Introduction
Having focused on references to anxiety in the previous chapter, we now offer an examination of the wider Anxiety Support forum as a type of computer-mediated communication (CMC), in order to identify the topics and forms of expression that are prevalent in online discussions of experiences of anxiety. We apply keyness analysis to the investigation of words and fixed phrases, as an exploratory, data-driven approach to identifying what it is that members of the forum choose to discuss. Through this analysis, we obtain a sense of what participants expect from the forum: in terms of both the kinds of support they seek from other members and the nature of the exchanges they have with them, as we identify interactional norms. In this way, we offer a summary of how the lived experience of anxiety is represented in the forum, as well as a view of the lived experience of the Anxiety Support forum itself.
Online Forums as Computer-Mediated Communication
The internet has become a key resource for providing health information and health-anxious individuals are particularly likely to use online resources to satisfy their health information needs (Reference Baumgartner and HartmannBaumgartner and Hartmann, 2011). Users have had to learn to navigate the myriad sources of health support and information that have emerged online in the form of webpages, blogs, social media, online forums and so on (Reference Sillence, Briggs, Harris and FishwickSillence et al., 2007) and digital tools also allow users to track personal health information related to, for example, heart rate, sleep patterns and ovulation through mobile apps (Reference Bach and WenzBach and Wenz, 2020). The development of ‘eHealth’ reflects broader trends in the evolution of digital and online technologies that have progressed from unidirectional repositories of information to participatory, collaborative spaces for open exchange, changing the way that individuals communicate, and contributing to addressing the health inequalities that manifest from uneven access to traditional health services and institutions (Reference Chou, Prestin, Lyons and WenChou et al., 2013). As such, investigations of how users generate their own content and communicate their own experiences through computer technologies can help us to develop tools and provide more targeted information to facilitate better health outcomes.
The late 1990s saw the emergence of studies of CMC in earnest, as the availability of home computers and the growth of the World Wide Web brought huge numbers of users to spaces where their interactions with others were interposed by a computer screen, or a mobile phone (Reference Herring and DonsbachHerring, 2008). Linguistics research at this point was preoccupied with defining ‘Netspeak’, setting out to position CMC on a continuum between spoken and written language and focusing on the textual features that distinguished forms of CMC from communication in other contexts, such as the use of acronyms (brb – be right back), homophones (2 for to/too, r for are), emoticons (:), :-P) and later, emoji (😊, 👋). However, as Reference CrystalCrystal (2001, p. 48) concluded, ‘Netspeak is identical to neither speech nor writing, but selectively and adaptively displays properties of both.’ This consideration of structural features is one of the domains of research that has typified linguistic interest in CMC, which Reference Herring and DonsbachHerring (2008) positions alongside:
classification research concerned with modes and genres of online communication;
discourse patterns, including interactional turn-taking, politeness, dialect;
internet behaviour in terms of identity, community, power;
language ecologies – for example, the multilingual Web, the spread of English.
One of the factors that has forced researchers to expand their investigations into CMC is the range of modes and platforms that – even in the 1990s – could be distinguished within the broad scope of CMC, including email, blogs, instant messaging, SMS texts, social media and virtual worlds. Each of these modes has their own structural characteristics that inevitably shape the textual features that we are likely to observe. Continued interest in finding the boundaries between these digital arenas has led to the recognition that there is a great deal of ‘hybridization’ and multifunctionality in CMC texts, whether conceptualised as genres that are defined by conventional organisational features (Reference Thurlow, Poff, Herring, Stein and VirtanenThurlow and Poff, 2013) or registers that are defined by situational characteristics and examined for linguistic indicators of such characteristics (Reference Biber and EgbertBiber and Egbert, 2018).
