The Ethoses of (Dis)Connecting with Friends on Social Media: Digital Cocooning and Entrepreneurial Networking among People with Eating Disorders

Abstract

Recent media studies conversations on disconnection or reducing mainly the quantity of engagement with social media so as to enhance well-being have suggested that these practices articulate a contemporary spirit focused on self-care and performance (productivity) that does not consider others or collective solutions. Drawing on and pushing forward disconnection research, we put forward a Foucauldian inspired concept of ethos that draws attention to qualitatively different principles and values characterizing social media socialities which users seek to foster and avoid. Interviews with people (n = 31) with eating disorders (EDs) featured what we call digital cocooning; that is, interaction with trusted real-life friends and family afforded by messaging apps characterized by mutual responsiveness, acceptance, and belonging. However, what we term entrepreneurial networking with wider acquaintances mostly on traditional social media was experienced as evaluative and competitive and fuelled a sense of non-belonging, prompting unfriending. Disconnection research has highlighted how social media (dis)connections are often underpinned by contemporary possessive individualism, obscured by the dominant research on ostensibly universal psychological processes. The concept of ethos pushes this research beyond criticism toward also highlighting alternatives or how social relations in social media and society could be imagined otherwise.

Introduction

Eating disorders (EDs) are mental health problems associated with preoccupation with body shape and size and characterized by problematic eating behaviors, such as restricting (Anorexia), binging (Binge Eating Disorder), and/or purging (Bulimia) of food (American Psychiatric Association, 2013). International studies have estimated that 8% of women and 2% of men experience EDs in their lifetime (Galmiche et al., 2019), although sub-clinical EDs are much more common with 16% of UK population and 28% of young women experiencing key ED symptoms (NHS Digital, 2020).

People with EDs can be considered vulnerable social media users. As body image disturbance is a diagnostic criterion, most research has focused on the use of visual platforms, such as Instagram, and concluded that they fuel eating disordered thoughts and practices (Holland & Tiggemann, 2016; Saiphoo & Vahedi, 2019). There is also research on online communities, especially controversial “pro anorexia” groups, which have been claimed to encourage EDs. It has been suggested that interacting with others with EDs online enables sharing feelings that are stigmatized and not understood by offline friends, but that sharing eating disordered thoughts and practices can fuel a group spirit to be thin/ill (e.g., Brotsky & Giles, 2007; Drtilová et al., 2022). Recent studies have drawn attention to how people with EDs have curated their Instagram feed by unfriending or blocking accounts that focus on slim bodies and restricted eating (Hockin-Boyers et al., 2021; Nikolova & LaMarre, 2023).

Thus, previous research on EDs and social media has largely focused on traditional platforms, such as Instagram, and has highlighted contradictions in online personal relations between people with EDs, mostly in relation to appearance and eating behaviors. There is hardly any research on how people with EDs use social media to connect with real-life friends. This is surprising as people with mental health problems often struggle with social isolation. Messaging apps, used especially with close friends and family, are also a stable part of social media use generally, with WhatsApp being the most popular app in the United Kingdom in 2023 (Dixon, 2024). Furthermore, while the primary focus of previous research has been on the body, exploring social relations draws attention to the important relation between the self and others. It raises the questions of what kinds of social relations users cultivate in conjunction with platform affordances and historically specific social norms and of their implications for mental well-being.

Social media use and mental well-being has primarily been studied in quantitative psychology, which has focused on associations between technology use and ostensible psychological universals, such as social comparisons (e.g., Seabrook et al., 2016). Here we take a lead from and push forward a more contextual approach articulated by recent media studies research on disconnection or practices of regulating engagement with (mainly time spent on) social media to enhance well-being (Beattie, 2020; Jorge et al., 2022, 2023; Syvertsen, 2023; Syvertsen & Enli, 2020). Disconnection research has been critical of such practices, arguing they articulate a contemporary spirit focused on self-care and performance (productivity) that does not consider others or engage in collective solutions (Baym et al., 2020; Beattie, 2020; Figueiras & Brites, 2022; Kaun, 2021; Syvertsen, 2023).

Research on practices of disconnection has usefully analyzed the individualistic social and historical values they articulate, which have been obscured by dominant research on psychological universals. However, disconnection research has focused mainly on criticism and on quantity of engagement. There is little research on quality of social media connections, or on what kinds of socialities users pursue or try to avoid. Exploring different qualities of engagement goes beyond criticism and highlights alternative ways of imagining social connections on social media. To make sense of such alternatives we draw on late Foucault (1990) and anthropology (Geertz, 1973) to propose the concept of ethos, which refers to the values and principles articulated in practices and relations of specific social groups in a specific social context.

