‘Feel like going crazy’: Mental health discourses in an online support group for mothers during COVID-19

Abstract

COVID-19 has become a mental health pandemic. The impact on vulnerable demographic groups has been particularly severe. This paper focuses on women in employment in Hong Kong who have had to balance remote work and online schooling for over 2 years. Using semi-ethnography and theme-oriented discourse analysis, we examine 200 threads that concern members’ mental health on a popular Facebook support group for mothers. We demonstrate that mental health messages are typically framed as ‘troubles talk’. Other support group members actively align with a trouble-teller through ‘caring responses’, namely expressions of empathy and sympathy. These are realized through assessments of the trouble-teller’s experience, reports of similar experiences; expressions of compassion and advice-giving. Mental health talk online is heavily mitigated, nevertheless the medium provides a space for expressing mental health troubles and providing informal psychosocial support. We advocate the importance of microanalytic discourse studies for mental health research to get insights into people’s lived experiences during the pandemic.

Introduction

In January 2021, the United Nations published a short article, stating that ‘the COVID-19 pandemic has generated a “parallel epidemic” of deteriorating mental health’ (United Nations, 2021). Attested to by an increasing number of publications in gray and academic literature, mental health has become one of the priority issues in public health and society, along with the ongoing global combat to develop life-saving prevention and treatment protocols for COVID-19 as new mutations of the virus continue to emerge (e.g. Cullen et al., 2020; Pfefferbaum and North, 2020). The mental health impact of the pandemic on certain demographic groups has been particularly severe. In this paper we examine the mental health impact of COVID-19 on women in employment who have historically been considered one of the vulnerable demographic groups in the workforce (Hyland et al., 2020; Mills, 2003). Specifically, we focus on the impact of the pandemic on working women in Hong Kong, the city which has arguably been subjected to the most stringent government restrictions in the world since the onset of the pandemic.

The geographical proximity of Hong Kong to Mainland China where the virus first originated back in late 2019 meant that Hong Kong faced a very high risk of a large-scale outbreak from the early stages of the pandemic. With memories of Severe Acute Respiratory Syndrome (SARS) in 2003 still fresh in the minds of many Hong Kongers,¹ the government took immediate actions to control the transmission of COVID-19 in the Hong Kong community. The closure of schools and other educational institutions, online learning, work from home, restrictions on public gatherings, quarantine, and other measures have begun to be implemented in the city in January 2020, several months before any other country in the world. To aggravate the situation further, the pandemic was preceded by months’ long sociopolitical unrest in the city in 2019 that also led to the suspension of face-to-face classes and remote work for families. Fast-forward to mid-2022, Hong Kong is still battling with the pandemic and many restrictions are still in place, despite the fact that many other countries around the world are learning to ‘live with the virus’.

The impact of the ongoing pandemic on the mental health of the Hong Kong population has been devastating (e.g. Choi et al., 2020; Mind, 2021). In June 2020, we conducted a survey in collaboration with the City Mental Health Alliance Hong Kong (CMHA HK), a non-for-profit organization working to improve workplace mental health in Hong Kong, and Oliver Wyman, a global management consulting firm. The survey focused on the mental health of the working population in Hong Kong. Out of 732 respondents, 26% (one in four) reported mental health problems in the 12 months’ period prior to the survey, and women in particular reported experiencing more severe mental health problems. This finding was not surprising as in general women are more likely to suffer from mental health disorders, such as anxiety (Remes et al., 2016). 9% of the respondents left their jobs due to experiencing mental health problems. We also asked the respondents to rank their stress factors. Fear of contracting Covid-19 emerged as a top concern for the largest number of people (66%), which was even higher for people who had children under 10 years old (71%). Moreover, for those with children under 10, childcare emerged as a top concern for 70% of people, compared to only 21% for those without children under 10. The results of our survey reflect the global situation of mental health among parents (Czeisler et al., 2021; McKinsey and Company, 2021).

Against the backdrop of the survey that outlined the general trends in mental health, in this paper we explore a popular Facebook parenting support group for mothers in Hong Kong to gain insights into these mothers’ lived experiences of mental health struggles during the pandemic. We focus on the discussion threads that address the challenges of balancing work with family and childcare responsibilities. We examine the linguistic and other resources that the participants draw on to talk online about their mental health struggles, and the role of parenting support group in supporting members’ mental health during the COVID-19 health crisis.

