-time block walking . . . I’m starting to get a little bit braver. I’m going to put a sign in my yard and I’m putting a “Turn Texas Blue” bumper sticker on my car so those are little ways that I’m being braver about it. (Beth)
Leanne also saw the switch from private and hesitant to “visible and vocal” in her Florida group. She attributed this mostly to the group who had partnered with other local organizations to coordinate events and phone bank.
Many of us are more visible and vocal than we have been in the past. We have connected and networked with like-minded organizations in the region to plan marches and rallies. We are writing letters and visiting our local legislators. We are speaking at city council meetings. We got involved in record numbers during this last election to canvass and phone bank. Several members worked on congressional campaigns, and I took a job as a field organizer for the Democrats in our county. (Leanne)
The push for public action was also evident online in TXP. TXP members frequently posted “calls to action.” These calls included requests for members to do something offline (e.g., vote, attend a meeting, attend a march or rally) or online (e.g., share an article outside the group, share information about an event outside the group, like a Facebook page), although online calls were much less common. Because the time frame of this digital ethnography fell during the primary election and the gear-up to the midterm election, a lot of this content included calls to vote for progressive candidate or calls to campaign for Beto O’Rourke who, at that time, was a Democratic candidate for U.S. Senate.
Beyond encouraging members to take action, TXP also mobilized members by providing information on how and where to engage in politics as well as how to discuss politics. For example, members would frequently post inquiries that featured some type of information request. Sometimes these were communicative inquiries, in which members asked others’ advice on how to respond to a particular argument made by the opposing side. These posts often asked how to counter a conservative argument. For example, members asked how to respond to people who are anti-abortion or how to respond to conservative claims about immigrants. On other occasions, members posted information inquiries, in which they requested information about a particular policy, event, or clarification on something they had seen, heard, or read on the news. For instance, one member asked a question about whether or not the United States was in fact withdrawing from the United Nations and another member asked for help clarifying a piece of misinformation about Nancy Pelosi they had come across online.
Communities of practice and the move from private coordination to public action are not new to the functionality of social movements and political activism (Blee, 2012). More novel, however, is the shift from intentionally concealed practice to public participation, especially among those in the mainstream. Shklovski and Valtysson (2012) found a similar community of practice among counterpublics in Kazakhstan, where individuals practice discussing politics before engaging offline. Likewise, the secret groups in this study helped build trusted communities online, and then offline, where they could practice discussing politics and learn about how to get involved.
Discussion
This study suggests that secret groups on social media have unique forms and functions. Most importantly, that they covertly support public political action while also fulfilling a larger social and communal void for partisans who feel marginalized in their community. Like other examples of secret organizing (Clark-Parsons, 2018; Pearce et al., 2018), I find that mainstream secret groups form in response to feelings of isolation, but that this isolation is driven locally rather than nationally. Members of the secret groups in this study were locally marginalized while remaining nationally mainstream. Social media, namely the secret group function on Facebook, was key to each group’s development in that it allowed members to organize and coordinate covertly.
Through interviews with leadership and observing a year’s worth of content on a secret Facebook group of progressives, I find that the secret groups observed here serve three functions for members. First, they offer a community of solidarity, in which individuals who previously felt politically isolated or outnumbered are able to find understanding and support, akin to online communities of solidarity commonly established by other marginalized groups (e.g., Jackson et al., 2018, 2020; McKenna & Bargh, 1998). I also find that the group is able to provide solidarity not only through ingroup favoritism toward progressives but through outgroup derogation toward conservatives. Second, I discover a community of contention. By their protected nature and emergence distinct from a political party, these groups are able to form a contentious narrative about local party structure and the failure of the party to adequately support and mobilize their region. Finally, I find that online secret groups serve as incubators in two ways. First, they offer a community of practice, taking members from unwilling partisans to public dissenters. And second, they offer people a place to transition from online to offline—a coordinating mechanism that takes into account not only the connecting qualities of social media, but its protective affordances as well.
It is not surprising that Facebook was the medium through which these groups developed. For one, it is a platform with widespread, almost ubiquitous usage, meaning that to initially join the group took little effort beyond clicking a button. But Facebook was also key for the development of the groups in this study not just because it is ubiquitous, but also because it is opaque. In its current form, groups on Facebook can be formed as entirely “public” where anyone can join or as “private” where members must be invited to join. They can also be formed as “visible” or “hidden,” in which the former is searchable by non-members and the latter, which is only searchable by current members. That these groups used Facebook to organize carries its own implications for their subsequent interactions. For instance, Facebook connects users’ actual identities with their communication, which can influence both what people say and the extent to which their communication is polite or civil (Halpern & Gibbs, 2013). The visual and highly personal nature of Facebook can also enhance the development of a collective identity within the group by allowing individuals to share and connect their own personal experiences with the group’s larger purpose (Khazraee & Novak, 2018), something especially evident in how members of TXP used the group to share selfies of active solidarity. Facebook was also an essential tool for development because some relied on larger existing Facebook groups to form their own groups connected to their immediate communities. Along the neo-federated model of political organizing (Karpf, 2012), Facebook allows individuals not only to join groups but to leverage those networks to create new, location-based groups.
Theoretically, this study extends previous research on online communities and provides evidence that online secret groups are worthy of their own theorizing. For one, these findings support Baym’s (1995) early contention that online and offline contexts are increasingly blurred just as public and private are increasingly intertwined (Papacharissi, 2010). All of these groups were tied to some isolating experience in an offline context. In addition, while some of them became more public or had branches that sponsored more public activities, at least a part remained secretive, clinging to the confidentiality that had been so formidable at the beginning. These findings only add to this conversation and offer further evidence of the overlap between offline and online and public and private that color the contemporary digital space.