Interviews we conducted with people with EDs featured an alternative ethos we call digital cocooning, referring primarily to the use of the affordances of messaging apps to interact with real-life friends and family that articulated principles of responsiveness to others’ distress, understanding, affirmation, belonging, and mutuality. In contrast, connecting with wider networks of acquaintances mostly on traditional social media, such as Instagram, was characterized by an ethos we call entrepreneurial networking, marked by a sense of evaluation, competition, and non-belonging, leading to often fraught attempts to disconnect, such as unfriending and withdrawal.

In what follows we will outline an approach to making sense of users’ practices of (dis)connecting with friends on social media in terms of different ethoses. We will then elaborate on the details of digital cocooning and entrepreneurial networking ethoses as experienced by people with EDs.

Social Media (Dis)connection

Quantitative psychological research has concluded that social media use in general does not have a negative effect on mental well-being, but that social connectivity enabled by social media has positive mental health implications, whereas upward social comparisons (considering someone to be doing better than you in some respect) have negative implications (Appel et al., 2020; Seabrook et al., 2016). Psychological research offers useful observations. However, since it is focused on associations between technology use and ostensibly psychological universals, such as social comparisons (Festinger, 1954), it is largely acontextual.

The concerns about the mental health implications of social media have sparked various initiatives to regulate engagement. Recent media studies research has critically analyzed such efforts to “disconnect,” mainly focusing on temporalities or efforts to reduce time spent on social media. It has been observed that texts promoting digital detox focused on users’ self-responsibility and a nostalgia for offline “authentic” experiences (Syvertsen & Enli, 2020), and that detox activists were skeptical about possibilities to change social media platforms and saw themselves as part of a social trend even if the activities focused on individual experiences (Syvertsen, 2023). Analysis of technologies to reduce time spent on digital media has concluded that they promote a new self-discipline to enhance productivity and well-being rather than challenge the platforms (Beattie, 2020; Jorge et al., 2022). Research on user experiences has also found that users seek to mainly limit time on social media (Nguyen et al., 2022), and adolescents wanted to disconnect from social media, which was perceived as boring and wasting time, as neither productive nor fun (Jorge et al., 2023).

Some work has explored disconnecting from social connections or sociality, in terms mainly of evading overwhelming volumes of messages (Jorge et al., 2023; Mannell, 2018). More generally, Kaun (2021) has argued that practices of disconnection articulate a contemporary spirit of self-centredness (Bauman, 2003; Illouz, 2018) with little consideration of how digital self-care may affect others, of how users could change their behavior to modify the social media experience for others or of what they could do together with other users (also Baym et al., 2020; Figueiras & Brites, 2022). From a psychological perspective, Turkle has suggested that social interaction through digital devices facilitates light touch relations that keep others at arm’s length, with conversations easily called off by not responding or ghosting if things become inconvenient (Turkle, 2011).

However, there is little research on what kinds of social connections users seek and what kinds they try to avoid. It has been observed that users feel pressured to live up to expectations, especially in terms of appearance, when engaging with friends on traditional platforms (Gill, 2021; Jorge et al., 2023). It has also been found that people mainly unfriend acquaintances rather than close friends, typically because they post too much, especially trivia (Sibona, 2014). Earlier research has also explored how users create different socialities across social media platforms by connecting and disconnecting with others (Light, 2014; Madianou & Miller, 2013). Research on online mental health communities has observed that they may develop, for example, affirmative ways of communication to counteract negative thoughts (McCosker, 2018). Das and Hodkinson (2019) found that it was easier for new fathers to disclose and discuss mental health problems with friends and family via messaging than in person and indirectly disclose their distress to friends on traditional platforms.

In political communication unfriending has been considered problematic, leading to increasing close-mindedness and polarization without civic dialogue. Yet, unfriending in the context of Israel and Gaza conflict (John & Gal, 2018), Hong Kong politics (Zhu & Skoric, 2021), and disability activism in the United States (Trevisan, 2020) has been observed to be motivated by distress and intended to create a safe space. Furthermore, unfriending has been found to be more common among minority political and stigmatized groups that felt offended by more powerful groups (Zhu & Skoric, 2021).

To conceptualize the scattered research on rewarding and distressing social media connections, we put forward the notion of ethos. The Greek word ethos translates to ethics and, for example, Bauman (2003), drawing on Buber (1970), argues that contemporary societies are characterized by an instrumental approach toward others as objects (I-It) to serve the interests of the self, as opposed to an I-Thou relationship that respects the other for who they are. Bauman’s discussion highlights different ethoses, and his observations on objectifying or self-centered relationships map onto critiques of disconnection focusing merely on self-care (Figueiras & Brites, 2022; Kaun, 2021). However, rather than seeking to identify ethical or ideal relationships, such as I-Thou, we draw on late Foucault’s contextual notion of the ethics of the self.