Online support groups

Mental health support provided by online groups has been the topic of much research across disciplines. In psychology and psychiatry, the literature on the benefits and the limitations of such support presents conflicting findings (see overview in Finfgeld, 2000). Some studies suggest that online support groups are a form of social avoidance and as such, they are detrimental to mental health (Lawlor and Kirakowski, 2014). Other studies, on the contrary, maintain that these groups foster mental well-being (Barak et al., 2008) and enable reaching out to individuals who are concerned about social stigma and seeking help off-line (DeAndrea, 2015). What comes across in this body of research is that online support groups may offer some therapeutic benefits to those seeking help online when and if used appropriately. Noteworthy is that this research focuses on specialized mental health support groups, such as online groups for depression, suicidal tendencies, or substance abuse.

Discourse-oriented research adds a unique and nuanced perspective to the study of online support groups by examining the actual lived experiences of the members of online support groups. While some discourse studies focus on specialized mental health support groups (e.g. Hunt and Brookes, 2020; Smithson, 2015; see also contributions in Koteyko and Hunt, 2018), there are also studies of mental health support provided by general online groups (such as, e.g. parental support groups). Both of these strands of discourse research are relevant to our work; however, the current study is particularly interested in the latter group of studies as our focus is on the mental health of the general working mothers’ population who are members of a general online group. These mothers may or may not have experienced mental health issues prior to the COVID-19 pandemic (any prior issues are not reported in the online discussions that we examine here).

Recently, there has been an upsurge of discourse-oriented research on online groups (such as blogs, forums, messaging apps, video-sharing platforms) for parents, and mothers in particular. This growing interest has led some scholars to suggest a specific interdisciplinary field of study of ‘motherhood online’ (Mackenzie and Zhao, 2021) that is concerned with the production of digital media by mothers about motherhood. These social online communities of and for mothers serve as platforms to negotiate gendered hegemonies, maternal identities, cultural norms, and expectations (see Mackenzie, 2018; Ringrow, 2020), and to offer help and support to its members. The role of online communities in supporting mental health is alluded to in many studies but there is still a scarcity of discourse-oriented research that addresses these issues more directly. Generally, emotional support by the online community is acknowledged by many discourse studies (e.g. Petersen, 2015); however, very few studies examine emotions in detail as part of online communication (Bailey, 2022; Pedersen and Lupton, 2018), and even fewer studies link emotions expressed online with mental health. This link, however, is well acknowledged in clinical research that maintains that regulation of emotional state has a direct impact on one’s mental health (Gross and Muñoz, 1995).

The relative scarcity of research on mother’s mental health and online communication is somewhat surprising, considering that contemporary motherhood is a highly emotional experience (Pedersen and Lupton, 2018). The dominant societal ideologies across cultural contexts place high expectations on mothers to embrace motherhood as a happy and fulfilling experience. Citing the work of a sociologist Arlie Hochschild, Pedersen and Lupton (2018) describe these expectations as the ‘feeling rules of good motherhood’. If a mother’s emotions do not conform to these societal and cultural expectations and ‘feeling rules’, then this is perceived as a deviation and an abnormal behavior. Pederson and Lupton’s point relates closely to the broader issue of perpetuating mental health stigma when individuals who display ‘attribute[s] that [are] deeply discrediting’ (Goffman, 1963: 3) are judged and condemned by society. Moreover, mothers are subjected to the ‘intensive mothering ideal’ (Bailey, 2022; Hays, 1996) which requires that all attention, energy and resources should be directed toward one’s child(ren) foregoing the needs of other family members (primarily, parents), and that mothers carry the main responsibility for parenting. ‘Intensive mothering’ could as well be an additional stressor in women’s motherhood experiences, leading to less favorable emotional responses from women.