One of the legacies of the positioning of CMC in relation to spoken and written language is to view online communication in terms of what it is lacking; for example, the resources that we have at our disposal in face-to-face communication that are not possible in spaces such as online forums. For instance, we can refer to non-verbal (e.g., facial expressions), auditory (e.g., intonation, volume) and tactile (e.g., a hand on the shoulder) signals when we speak to someone face to face. Of course, these are variously available depending on the type of CMC, with facial expressions and prosodic aspects discernible in video conferencing, for example. Textual features of CMC such as emoticons/emoji, non-standard orthography (whaaaat?) and stylised metacommentary (*checks notes*) have, nevertheless, been described as compensatory, in lieu of paralinguistic features (as discussed by Reference Bieswanger, Herring, Stein and VirtanenBieswanger, 2013). However, this view diminishes the creative and ludic aspects of what are highly flexible features of CMC and other, iconic aspects of CMC, such as memes, hyperlinks and GIFs are near-impossible to replicate in other modes of communication. Furthermore, researchers have stressed that such resources are multifaceted and continually adapted in CMC towards new meanings and functions (e.g., Reference Thurlow, Poff, Herring, Stein and VirtanenThurlow and Poff, 2013). Reference Dresner and HerringDresner and Herring (2010) stress that emoticons have multiple functions beyond their association with conveying emotion, most notably as illocutionary force indicators, prompting recipients to interpret a message as ironic, or playful.
While micro-linguistic features such as emoticons, non-standard spelling, abbreviation and non-standard punctuation have long been held as characteristic of CMC, Reference Bieswanger, Herring, Stein and VirtanenBieswanger (2013, p. 478) makes the point that variation in the use of such features occurs ‘according to the language and script used, the mode of CMC employed, and other use- and user-related factors’. In this way, we can consider online forums as fundamentally different from email, social media and SMS text messaging. Furthermore, even within sub-genres of CMC, there is heterogeneity; Reference Bieswanger, Intemann, Luginbühl and PerrinBieswanger and Intemann (2011) found that on average 2% of the words across English-language online forums were emoticons, but that this varied across individual discussions from an average of 0.3% in one forum, to 3.1% in another. They ultimately conclude that there is no such thing as a language of online discussion forums (Reference Bieswanger, Intemann, Luginbühl and PerrinBieswanger and Intemann, 2011).
Instead, studies of online forums are arguably better served to evaluate them in terms of their specific topic focus and in relation to the membership of their respective communities. Reference Brookes and DemjénHunt and Brookes (2020, p. 63) agree, stating that forum posts ‘must be understood as contributions to online communities that have established norms and expectations for communicating about illness’. Indeed, HealthUnlocked provides explicit guidance to users about the manner in which they articulate their recommendations, with the example ‘“For me, this worked … ” rather than “You should do this … ”’ (https://healthunlocked.com/anxietysupport/about). Less explicitly, contributions from established and regular posters offer a model to novice members of what is germane and appropriate to the forum. As Reference Brookes and DemjénHunt and Brookes (2020, p. 63) assert, ‘Expectations based on existing interactions in an online community and an individual member’s status within it can therefore play a role in the ways in which support group members linguistically frame their messages’, in order to show to other members that they belong. Corpus linguistics, which is predicated on identifying patterns of language use, is therefore a fitting approach through which to uncover the regular vocabulary and formulations which contribute to the linguistic fingerprint of the Anxiety Support forum.