Foucault’s ethics of the self refers to the ancient Greek reflexive sexuality that sought to create a distinct style rather than be guided by strict interdictions, as in the later Christian period (Foucault, 1990). Foucault was drawn to the Greek ethos, as he considered it to offer more latitudes of freedom to create different subjectivities and relations. Yet, he emphasized that it articulated a historical context of elite free men fashioning themselves to acquire the respect or admiration of their peers and legitimacy to govern others in the city states, and that the practices changed in the context of Roman Empire (Foucault, 1984). It has been noted that Foucault’s notion of ethics of the self resembles the contemporary fixation on self-regulation, also characterizing disconnection initiatives (Beattie, 2020; Figueiras & Brites, 2022). Thus, to better gauge the social dimension of different ethoses of social media sociality, we also draw on the anthropological notion of ethos as principles and values guiding the habits, practices, and relationships of social groups (Geertz, 1973). This also highlights that different groups may embrace different ethoses, and also captures their everyday, habitual aspect, not necessarily guided by meticulous self-reflexivity. Yet, Foucault’s strength is his keen awareness of how ethos is always informed by the historical spirit of the times and its social, political, and economic ramifications, which is often bypassed in anthropology.

Turner’s (2006) history of the origins of social media in the Californian countercultural communities and media (Whole Earth ‘Lectronic Link, WELL) helps to understand the ethos dominating connections on social media. Turner chronicles how the 1960s counterculturalists, who sought to create alternative communities, lifestyles, and economies, later adapted them for self-promotion and entrepreneurial networking to get ahead in the 1980s Silicon Valley freelancing and project-based tech economy. Turner’s history highlights both the appealing side of social media in terms of connecting like-minded people and its dark side in that these connections are often self-interested and competitive. Turner’s discussion also illustrates that social media is not an isolated/isolatable force but an integral part of increasingly individualistic and volatile post-capitalist social and economic context (Bauman, 2003; Boltanski & Chiapello, 2005; Illouz, 2018).

Research on disconnection has argued that it frequently remains wedded to the same individualist and entrepreneurial ethos permeating the platforms seen as the problem (Beattie, 2020; Jorge et al., 2022; Kaun, 2021; Syvertsen, 2023). However, in focusing on critique, research on disconnection has hardly considered the quality of connections or what kinds of socialities users pursue and seek to avoid on social media. Against this background, we will explore, through the lens of ethos, what kinds of values and principles people with EDs experience, foster, and seek to avoid when they connect and disconnect with real-life friends and family using social media.

Methods

The research was based on 31 qualitative interviews with people who self-identified as having past or present experience of EDs. As most research on social media and EDs has focused on body image, we wanted to explore, in an open-ended manner, both positive and negative experiences with social media. The interview schedule focused on participants’ general life context and ED, and general digital media use including positive and negative experiences. We also asked participants whether and how they had regulated their digital media use and what advice they would give others on using digital media. We specified that by digital media we referred to social media as well as, for example, self-trackers, mental health apps, and online information and support. In this article, we focus only on social media, including both traditional platforms, such as Instagram and Facebook, and messaging apps, such as Messenger, WhatsApp, and Snapchat.

The interviews were conducted between January and December 2020, most being conducted between late May and June and between early September and early November. This meant that most interviews were conducted when the United Kingdom was between the two main covid-19 lockdowns (March-May 2020; January-March, 2021). The situation likely intensified the role of social media for interpersonal interaction, although messaging is widespread (Dixon, 2024).

First, the third author (A.B.) recruited four participants from an ED treatment service for a pilot student project in January. In May we posted a recruitment ad for people self-identifying as having a current or past experience of ED on the website of the UK national charity Anorexia and Bulimia Care (now Talk ED), which was shared on the charity’s and the first author’s (P.S.) social media (Twitter and Facebook) several times. Ten participants were recruited through the charity’s and three through personal social media. A further 14 participants were recruited via a credit bearing scheme for psychology undergraduates in the South of England in September–October to reach robust thematic saturation. The different methods of recruitment diversified the sample. The participants recruited from the treatment center and ED charity included both young adults and middle-aged people and activists, whereas the student sample were all young adults, typically not involved with ED charities. Participants had experience of anorexia (n = 10), bulimia (n = 6), binge eating disorder (n = 3), other EDs (n = 2), and a combination of EDs (n = 10). Some described their ED as acute; others considered themselves fully recovered; the majority defined themselves as in recovery. Eighteen participants were between 18 and 24 years, 3 were between 25 and 29 years, and 10 were between 30 and 49 years. The participants included 28 women and 3 men; 25 participants identified as White and 6 identified as (South) Asian or of mixed-race ethnic heritage. All participants were either university students or had a university education, which should be borne in mind when interpreting the findings.