Indeed, studies note that mothers’ negative feelings regarding motherhood experiences, such as guilt, anger, maternal regret, ambivalence, resentment, anxiety, shame, hatred, are quite common in the online support groups (see Matley, 2020; Pedersen and Lupton, 2018; Petersen, 2015; Ringrow, 2020), but the extent to which these feelings are accepted and supported by the respective online communities varies. If we view mental health as a spectrum, studies show that online communities are generally happy to be emotionally supportive of women experiencing ‘milder’ mental health issues, typically described as ‘struggles’ or ‘difficulties’. There also appears a strong level of support to women going through key transitional stages of motherhood, such as childbirth, even if these transitions result in more severe mental health issues, such as postnatal depression. For example, Archer and Kao (2018) discuss the social support of mothers of newborns and young children with symptoms of stress, anxiety, and depression on social media. Jaworska (2018) in her work on stories about postnatal depression (PND) that women tell online notes the transformative and the therapeutic effects of sharing these stories. However, she also highlights the tensions that arise between ‘hegemonic discourses about motherhood and personal PND stories in which women appropriate and re-work these discourses to break silence and exercise agency’ (Jaworska, 2018: 26). Research suggests that when more serious expressions of emotional state are not related to clear triggers such as key life events, or when they greatly exceed the accepted ‘normal’ state (e.g. in the case of contemplating suicide) general online communities are not willing to engage with them closely, and instead suggest seeking professional help (Pedersen and Lupton, 2018).

In this paper we demonstrate that in the context of the COVID-19 pandemic, online communities have become indispensable in supporting mothers’ mental health. The pandemic and the imposed restrictions on direct social contact, have impacted the perceptions of physical proximity and online ‘dispersion’. Namely, for many people online interactions have become the ‘most proximate’ or sometimes the only experiences of social interactions, taking over the real physical face-to-face interactions. As our analysis below demonstrates, online spaces have also become the loci of mental health struggles and informal psychosocial support by online community members.

Data and method

This paper is part of a larger-scaled project on the impacts of the COVID-19 pandemic on the mental health of women in employment in Hong Kong. The project was approved by the Human Research Ethics Committee (HREC) of the University of Hong Kong (project reference number EA1808021).

The data that we draw on in this paper come from the largest Facebook support group for mothers in Hong Kong with over 65,000 members. The terms and the conditions of the group stipulate that all communication must be in English, although as the members’ profiles suggest, they originate from diverse sociocultural backgrounds. This reflects the general sociocultural diversity of Hong Kong (Census and Statistics Department, n.d.).

Several research methods were employed. As the first two authors are members of the support group, our research followed a semi-ethnographic method (Baxter, 2018): our participation in the group was initially not research-oriented, but we later became aware of the recurrent discussions surrounding group members’ mental health issues during the COVID-19 pandemic. We then read through all discussion threads posted from January to July 2020 and identified over 200 threads in which participants talked about their mental health. Through repeated readings of the selected threads we then mapped the data thematically along the main mental health themes that were raised in these threads, and analytically, along the main linguistic and other resources (e.g. photos, videos, hyperlinks, ‘likes’, and emoji ‘reactions’) that the participants employed to talk about mental health. We applied thematic and analytic mapping to both the initial posts and the comments. Our analysis of the data was also informed by research on computer-mediated communication and its affordances (and constraints) (Herring, 2004, 2007). In all examples examined in this paper, the stylistic and the grammatical features of the messages have been preserved. The analysis presented in this paper focuses on one representative thread in terms of the linguistic and other resources that participants draw on to talk about their mental health.

The Ethical issues surrounding the use of online data remain contentious (Bolander and Locher, 2014). The Ethical Guidelines of the Association of Internet Researchers (Franzke et al., 2020) outlines several options with regards to informed consent: when first-degree informed consent is difficult to obtain (as is the case, e.g. with a large number of participants), identifying information is anonymized when storing and processing the data. The guidelines also suggest that informed consent could be delayed to the stage of dissemination when selected participants are contacted for consent. Another approach is posting a message about the research project on the online platform and offering the participants an option of withdrawing from the research (Koene et al., 2015) which may also prove difficult in large-scale projects. In this project, as obtaining consent from all members was impractical, we sought consent from the group moderators. In addition, we have anonymized all identifying information in the data corpus and assigned pseudonyms to the participants in the threads to protect their anonymity.

Data analysis

In identifying the relevant threads in the group, we focused on lexical items that denote a certain mental state. Examples include: my anxiety; I am loosing/lost my sanity; I get very annoyed and angry; I am not happy; I feel like I am going crazy; it’s driving me crazy; one crazy mom; I am having a nervous breakdown; it’s stressful; it’s taking its toll; I’m struggling. As seen from these examples, mental health issues that are raised in this group are typically ‘mild’ (as opposed to more serious mental health issues, such as psychosis or suicidal thoughts); this finding aligns with the previous research on ‘motherhood online’ that we cited earlier in this paper.