In Chapter 1, we established that online support forums are characterised in terms of the information and emotional support they offer (Reference YipYip, 2020), that members report benefits according to group belonging and social connectedness (Reference Naslund, Aschbrenner, Marsch and BartlesNaslund et al., 2016) and that many users of online forums feel more comfortable discussing sensitive and personal issues in online spaces that afford them relative anonymity (Reference Webb, Burns and CollinWebb et al., 2008). Linguistic investigations of online health forums have highlighted aspects that are more typical of the platform when compared with other modes of communication, such as members constructing their identity through diagnosis, advice-seeking and advice-giving (Reference Stommel, Lamerichs, Hamilton and ChouStommel and Lamerichs, 2014). Discussions in online forums are reported to be characterised by ‘problem’ messages, discussing issues with diagnosis or medication for example, and disclosing personal troubles is an important component in the enactment of social support (Reference GoldsmithGoldsmith, 2004). It is argued that the communicative context of internet forums encourages people with mental health concerns to offer sensitive self-disclosures ‘with a level of candour that would be unlikely in more inhibiting offline settings’ (Reference Brookes and DemjénHunt and Brookes, 2020, p. 62), however, with advice-seeking and advice-giving there are found to be varying degrees of directness (Reference Stommel, Lamerichs, Hamilton and ChouStommel and Lamerichs, 2014). Reference YipYip (2020, p. 1216), for instance, found that 70% of thread openers in forums for dealing with anxiety and depression requested support but in indirect means, such as through self-disclosure. This demonstrates the multifunctionality of such communicative acts, which encourage social connectedness in their frankness – that is, by providing intimate personal details – but also through their invitation for reciprocity and wisdom.
Tied to this kind of problem-solution exchange are concerns for credibility and legitimacy, which can be established through accounts of long-standing membership to the forum, of a history of interactions with medical professionals or a knowledge of treatments, for example (Reference Galegher, Sproull and KieslerGalegher et al., 1998). There is a concern, however, that users’ commitments to demonstrating membership subsequently limit the range of perspectives communicated in online forum spaces, that contributors are encouraged to not only express but, indeed, even hold the ‘right’ point of view. Reference WeberWeber (2011), for instance, found that senior members of a sexual abuse support group reformulated or recast new members’ posts to bring them into alignment with normative values. Newcomers to an online support group for those recovering from eating disorders who had reported previously belonging to a pro-anorexia community were obliged to disavow that ideology as a precondition for membership (Reference Stommel and KooleStommel and Koole, 2010). Reference McDonald and Woodward-KronMcDonald and Woodward-Kron (2016) show that the socialisation into a bipolar disorder support forum community and the moderation of views expressed in that forum can be evidenced in lexicogrammatical and discourse-semantic choices, modelled by senior members and adopted by new members. As such, we investigate textual features with the aim of understanding what it means to be a participant in the Anxiety Support forum, as a platform for discussing experiences of anxiety, as well as a type of CMC.
In addition to textual units such as abbreviations, hyperlinks and emoji characters, we can also ask questions about the structure of texts composed in asynchronous modes such as online forums. Compared with the often quick exchange of short turns in spoken conversation, users of online forums may deliberate longer on the content and structure of their posts as they strive to optimise the meaning of their communication and elicit multiple responses, albeit staggered over longer stretches of time. Furthermore, Reference Stommel, Lamerichs, Hamilton and ChouStommel and Lamerichs (2014, p. 201) observe that asking and advice-giving are ‘sequentially related’ and that the interactional aspects of a discussion thread are important for understanding how communicative goals are achieved and negotiated with other members. We will discuss these structural and sequential aspects of the forum posts more in Chapter 4. In this chapter, we are first and foremost concerned with harnessing a view of how the content of the Anxiety Support forum, as seen through recurrent lexical items, reflects what it is like for participants, in terms of the topics they discuss and the ways in which they formulate their messages to one another. Guiding our discussion of recurrent features are the following questions:
What do recurring features of posts tell us about why members participate in the forum and what they set out to discuss?
What do patterns of language tell us about how lived experiences of anxiety are typically articulated in the context of this forum?
Before we report our observations of the Anxiety Support forum data, we will briefly introduce and outline the methodological concepts and procedures that we have applied to identify recurrent linguistic patterns. We begin with a discussion of one of the fundamental principles of corpus linguistic analysis, ‘keyness’, which provides the basis on which we identified not only the findings reported here but also in later chapters as we consider how patterns in language use relate to different groups of contributors at different stages of the life of the forum (Chapters 5–7).