The qualitative interviews were conducted mostly through video call (mostly with camera but on a few occasions without camera and once via chat based on participant preferences). The pilot interviews and a few others were conducted by phone due to technical issues. Most interviews lasted approximately an hour. The interviews were transcribed verbatim, anonymized, and coded for themes using the constant comparative method (Glaser, 1965), facilitated by NVivo 2020 software. The coding scheme was developed collaboratively between the three authors, following repeated reading of transcripts. The thematic analysis was abductive (Timmermans & Tavory, 2012) in that literature on, for instance, disconnection was brought to bear on the analysis as it was relevant to the preliminary themes identified.

We gained ethical approval from our universities’ ethics committees and followed standard ethical procedures. After potential participants had emailed us in response to recruitment ads, we sent them a participant information sheet describing the study including helpline numbers and recommending that they not take part if the study was likely to cause distress. Participants were also asked to sign a consent form before the interviews, and during interviews, we did not probe for details of ED experience but steered the conversation toward their social media use. Audio(visual)-recordings and consents were uploaded to a password-protected university cloud storage and anonymized transcripts were analyzed.

We identified three broad thematic areas in the interviews: (1) connections with real-life friends and family, (2) content consumption, and (3) help-seeking. This article focuses on connections with real-life friends and family, within which are two main subthemes: (1t) digital cocooning and (2) entrepreneurial networking. Participants were assigned a random identifying number between 17 and 48, and these numbers are used to identify participants in the findings.

Findings

Cocooning

One-to-One Cocooning

Participants frequently discussed how the affordances of messaging apps helped them to interact with close friends and family when not feeling well, which they considered helpful for their mental well-being. Messaging apps were used, for example, by participants doing undergraduate studies at a university to receive support from parents and other close relations back home:

If anything was bothering me . . . and I was like very anxious about something, I’d call my mom on like Messenger . . . and it helped me a lot (P48).

A lot of us would be quite helpless and hopeless if we didn’t have the technology . . . If I didn’t have my phone, I wouldn’t be able to text my family or call my mum or text my boyfriend or my friends at the other universities . . . it definitely helps. (P37)

The comments on using apps to interact with parents and other close relations highlighted how the technology helped the participants to receive timely support when experiencing, for example, a bout of anxiety. The importance of this support was underlined by the fact that a couple of the participants had previously dropped out of undergraduate studies due to mental health problems in their first year.

The principle or ethos of these relations was to be able to count on the responsiveness of others at a time of distress. These interactions were also characterized by what Illouz calls the therapeutic ethos, whereby difficult emotions are aired and discussed to facilitate adjustment to challenging and/or changing circumstances, such as educational transition, typical of our historical times (Illouz, 2008).

Participants also related that messaging apps afforded them creative indirect means of communicating with close friends, when feeling unwell or body conscious:

I’m more likely to send a WhatsApp message to my friends, and sometimes I’ll send them a WhatsApp video or WhatsApp audio recording. So, I might not want to ring them, but I will send them a little two-minute video of me going “hi.” (P24)

Interviewer: Do you use the camera on Messenger?

Interviewee: No, but you can . . . I don’t because of confidence issues, because I have gained a substantial amount of weight. (P40)

It has been argued that messaging apps facilitate light touch or quick communication easily terminated when convenient by non-response (Turkle, 2011), which is echoed in promotion of more “authentic” in-person interaction (Syvertsen & Enli, 2020). However, new fathers found it easier to disclose and discuss mental health problems using messaging (Das & Hodkinson, 2019). Similarly, participants with EDs found messaging offered easier indirect means to keep in touch with friends, when direct communication felt too “demanding”; it also enabled participants to shield their body.

Group Cocooning

Many participants interacted with close friends who also experienced mental health problems using messaging apps, either one-to-one or in a group, as illustrated by comments by a young woman (P40) and a young man (P21):

I have a friend, who has similar issues with me with eating disorders and depression. So, there was a meme we shared earlier that says . . . Catch me spiralling into the endless void . . . It like allows you to express your mood . . . without it being like a lengthy conversation, we’re just like “ahh, I relate with that, too.” We’re not alone. (P40)

I’ve got a lot of friends who are open about their mental health, friends who have had an eating disorder . . . I think [it] is vital because you can get like an immediate response to anything you want to talk about. . . . I’ve got like a WhatsApp group with friends, I know it’s always a safe place to chat to people. (P21)

It has been suggested that limiting online discussions to like-minded people can lead to insularity (John & Gal, 2018). Yet, minority and stigmatized groups, such as those with a disability, may face a lack of understanding or hostility both online and offline, and interacting with similar people fosters a sense of belonging, understanding, and acceptance (Kaur & Saukko, 2022; Trevisan, 2020).