Thematically, the analysis of discussion threads has revealed striking similarities with the survey results described in the Introduction: among the common identified themes are fear of the COVID-19 virus, childcare concerns, including home-schooling, and remote work arrangements. In the online data, these themes are typically framed as troubles-talk (Jefferson, 1988). Troubles-talk has been defined as ‘portrayals of (and interactions about) aspects of people’s lives as undesired and, perhaps, warranting change in behavior or perspectives’ (Miller and Silverman, 1995: 725). Jefferson and Lee (1981: 411) make a distinction between troubles-talk that focuses on the ‘teller and his experiences’, and advice-seeking that deals with the ‘problem and its properties’. Since troubles-talk portrays ‘undesirable aspects of people’s lives’, it is often followed by advice-giving, regardless of whether the advice is solicited or not (Pudlinski, 2005; Smithson et al., 2011). The concept of troubles-talk provides a useful analytical lens for examining how mental health-related talk is initiated by troubles-tellers, and how other members of the support group affiliate (or disaffiliate) with this kind of talk, and the mental health support that they provide to each other through their verbal and non-verbal comments.

In the thread that we examine in this paper, Sarah, the author of the initial post, is feeling stressed by home-schooling and remote work arrangement.

Example 1

Sarah White: Home schooling is stressful (during school suspension period), anyone? Or it’s only me. . . P.S. My girl is 5 and she’s going to enter a local Primary school. For the entire month, I will work from home. I also have a 2-year-old at home (I have a helper). Still, I feel every day, time flies and I always feel I could have done more with her. Is this normal?

Sarah’s post begins with a very brief inquiry into other mothers’ troubles-status. Her own troubles are mainly relayed in the postscript (P.S.) to the main message. This sequential organization of troubles-talk may not be coincidental. By seeking to generalize her personal experience to a larger group of the support group first, Sarah thereby also seeks to normalize how she is feeling, before going into any detail about her personal experience. Even mild mental health issues are perpetuated by stigma in the modern society (Sickel et al., 2014), and by admitting to being stressed Sarah risks to be ‘labeled’ and potentially stigmatized as a ‘bad mother’. She breaks the ‘feeling rules of good motherhood’ (Pedersen and Lupton, 2018: 2) and the intensive mothering ideal (Bailey, 2022; Hays, 1996), and doing it publicly to a large group of trouble recipients arguably exacerbates the sensitivity and the face-threat of Sarah’s message.

The postscript details Sarah’s clashing responsibilities of working from home and looking after her two children, and feeling like she could have been doing more with her 5-year-old daughter who is about to start primary school. Local Hong Kong schools are notoriously academically demanding and while Sarah has a domestic helper (which is a common practice in many Hong Kong families) she feels that ‘time flies’. This detailed account (Scott and Lyman, 1968) of home and work responsibilities in the postscript serves to mitigate and, to an extent, justify how Sarah is feeling. In the concluding sentence of this post, Sarah seeks to normalize her experience again with an open question to the mothers’ online community whether what she is feeling is ‘normal’. This example shows that while Sarah opens up about her mental health struggles, she mitigates her message in multiple ways through sequential organization of her message, seeking to generalize and normalize her experience and providing a detailed account of her troubles.

Sarah’s initial post attracted 41 ‘likes’ and 61 comments. This large number of comments to the posts that concern mental health are a common feature in the data. The comments focus on expressing what Pudlinski (2005) in his work on a peer support line describes as ‘caring responses’, namely the expressions of empathy and sympathy. In the online data that we examine here we see responses that: (a) provide assessments of Sarah’s experience (Example 2), (b) report similar experiences (Example 3); (c) convey compassion (Example 4); and (d) give advice (Example 5). The actions of expressing sympathy and empathy bear some similarities in telephone and online communication, however, the medium of the online communication allows members to exploit a broad range of non-verbal resources.

Example 2 lists members’ assessments of Sarah’s experience. Pudlinski (2005: 271) maintains that assessments, as a way of doing empathy and sympathy, ‘attempt to quantify how bad one thinks the news is’. In Sarah’s case these are assessments of whether what she is feeling is ‘normal’ or not.

Example 2

• (a) Anita Choi: Yes it is.

• (b) Mary Chung: Normal!