Keyness
The notion of keyness is often described in relation to ‘aboutness’ (Reference Gabrielatos, Taylor and MarchiGabrielatos, 2018), in that the procedure is generally designed to highlight features that are indicative of both the topic and the style of the content of a corpus, telling you what it is ‘about’. As we will demonstrate in this chapter, keyness can be applied not only to lexical items, but also to any feature of a text for which we can determine frequency. However, for simplicity, we will briefly summarise the procedure for keyness analysis in terms of keywords. Reference ScottScott (1997, p. 236) reports that a keyword may be defined as ‘a word that occurs with unusual frequency in a given text’ and that, crucially, we are interested in unusual frequency, by comparison with an established norm. As discussed in Chapter 1, the selection of a reference corpus has implications for what emerges from the analysis and we have selected the English Web 2020 for comparison in order to establish a norm for web-based English, thus foregrounding the distinct – principally topical – features of the Anxiety Support forum.
Once a reference corpus is selected, keyness is determined through a comparison of the relative frequencies of select features (e.g., words) and corpus linguists typically establish a threshold for what is unusual according to a frequency or statistical measure. Reference ScottScott (1997) writes that the calculation of keyness is done using the chi-squared statistic and this has historically been computed in the form of Log Likelihood (Reference DunningDunning, 1993). However, as Reference Gabrielatos, Taylor and MarchiGabrielatos (2018) discusses, there is a range of metrics available for determining keyness, each of which attends to a different aspect of the quantification of the difference – and by implication, similarity – between the two relative frequencies. The Log Likelihood calculation, for example, determines statistical significance, that is, that we have sufficient evidence, by way of the size of the corpus and the number of occurrences, to confidently assert that the observed differences are not down to chance. We can also apply effect size measures to determine the magnitude of the difference between relative frequencies, which can be combined with a measure of statistical significance to establish a level of confidence in those results. We applied the Simple maths calculation (the default measure in the corpus analysis tool Sketch Engine) to determine keyness, adjusting the smoothing parameter to prioritise higher frequency words (k=100) (Reference Kilgarriff, Mahlberg, González-Díaz and SmithKilgarriff, 2009). This is a fairly accessible measure, which provides an indication of how many times more frequent an item appears in one corpus compared with another. We also wanted to ensure that the features we investigate can reasonably be considered to be representative of the corpus as a whole and so we have excluded words for which 50% of occurrences come from one poster. These keyness parameters have also been applied in subsequent chapters, unless otherwise stated.
N-grams
Our analysis begins with a discussion of keywords. However, what also emerged through our investigation was the importance of multi-word phrases that occurred repeatedly in posts to the Anxiety Support forum. Linguistic interest in repeated sequences of words is captured in the concept of ‘n-grams’, which uses the mathematical notation n to indicate an unspecified number of elements in a string of lexical items. In other words, we can search for repeated strings of two (2-grams), three (3-grams), four (4-grams) or more words. Investigations of n-grams have helped linguists to demonstrate that much of language is phraseological, that speakers have numerous ready-made constructs at their disposal, which they frequently utilise, and that readers process these frequently repeated strings more quickly (Reference Underwood, Schmitt, Galpin and SchmittUnderwood et al., 2004). N-gram analysis has been shown to be effective in offering a bottom-up approach to distinguishing (sub)genres of English, with 3-grams highlighting differences between spoken and written language, as well as distinguishing academic writing from creative writing (Reference Gries, Newman and ShaoulGries et al., 2011). For example, Reference BiberBiber (2009) has shown that I don’t know why and the way in which are characteristic of informal spoken conversation and academic writing respectively, and that corpus analysis tools are effective for identifying n-grams (or ‘lexical bundles’) as indicative of the communicative goals and conventions of categories of text types. As such, we report key n-grams from the Anxiety Support forum, as determined by their unusually high frequency when compared with the English Web 2020 corpus.