P40 related sharing of negative affect (memes about falling into a void) via social media. This could lead to holding onto rather than letting go of the sentiment without outside input as has been suggested in relation to pro ana groups (Brotsky & Giles, 2007). Instead, though P40 conveys relief in shared understanding. And, as feminist media studies research has suggested, although especially girls’ displays of negative affect online have caused moral panics, they serve as a counter-discourse to the pervasive, compulsory positivity on social media (Dobson, 2015). Thus, sharing feeling unwell allowed people with EDs to cocoon themselves with others with mental health problems from misunderstandings, stigmatization, and the onslaught of positivity on social media and to feel connected and understood.

Participants also used messaging apps to interact with groups of trusted close, real-life friends that did not necessarily have mental health problems through group calls, chats, and posts:

On Snapchat it’s not at all difficult to like swipe onto someone and have a chat with them about it . . . Like quite a few people put it up when they are like feeling down, and then loads of people pop up to them and message, like don’t be, you’re okay, and stuff like that, that’s like a positive. (P44)

I’ll just go on my like Discord service and go into voice chat and then other people can see I’m there and then other people will join. . . . The platform itself is focused around playing games together, so we like play games together and stuff. (P35)

The account of P44 illustrated another instance of friends modeling popular therapeutic interaction (Illouz, 2008) online by affirmation or giving positive feedback (you’re okay) to friends having negative thoughts, which has also been observed in online mental health groups (McCosker, 2018) even if validation may also be more complicated in online support groups (Moreton & Saukko, 2024). Messaging has been observed to enhance people’s perception of social support, including positive feedback (Frison & Eggermont, 2015), although sociologists have criticized positive psychology for promoting changing one’s thoughts rather than addressing the social ramifications of mental distress (Frawley, 2015). Nevertheless, affirming the value of a person as they are can also be seen as an antidote to ubiquitous discourses on social media framing users as in need of self-improvement in terms of appearance, body weight, performance, and so on (Banet-Weiser, 2018). The account of P35 illustrated a more light-hearted, enjoyable interaction with close friends to relax and take one’s mind off things. Both users used messaging apps to shift between one-to-one and group interaction, highlighting the fluidity of the boundary between semi-public and private social media.

Real-Life Friends and Preliminary Thoughts

Participants often emphasized that limiting interaction to close, real-life friends afforded by messaging apps was key to why they were beneficial for mental well-being. For example, as research has suggested that Snapchat use is associated with body dissatisfaction (Saunders & Eaton, 2018), we double-checked with several participants that they considered Snapchat positive, in response to which participants emphasized they only used it to interact with close friends: “I just have my friends on [Snapchat] and they don’t influence me in any bad way” (P36) and “It’s less toxic . . . I just use it to communicate with my friends and stuff” (P46). While most participants had mixed or negative views of Facebook, a few had connected with “only friends and family” (P26) on the platform and did not experience it in a negative way.

The benefits of connecting with close friends were likely related to the fact that they had developed a supportive ethos of communication with them. This was different from acquaintances (discussed later), although participants also mentioned that they avoided some close friends and family on social media, who were perceived as, for instance, judgmental.

Many emphasized that keeping in touch with friends was important as EDs could be isolating. Furthermore, some also emphasized an ethos of mutuality in social media interaction, of giving support in return, countering criticisms that online self-care is self-centered (Kaun, 2021):

It is very important to keep in contact and always check up on your friends, because it’s very easy in an eating disorder, you get too lost in yourself. Yeah, and because so many people are helping you, you kind of forget to help others, so it’s always important to let other people know . . . that I can still support. (P39)

Participants did not discuss many drawbacks of interacting with close friends on messaging apps. A couple mentioned they could sometimes feel overwhelmed and needing to regulate their messaging:

I have friends who have eating disorders and . . . sometimes you just have to take a step back. . . . Everyone says all the time: “Oh, my DMs are always open.” You don’t need to be like that. (P35)

It has been noted that users often experience the volume of messages as overwhelming and develop ways to evade them (Mannell, 2018). However, P35’s comment mainly illustrated the taxing nature of the emotional labor (Hochschild, 1983) involved in cocooning in terms of helping friends with their mental health problems, which is also noted in relation to gendered modes of disconnection (Van Bruyssel et al., 2023) and online support groups (McCosker, 2018).