• (c) Marina Richard: Totally normal! If it helps, understand that homeschool world 2 hrs of homeschool typical equals a whole school day because they’re getting more personalized attention and you don’t have all the stuff [2 likes]

• (d) Liana Wong: Feeling you could have done more is. . . normal. There is always more that can be done. But, doing more doesn’t mean better. Enjoy this time you have with her before the school system takes up all of their time and energy. . .. [7 likes]

• (e) Christina Yau: It is definitely not normal that time flies. I have 4 children at home and each day last week felt like a year. [12 likes; 1 ‘shocked’ 19 ‘haha’/‘lol’ responses]

• (f) Carina Ma: Yes, definitely not normal!! I have one 2 years old, this past 10 days felt like a month. . . ((5 ‘loudly crying face emojis’)) [6 likes]

• (g) Matilda Wittkemper: It’s a bit overwhelming for sure. . . . . can’t believe all the apps, videos, worksheets etc a first grader can have. . .. and my 5th grader and I can get lost in all the assignments.

The assessments range from brief comments (responses a, b), to elaborate assessments that include accounts of why certain assessments have been made (responses c–g). These accounts justify and support the assessments. The accounts include explanation-giving (response c, d), advice-giving (response d), and reports of similar or different experiences (responses e–g).

In giving explanations the participants employ direct and personalized statements (understand that. . ., response c), numerical information (2 hours of homeschool. . . equals. . ., response c) that make explanations more targeted, personal and categorical; however, there are also generalized and idiomatic statements (there is always more than can be done, response d). The advice to ‘enjoy th[e] time’ with Sarah’s 5-year old daughter is also very general. Recurrent non-verbal feature is a series of three dots (. . .) that conveys different meanings in the comments: a pause that visually separates and, therefore, foregrounds the main message (feeling you could have done more is . . . normal, response (d)); an elliptical statement that either invites the recipients of the message to stop and think over what is implied (before the school system takes up all of their time and energy. . .., response (d)) or that contracts an otherwise too lengthy statement, a graphic substitute for ‘and so on’ (all the apps, videos, worksheets etc a first grader can have . . .).

The negative assessments of Sarah’s experience (responses e, f) as ‘not normal’ can be construed as non-affiliative responses. However, these responses are framed and interpreted by other members as humorous. This is conveyed through the use of emojis that embed heighted emotions or reactions (such as ‘shock’ or ‘loud cry’); these emojis are another non-verbal resource that the members employ to convey a certain mental health state in the group. While the members tell a different experience of finding that the time goes past very slowly, their ‘troubles’ are similar in that they too struggle with children at home during the lockdown. The last response in Example 2 reports a similar experience to Sarah’s: this mother feels ‘overwhelmed’ by the amount of assignments, apps, videos, and worksheets that her two children get. These ‘tellings’ of members’ own experiences convey their understanding of what Sarah is going through; they ‘display a shared understanding that only actual similarity of feeling and experience can bring’ (Pudlinski, 2005: 280). Similarity of experiences forges a social connection between other members and contributes to establishing rapport.

There is thus a wide range of verbal and non-verbal responses that the participants draw on to support their assessments of Sarah’s ‘trouble-telling’ to help her take a different perspective on the ‘trouble’ (Miller and Silverman, 1995) and cope with it.

When commenters respond by telling their own troubles, these are typically not reciprocated any further, that is they do not receive further comments or reactions. One reason for that could be that the primary function of reciprocated trouble-tellings in comments is to provide support to Sarah through affiliation and experience sharing rather than to seek support for themselves. Example 3 includes comments that report similar experiences to Sarah’s.

Example 3

• (a) Kamila Sarangi: 4 years old and 19 months old feel like going crazy

• (b) Anny Wong: I hope time flies. I am struggling with things to do with my 4 yr old and is going to be out of stock

• (c) Hannah Reids: I cant remember whice a day and date I feel the same. . .not day off for me. [‘smily face’] my 3 years old and 1 year old with out helper

• (d) Cheryl Huang: Totally understand. 6 and 4 year old and me working from home. In between keeping an eye on work + e-sch for the 6 year old, I’d completely forgotten about 4yo! Fortunately he’d taken the iPad and went off to do his own ‘work’ he said.

• (e) Koren Bingo: It is so hard to home school them and work from home. This last week has been very stressful. Not sure how I am going to do another 3+ weeks!!!