The groups in this study were formed by women, and membership comprised mostly women. Given previous research exploring closed groups on social media (Clark-Parsons, 2018; Pruchniewska, 2019), particularly related to politics (Gose & Skocpol, 2019), this gender composition is not surprising. Just as Pruchniewska (2019) finds that private feminist Facebook groups helped offer solidarity and incubate action related to women’s professional identities, I find that private political groups of women helped offer solidarity and incubate action related to their political identities. Beyond building communities of solidarity and practice, it is also intriguing that groups of women were the ones to establish communities of contention in secret, which like Gose and Skocpol’s (2019) work, suggests that women are often at the forefront of efforts to organize outside of the existing political structure or recreate the Party altogether.
Curiously, the women I spoke to rarely mentioned gender as a crucial feature of their organization. That the groups were led and dominated by women without it being explicit in their creation is perhaps evidence that women are more likely to be concerned with the visibility of their politics or more likely seek out secret groups in the first place. It could also be that, given the groups’ recruitment was word of mouth, women are simply more likely to interact with other women when it comes to political topics. Regardless, it is clear that the women’s experience of political marginalization was in some way confounded by their experience of being marginalized based on their gender, and more broadly, that mainstream political marginalization intersects with other marginalized identities in important ways. Previous research has long documented the marginalization of identities connected to race/ethnicity and gender/sexuality (Crossley et al., 2012; Herbst, 1994; Jackson et al., 2018), yet this study points to the importance of addressing the intersection of political identity and other lines of marginalization to understand how people organize in both a polarized and digital democracy.
In some ways, the groups in this study can be considered distinct from the subaltern counterpublics of earlier theorizing (Fraser, 1990; Warner, 2002). Most obviously, they were formed not as alternatives to the mainstream, but as a hidden part of the mainstream. In other ways, the groups explored here are similar to the counterpublics of contemporary movements (Jackson et al., 2018, 2020; Shklovski & Valtysson, 2012), in that they do more than just organize people, but also support and mobilize them. Members used their group as a community to practice politics. Like other communities of practice that consider common identity as a pathway to knowledge acquisition (Duguid, 2005), these groups provided identity and connection that then facilitated learning “to be” a partisan rather than “learning about” (p. 113) partisanship. Yet the rich affordances, features, and qualities of secret online communities suggest that these groups reach a different audience, serve a different role, and have a different effect than more public political groups and communities where political “practice” takes place. The secret groups of this study comprised different partisans than the local party included, challenging the idea of a secrecy-efficiency trade-off (Crossley et al., 2012) for recruitment and suggesting that secret political groups are worthy of study in their own right.
Methodologically, this study advances an approach to ethically and comprehensively explore secret groups on social media, which involve different considerations than public organizations and communities. There are many ethical concerns in web scraping or quantitative content analysis (Freelon, 2019), in which content associated with users’ personal accounts is both collected and stored. Digital ethnography serves as an ethical starting point for those looking to explore secret groups because this method observes, but does not collect, user data. This is not to say that all digital ethnographies should be treated equally. The challenges to studying protected or secret communities online are myriad and require a systematic and consistent empirical approach, something that current institutional practices through the IRB or similar bodies have yet to set or routinely implement. Because of that oversight, studies of secret groups like this one can and should draw from online research ethics when planning and executing their research (e.g., Roberts, 2015), particularly as the boundary between private and public on social media becomes increasingly unclear.
The form and function of secret groups on social media also hold implications for those looking to conceptualize the public, particularly journalists, pollsters, and political campaigns. For one, secret online groups can shield political organizing from the scrutiny of journalists, other politicians, and oppositional members of the public. This obfuscation makes it hard for opposing partisans to usurp their opponent’s efforts and for journalists or pollsters, whose job it is to uncover the opinions of the electorate (McGregor, 2019), to find a full picture of the public’s attitudes. Although the groups in this study likely contain members that express similar sentiments elsewhere, these groups also hold individuals whose membership is exclusive.
It is important to note that because I am limited by the confidentiality of these groups, my data consist of interviews and observations of a small slice of the potential mainstream secret groups that exist online. Moreover, because the groups are secretive, my access to and contact with each relied on their willingness to reach out to and speak to me. This resulted in a sample of marginalized progressives and data that is not representative of all mainstream political ideology in the United States. Yet this point in itself is not irrelevant to the conclusions of this study. That those who felt I share their political beliefs were more available than those who felt I do not only exemplifies the point that whose opinions we see are largely influenced by how they see our own.
Because this study is one of the first to explore secret groups in the political mainstream, there is a wealth of research still to be done. For one, it will be important to know the true causal effect that group membership has on the transition from secret to public. Although this study provides observational evidence that the group may incubate by offering a community of political practice, the degree to which this is a causal effect of the group will be important to understanding how secret groups truly shape public affairs. It will also be important to understand the potential differences in how different mainstream partisans organize, for instance, whether or not conservatives utilize secret groups online in ways similar to the progressive secret groups in this study.
Despite their limitations, these findings convey important and novel information about the current state of social media and its related research, most importantly that the simultaneous rise of political polarization and private social media has both necessitated and made possible mainstream political secrecy in the United States. This study showcases the unique form, function, and membership of secret groups on social media and the growing need for managing online visibility, not only in authoritarian contexts, but in liberal democracies as well.
Footnotes
Appendix
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