Overall, digital cocooning illustrated a communicative ethos that participants experienced as conducive of mental well-being. It was related to disconnection in that the affordances of the apps enabled participants to restrict interaction to a small group of trusted friends. Although it had elements of self-care, it was not primarily self-centered as has been argued in disconnection research (Baym et al., 2020; Figueiras & Brites, 2022; Kaun, 2021). Rather, digital cocooning articulated a qualitatively different sociality or ethos, based on being responsive to others’ distress, understanding, acceptance, belonging, affirmation, and mutuality.

Entrepreneurial Networking

Social Comparisons?

Different from cocooning, participants experienced connecting with a wider network of acquaintances on traditional social media, such as Instagram and Facebook, as contradictory and negative and attempted to avoid such connections. However, many appreciated the observed quality of traditional platforms for enhancing connectivity (Ellison et al., 2007) with a broad network of friends, including a middle-aged participant:

[You can] catch up with your friends no matter where they’ve like moved to, even friends that I’ve got in America. And you can contact them for free, and you can talk to them on social media, and you can see each other’s pictures and what they’re up to. (P18)

P18 highlighted the benefits of Instagram and Facebook in terms of keeping in touch with friends that might be far away in space or time (old friends) and being able to shift between browsing and messaging. However, in the next breath, P18 related that browsing on the platforms also led to a sense of inferiority:

Even my husband, he’s got a Facebook page, and I see pictures that are shared on there or on his Instagram. And I’m thinking well, I don’t look anything like that, and look how fit [his female friends] are . . . and I’m his wife, and it has that knock on effect and makes me feel that he’s going to think that I’m overweight. (P18)

Thus, P18 got caught in surveilling and comparing herself with her husband’s female friends’ appearance on traditional platforms and felt negatively about herself, as has been widely reported in psychological literature (e.g., Seabrook et al., 2016).

Many participants reported feeling inferior and pressured browsing acquaintances’ posts, including checking old school friends “weight and how they are doing career-wise” on Facebook (P25) or how their peers on LinkedIn “are working on really nice terms and earning a lot of money, whereas I don’t” (P28). These comments could be read through a psychological lens as illustrating social comparisons, but taking a contextual perspective they reflect distress with the dominant contemporary sociality, underpinning the affordances and ethos social media platforms, characterized by self-promotion, competition, and entrepreneurial networking to get ahead in social and professional lives, as discussed by Turner (2006).

Participants could use the concept of social comparisons to make sense of their experiences, highlighting how psychological notions become part of people’s self-understanding (Saukko, 2008; Tiidenberg et al., 2017). The theory asserts that upward social comparisons may fuel feelings of inferiority but also aspiration to improve oneself (Festinger, 1954). Thus, a participant was disappointed with herself for not experiencing’ others happiness on social media as aspirational/motivational:

I find Facebook very harmful to myself exactly because I’m not happy with where I am. . . . They’re happy I’m not, they went on a holiday I didn’t, they’re having a family I don’t, and this kind of comparison rather than motivating . . . is putting me in a bad mood. . . . I know it could be motivational in the sense that if you’re unhappy then do something about it, and I think I’m currently in that stage where I am trying to push myself, and then hope that I will open up to this side of social media more once I’m happier. (P41)

The comment by P41 highlights how she first felt falling short of the ideal of being happy and second of the ideal of being motivated by others’ happiness to change her attitude and/or life. This demonstrates how psychological theories locating the problem with social media use in the individual may further fuel a sense of inadequacy, which is especially problematic in the context of traditional social media platforms underpinned by an entrepreneurial logic of self-marketing, rife for making vulnerable users feel negative. The participants thus often embraced the idea that it was their responsibility to regulate their social media use (as well as themselves) as suggested by critical literature (Baym et al., 2020; Kaun, 2021; Syvertsen, 2023). However, rather than articulating an entrepreneurial spirit toward managing their use and/or their attitude, participants often lamented their limited ability to do so.

Unfriending and Preliminary Thoughts

Many participants had unfriended acquaintances on social media that had made them feel anxious. Instagram was considered the most problematic app. Participants typically followed a variety of (micro-)celebrities as well as real-life friends often showing off “amazing bodies” (P21) on it, highlighting problems with platforms, where connecting with friends converge with traditional broadcast media and its celebrity aesthetic code (Marwick, 2015). Many participants had blocked some accounts (also Hockin-Boyers et al., 2021), and a couple had deleted the app temporarily at the time of interview.