• (f) MC Chau: My 8 yro gal’s school loaded heaps of lessons ppts and assignments for my gal to study and call this online learning. I found that instead of teaching by different teachers, parents have to pick up the teachers roles and teach across subjects! This is stressful as each subject is followed by quizzes and assignments. The teachers throw us week-length lesson materials every Friday. . ..

• (g) Elena Kovalenko: Yeah very stressful (passing 2.5h a day with my son, and I’m getting crazier than him).

• (h) Carmen Lee: I can’t get any work done bc my 9yo requires a lot of supervision. It’s really stressful and frustrating. [1 like]

In these responses, mental health struggles are apparent as the mothers tell about ‘going crazy’, ‘struggling’, and ‘feeling stressed’. The trouble-tellings by these different women are very similar: the children under the age of 10 years old are locked at home, the mothers are struggling to find what to do with the children, or conversely, to cope with the workload assigned by the schools. To add to these troubles, some mothers are also working from home while supervising their kids. The trouble-tellings are brief, which, on the one hand, could be attributed to the medium of the Facebook group; on the other hand, it could also be due to the limited time that these mothers can afford to communicate with each other amidst all of their responsibilities. Typos, contractions, elliptical syntactic structure contribute to the brevity and the time constraint impression. The brevity of the majority of trouble-tellings in responses is in contrast to the detailed narration of troubles in Sarah’s original post. This supports our observation that by posting their comments members seek to help Sarah cope with her troubles through establishing similarity of their experiences rather than to seek support for themselves by providing very detailed accounts. Similar to Example 2, there is a humorous use of emojis (response c). Another example of humor is Cheryl’s (response d) anecdote (Norrick, 1993, 1994) about her 4-year-old who she forgot about and who occupied himself with an iPad ‘to do his own “work”’. Humor here contributes to establishing solidarity (Dynel, 2009); it also mitigates these mothers’ admission to struggling mentally and not coping with the mounting responsibilities. While these admissions could be seen as challenging the hegemonic ideologies of ‘good motherhood’ (Schnurr et al., 2016) and ‘intensive mothering’ (Bailey, 2022; Hays, 1996) in non-crisis times; the context of the pandemic and the fact that these experiences appear to be so common among the forum members contribute to normalizing them.

Another type of comment that is seen in the data are compassionate statements that are reminiscent of the Hong Kong government COVID-19 slogan ‘Together, we fight the virus’. Example 4 presents one example of a compassionate statement:

Example 4

Marina Reidy: You are doing your best Mama, that is all that matters. We are all in this together [‘thumb up’ emoji] [4 likes]

This brief comment provides emotional support and reassurance to Sarah (You are doing your best; see also the ‘thumb up’ emoji). The supportive slogan (We are all in this together) and the informal address term (mama) create alignment between participants. The slogan emphasizes that the experience that Sarah is going through is shared by other members of the mothers’ community. Similar to other comments examined in Examples 2–4, establishing similarity of personal experiences among the support group members creates social connections and rapport.

The last set of examples that we examine here are advice-giving comments. As mentioned earlier in the paper, troubles-talk is often followed by advice, regardless of whether that advice was solicited or not (Pudlinski, 2005). While some advice sequences in the data are brief, and focus on maintaining Sarah’s mental health (e.g. relax, don’t stress out, etc.), there are also examples of more extended advice sequences, as Example 5 illustrates.

Example 5

• (a) Melissa Kong: Hiya I hear you - I have 2 boys 8 and 9 years old with no other help. I only focus on the core subjects like maths and English. They get online chinese class 3 times a week and play football on the beach everyday. I squeeze in the artsy subjects (like drama, art and music) if we have time and the energy. Choose your battles and keep your sanity. Do what you can and don’t be so hard on yourself. We are living in a situation beyond our control [‘flexed biceps’ emoji] you are doing great mama.

• (b) Alexandra Mills: Chill! She’s 5. Did any lessons you had when you were 5 shape your future? Do what you can to keep her entertained, happy and, for your sake, quiet. And don’t worry how it’ll affect her. It won’t. Stay sane.

The first example of advice (a) comes from Melissa who acknowledges Sarah’s experience (I hear you) and then shares a short personal narrative of her own experience. Like Sarah, she has two children and no help. She proceeds to tell Sarah how she organizes her children’s schooling and social activities. This sharing of experience in addition to creating alignment and rapport, also serves as an indirect advice to Sarah about how schooling and extra activities can be organized and what should be prioritized. The indirectness of this practical advice mitigates its any potential face-threatening connotations that Melissa has managed to organize her children’s day better than Sarah. The indirect advice is followed by direct advice that focuses on Sarah’s mental wellbeing (keep your sanity, don’t be so hard on yourself, lines 4–5). Arguably, the directness of this second piece of advice is necessitated by the importance of the message to prioritize one’s mental health. Melissa concludes by a generalized statement about the pandemic situation, the collective pronouns (we, our) here contribute to creating alignment. A compassionate statement (you are doing great mama) in the end is an expression of emotional support and reassurance to Sarah.

The second example of advice (b) from Alexandra is similarly direct in relation to Sarah’s mental health (chill, don’t worry, stay sane). Alexandra asks a rhetorical question about Sarah’s own experiences at the age of five that implies that nothing at this age shapes a child’s future. Alexandra’s message echoes Melissa in that she suggests prioritizing certain things over trying to do everything. The statement that Sarah’s experience will not affect her daughter is brief and categorical, and as such conveys a strong message that Sarah needs to focus on her own mental wellbeing in the circumstances.

Like other threads in the data corpus, this thread has not ‘closure’ in the sense that Sarah does not engage with the comments that her message receives. It appears that after a certain time, the relevance of the thread is exhausted, and it attracts no further comments. The lack of return comments from Sarah could be attributed to a number of factors, such as the medium of communication, the quasi-synchronous nature of information exchange on Facebook, and the lapse of time between the original post and the multiple comments it receives. The lack of a ‘return comment’ from Saran could also be due to the fact that her concerns have been addressed in the members comments. It could also be due to the non-anonymous form of communication on Facebook and the extent to which the authors of original posts are prepared to open up any further about their mental health struggles. The support groups like the one we have examined in this paper then provide an informal first stop space for members to raise their mental health concerns and receive empathy/sympathy and advice. They are, however, not a substitute for continuous communication and mental health support, including support for more serious mental health issues. However, this informal psychosocial support has been particularly important when formal services may not have been readily available during the pandemic.

Discussion and conclusions

COVID-19 has brought a new meaning to the WHO’s statement, ‘there is no health without mental health’ (WHO, 2018). Deteriorating mental health of the population has emerged as one of the gruesome impacts of the health pandemic, and hence there has been an increasing research interest to how the negative impact could be addressed and ameliorated. This paper has aimed to contribute to this body of research by focusing on the mental health of one of the vulnerable demographic groups. Specifically, we have examined the impact of COVID-19 on working women in Hong Kong who due to the government-imposed long-term restrictions have had to deal with remote work, family and childcare responsibilities and home schooling for over 2 years. While the bigger study is interdisciplinary in nature, in this paper we have focused on online interactions of working mothers in one Facebook support group for mothers. In this discussion, we first outline our findings and then make some methodological points about using online discourse data in research on mental health in interdisciplinary projects like ours.

As a springboard for our analysis, we have used the findings of a large-scale survey on the impact of COVID-19 on the mental health of the working population in Hong Kong. Broadly speaking, the quantitative data from the survey have outlined some general trends of the current mental health landscape in Hong Kong, while the qualitative data has allowed us to explore the actual lived experiences of working women as reflected in their online interactions in the support group. Strikingly, the qualitative investigation of the online data has highlighted the same focal themes with regards to mental health as in the survey. Specifically, both in the survey and the online support group, mental health struggles related to fear of COVID-19, and managing childcare and homeschooling while working from home have been identified. In the Facebook interactions that we have examined, two omnipresent ideals of a ‘good mother’ and a ‘good worker’ have emerged that the women have tried to balance during the pandemic. It is arguably the burden of trying to conform to these two ideals that have contributed to the immense burden on the mothers and the deterioration of their mental health.

We have observed that mental health talk is framed as troubles talk in the online interactions, troubles talk thus served as the analytic lens through which we examined how the mothers talked about their mental health and sought support from other members. While troubles talk typically concerns ‘undesirable aspects’ that ‘warrant change in behavior or perspective’ (Miller and Silverman, 1995: 725), in the context of the ongoing pandemic the ‘trouble’ may not be as easily addressed. The troubles-talk is arguably of a different nature, where a source of trouble is external, of a global magnitude, life-threatening, difficult to control, or, drawing on the most used word since the start of the pandemic, ‘unprecedented’. Rather than changing the trouble, the women go online to make sense of their experiences, to normalize them and to seek informal psychosocial support from other members going through the same trouble.