Some participants had taken more drastic action and had a “massive cull” on Instagram from about “100 followers to only 10 actual friends” (P24), whereas others more discreetly “used the mute button a lot,” so their friends would not know they had hidden their posts (P35) (also Mannell, 2018). Participants sometimes unfriended specific connections that had a negative impact on their mental well-being, such as “old school friends that put things into your head” (P44) or people from inpatient ED treatment, because they “did not want to hold onto that part of illness” (P22).

One participant had joined a Messenger group with old school friends during covid-19 lockdown, but she left it because:

So, they started these conversations . . . like what worries them. And I’m still struggling with my mental health . . . with getting out of bed in the morning. [And they were talking about] what matching colours to put into my new extension and having to homeschool my children. (P26)

The account above highlights that interaction with friends through messaging apps could also be experienced as upsetting. This could be, as in P26’s account, because connecting with more distant friends could end up fomenting a sense of non-belonging and lack of mutual understanding.

While many participants unfriended some real-life connections, severe curtailing of connections was rare. Participants who were working often noted that maintaining connections and digital (self-)marketing was part of their job as illustrated by an early-career academic discussing Twitter:

I don’t love to go on Twitter. I feel it’s quite self-promoting and I find that quite hard, especially in the academic space. . . . And I also find it really hard to express myself in that many characters or say anything interesting. I do actually avoid it if I can, but I have to go on it. (P23)

Especially, younger participants were less likely to have taken major actions to restrict friends, sometimes mentioning they had gradually moved toward healthier ways of using social media:

[So] I would wake up in the morning and like look forward to having like 100 likes. And I thought, OK, this has to stop, because I was working, and I didn’t have enough time. And I also had a breakdown. (P38)

However, later in the conversation, P38 mentioned that she liked Snapchat because of the “good camera” and “filters,” and that she had recently posted a “funny” self-image that had garnered a lot of attention. Thus, participants often mentioned they did not use Facebook much or post edited selfies only to state in the next sentence that they surveilled people on Facebook or just posted a selfie. So, many participants continued habitual use of social media, highlighting how it is a landscape trap or an everyday infrastructure (Baym et al., 2020), and frequently found themselves in situations that upset them.

Thus, the sociality that participants encountered when connecting with friends on traditional platforms was one of entrepreneurial networking and self-promotion to foster connections and favorable attention in social and professional lives through foregrounding positive experiences and qualities of beauty, success, and happiness. The participants with EDs experienced this as distressing, fuelling a sense of inferiority, pressure, and non-belonging. The affordances of traditional platforms, such as Instagram, fuelled the competitive ethos through features, such as likes and visual displays, but the ethos also characterizes the wider individualist, post-capitalist societies (Turner, 2006).

Research on disconnection to reduce time spent on social media has argued that it subscribes to the same values or ethos as the platforms that are seen as the problem by being self-centered and focused on improving performance (Beattie, 2020; Kaun, 2021; Syvertsen, 2023; Syvertsen & Enli, 2020). The participants with EDs did consider their own well-being when unfriending on traditional platforms, but it could be asked if disconnecting from competitive connections is detrimental for common purpose. The participants did not refer to productivity or the quantity of posting (Sibona, 2014) as motivations to unfriend. The efforts of people with EDs’ to disconnect from friends as well as other content on social media was almost always motivated by mental distress rather than what could be referred as annoyance, such as time wasting. Unfriending could be informed by appearance-based distress (also Hockin-Boyers et al., 2021; Jorge et al., 2023), but more broadly the accounts of people with EDs highlight that they felt deeply uncomfortable with and sought to evade the general competitive individualistic sociality or ethos pervading the platforms. People with EDs could be more vulnerable to the pressures of this ethos than average users; also displays of happiness and success could be doubly hurtful for people, whose life, career, and relationship trajectories may have been interrupted by mental health problems. What the concept of ethos seeks to do is to direct attention toward the mental health toll of the harshly competitive value system that underpins the vulnerabilities and technologies.

Discussion

This article draws on and pushes forward disconnection research (Baym et al., 2020; Beattie, 2020; Figueiras & Brites, 2022; Jorge, 2019; Jorge et al., 2022; Kaun, 2021; Syvertsen, 2023; Syvertsen & Enli, 2020), which has argued that both disconnection and connection with social media are underpinned by contemporary individualist and performance-oriented spirit. Disconnection research has offered a useful historical and contextual perspective on conversations on social media and mental well-being, which has been dominated by psychological research, and which has focused on ostensibly universal psychological processes, such as social comparisons, associated with individual propensities and/or technological features (e.g., Seabrook et al., 2016).