We have observed distinct ‘grammar’ of mental health talk in the data: use of lexical items denoting (typically mild) mental health states; use of generalized and idiomatic language, elliptical sentences, backgrounding the details of mental health troubles, humor, non-verbal features (e.g. emojis). Some of these features could be attributed to the online medium of communication; however, another explanation is the stigma surrounding mental health talk, and the degree to which the participants of the interactions are prepared to disclose their mental health status to the online community. The issues that the women have disclosed are relatively mild and could be framed as ‘struggles’ or ‘difficulties’. The women have mitigated their messages using specific sequential organization and backgrounding the details of the trouble, attempting to generalize and thus normalize their experiences, and providing detailed accounts of their troubles as a way of justifying their mental health state. While online communication about one’s mental health may require less emotional investment than face-to-face communication, the medium of Facebook does not provide complete anonymity, particularly in a relatively small place like Hong Kong. Moreover, open ‘trouble-tellings’ also challenge the dominant ideologies of ‘good motherhood’ (Schnurr et al., 2016) that equates to ‘intensive motherhood’ in contemporary societies (Bailey, 2022; Hays, 1996). A question then arises why these women choose to take their troubles online. We would like to argue that the pandemic has changed the perceptions of proximity and distance, and perceptions of online communication. When face-to-face communication became unavailable due to massive lockdowns, online communication has become the most ‘proximate’ available way of communication. Online interactions on support groups have thus become indispensable in receiving informal psychosocial support. As we have seen in the data, the women in the group actively align with the ‘trouble teller’ and construct solidarity and rapport. This is achieved through various mean of expressing empathy and sympathy, such as assessing and normalizing the trouble-teller’s experience, reporting similar experiences, compassionate statements and advice-giving that downplays the trouble and suggests to the woman to prioritize her own feelings and mental health over competing responsibilities of remote work and home-schooling.

We strongly believe that the informal psychosocial support afforded by the online communities in crisis times like the COVID-19 pandemic should be given more attention in mental health research and practice. As our analysis has demonstrated the online groups are an important first stop resource of mental health support.

Finally, there are two methodological points to be made here. First, as we noted earlier, our qualitative analysis has focused on the exploration of lived experiences of working women in Hong Kong. Research on lived experiences of people with mental health issues is increasingly advocated in psychology as a useful bottom-up approach to explore how people make sense of these experiences (e.g. Reid et al., 2005). This is where discourse analysis could lend useful methodological insights from a long-standing research tradition of exploring lived experiences that would allow psychologists to gain a deeper understanding of how people tell their life stories, what form these stories take and what functions they perform (Zayts and Norrick, 2020). Second, in discourse analysis, discourse (in its many forms, including written and oral) is conceptualized as a practical, social activity that accomplishes social life and constructs different versions of knowledge. Discourse data are not just benign descriptions of what has happened when, they are rich ‘first-hand accounts of mental health knowledge’ (Hunt and Brookes, 2020: 2), and as such they present a particularly valuable resource for researchers across disciplines, workplace practitioners (such as human resource managers) and mental health professionals with regards to what support is needed. The value of the data, such as the online interactions examined in this paper, should be highlighted to non-discourse analysts: for example, in mental health research, online data like the Facebook group interactions that we have examined here could provide insights into specific mental health landscape, the issues that people report, the informal psychosocial support, etc. These insights could lead to targeted practical interventions to support mental health of specific demographic groups.

Author biographies

Olga A Zayts-Spence is an Associate Professor at the University of Hong Kong. She directs the Research and Impact Initiative on Communication in Healthcare (www.hkuriich.org). She is the author of Language and Culture at Work (with S. Schnurr) and has published widely on various aspects of healthcare communication in Asia.

Vincent Wai Sum Tse is pursuing his PhD at Monash University (Australia). He is also a member of the Research and Impact Initiative on Communication in Healthcare at the University of Hong Kong. His research interests lie in critical discourse studies and institutional communication (especially in education and healthcare).

Zoe Fortune is an Adjunct Assistant Professor at the University of Hong Kong. She has a background in Psychology and received her PhD in Health Services Research from the Health Services & Population Research Department at King’s College London. Her research focuses on workplace mental health.