However, disconnection research has mainly focused on critique, perhaps as it has often analyzed marketing or marketers of the practices (Beattie, 2020; Jorge et al., 2022; Syvertsen, 2023; Syvertsen & Enli, 2020). Research on users has been more complex. However, since it has involved “average” or healthy users, their main motivation for disconnecting has often typically been to reduce the quantity of time or engagement (Baym et al., 2020; Jorge et al., 2023; Mannell, 2018; Nguyen et al., 2022). Our contribution to this scholarship is to highlight that users may seek to foster and avoid qualitatively different social media socialities, which pushes beyond criticism and draws attention to alternatives.

Earlier research has found that users created different socialities across platforms through connecting and disconnecting with others (Light, 2014; Madianou & Miller, 2013). This research usefully highlights creativity of users in repurposing platform affordances. However, it hardly considers mental health and does not discuss the practices in relation to the broader historical context.

Our conceptual contribution is to put forward the concept of ethos to make sense of socialities underpinned by different values on social media. Foucault’s notion of ethics of the self (Foucault, 1990) draws attention to how users may self-reflexively create and apply principles, such as the principle of mutuality or returning support, characterizing cocooning, to guide their relations with others on social media. However, cocooning was also characterized by less self-focused and reflexive and more social, everyday inventing of intimacies together with other users, as discussed in anthropology (Light, 2014; Madianou & Miller, 2013).

Furthermore, Foucault emphasized that the ethos that individuals create always exist in a relationship with dominant socialities and the historical context. So, many of the principles of cocooning were defined as much by what they were not like as what they were like. Being responsive to others, able to share negative feelings and affirming others stood in contrast to evading or ghosting others if inconvenient, compulsory positivity, judging, and competitiveness, all characteristic of dominant sociality in social media that we have called entrepreneurial networking (Banet-Weiser, 2018; Dobson, 2015; Turkle, 2011). Yet, the cocooning practices of sharing difficult emotions to overcome them and affirmation embody what Illouz calls the therapeutic ethos, and such practices also characterize forms of treatment and online support for EDs that many participants were familiar with. Illouz argues that a therapeutic ethos helps individuals to adapt to rather than challenges contemporary competitive individualism (Illouz, 2008). Helping a person overcome, for example, a bout of anxiety may in both digital conversations between friends and therapy seek to get them back on their feet again rather than aim to revolutionize the system. Yet, listening and validating friends through digital networks also embodied an alternative to competitive individualism. As such, cocooning had elements of what Buber calls I-Thou relationship that seeks to respect rather than use or change the other (Buber, 1970). Yet, Buber puts the I-Thou forward as the ideal relationship, whereas the concept of ethos acknowledges contradictions. Users’ socialities do not exist outside of power or history and will always be shaped by prevailing values, opposing as well as aligning with them.

The ethos that participants with EDs perceived dominating social relations on traditional platforms, entrepreneurial networking, was experienced as anxiety producing and alienating. Participants had sought to evade this sociality by unfriending select acquaintances. Unfriending has been considered self-centered as well as leading to closed-mindedness (John & Gal, 2018; Kaun, 2021). However, it could be asked if rejecting the individualist competitive ethos of entrepreneurial networking is antithetical to commonality. Furthermore, it has been observed that marginalized and stigmatized groups unfriend more commonly than others (Zhu & Skoric, 2021), and such groups, including people with EDs, may sense the onslaught of dominant values (also Das & Hodkinson, 2019) more strongly than those groups that embrace and find it easier to fit into them. Furthermore, withdrawing from extensive connections on social media can also be seen as going against the grain of the dominant networking sociality, boldly open to the world, adapting to changing situations, and seizing their social and professional opportunities (Illouz, 2008). The difficulties the participants with EDs experienced with withdrawing from extensive networking testify for the dominance of this sociality.

Yet, even if participants with EDs could be considered marginalized due to mental health problems, they were also all educated, and cocooning may well tell about their privilege in terms of reflective habitus and availability of equally reflective friends and family. Furthermore, aspects of cocooning, such as being responsive to others, have been observed to be common among women (Van Bruyssel et al., 2023) and may be gendered, as most of our participants were women. However, the experiences of the few male participants were not different, and we would be cautious about suggesting that men with EDs have a different mode of interaction online without properly studying it. However, it is useful to bear in mind the positionality of different ethoses of social media use. Our research has suggested that people with EDs experience (dis)connecting with social media differently from the average users that are usually studied and begs for further research on differences and inequalities between groups of people and their use of social media in relation to mental well-being (Helsper, 2021; Jorge et al., 2023).

Overall, the perspective of ethos highlights the way in which social media relations articulate historical values, calling for changing such values and not just individual uses and attitudes or even technologies. The perspective also draws attention to alternative ethoses that articulate a discontent toward, attempts to adapt to, shield from and create alternatives to dominant forms, offering glimpses of how social (media) relations could be imagined otherwise.