Emoticons and Emoji
We discuss emoticons (comprising text characters such as :)) and emoji (pictograms such as 😊) separately from keywords and key n-grams, since their semantic and pragmatic meaning is arguably more flexible, depending on how they are combined with other characters or lexical items. Reference Gawne and McCullochGawne and McCulloch (2019), for example, discuss the illustrative (Coffee? ☕), metaphoric (Stay strong! 💪), deictic (☝ This is perfect), emphatic ‘beat’ (WHAT 👏 ARE 👏 YOU 👏 DOING) and illocutionary (That was fun 
) functions of emoji, which cannot readily be discerned from looking at the individual emoji character. We report our observations of frequently occurring emoticons, discussed in relation to how they were used by different categories of contributor to the forum. Studies have observed, for instance, that females produce emoji and emoticons more frequently than males while younger users post more emotions than older users (discussed in Reference Herring and DainasHerring and Dainas, 2020). While emoticons are associated with CMC, Reference FrehnerFrehner (2008) observed a higher relative frequency of emoticons in SMS texts (4.88 per 1,000 words) compared with emails (3.18 per 1,000 words), indicating that their use differs according to the type of CMC and providing a point of comparison for what we observed in the Anxiety Support forum.
What Do Members of the Anxiety Support Forum Set Out to Discuss?
Keywords and Key Terms
We divided the Anxiety Support forum data into two categories of messages: Main posts and Replies. This split was based on observations that Main posts (i.e., posts that initiate a discussion thread) were characterised by agenda/topic-setting, while Replies (i.e., the subsequent responses in a discussion thread) appeared to function very differently, based on their adoption of the ideas and experiences established in the Main post. Other research has made similar observations; for example, Reference CollinsCollins (2019) found differences in self-initiated posts and replies in an online learning discussion forum, with more expressions of affiliation in replies to existing posts. There were 42,239 Main posts in the data, constituting 14.36% of contributions and 5,133,149 words. Correspondingly, there were 251,843 Replies (85.64% of posts), containing 15,902,860 words. Establishing these sub-corpora and comparing each with the English Web 2020 corpus allowed us to identify similarities and differences in the prevalent terms of each message type and determine which patterns were characteristic of the Anxiety Support forum overall. The corpus analysis tool we used, Sketch Engine, offers the capacity to identify terms as well as individual words, capturing modified noun phrases such as anxiety attack or shortness of breath. These are distinct from n-grams, in that they represent singular objects and concepts, whereas n-grams are typically not complete or isolated structural units (Reference BiberBiber, 2009).
For both the Main posts sub-corpus and the Reply post sub-corpus we derived a list of the top 100 keywords and the top 100 key terms. We examined occurrences of each of the key items in their original context to determine their prevailing semantic meaning. There were 113 items that were key for both Main posts and Replies, meaning that there were 87 keywords and key terms for Main posts that did not appear in the top 100 lists for Replies and, equally, 87 key items that appeared only in our lists for Replies. We were able to categorise both the similar and the distinct key items to our thematic groupings, indicating that even when the precise terms differed, the key items for Main posts and Replies related to a restricted number of semantically related concepts. We discuss the similarities and differences in those terms according to our thematic categories, which were as follows:
health issues and symptoms
time
persons and things
medical tests, remedies and coping strategies
feelings and emotion; cognition
forum relations
personal circumstances
body parts and processes
These categories broadly reflect a focus on the physical manifestations of anxiety, which is reported in terms of various timeframes, affecting different parts of the body and disrupting natural bodily processes, such as sleep. Furthermore, there are frequent disclosures concerning the interaction with health professionals as members seek diagnosis and treatment and forum members are also invited to give their perspectives on identifying and coping with anxiety disorders.
We begin by looking at the keywords and terms we categorised as relating to Health issues and symptoms, which are presented in Table 3.1 along with their frequencies. The key words and terms are shown in order of keyness score. For example, the first keyword, anxiety, occurs 44,511 times in the Main posts, or 8,671.28 times per million words. In the English Web 2020 corpus, anxiety occurs 948,900 times or 26.96 times per million words. In other words, anxiety occurs 321 times as often in the Main posts compared to the English Web 2020 corpus. Compare this to the keyword for the Main posts, suffer, which occurs 28.13 times per million words in the English Web 2020 corpus and 41.02 times per million words in the Main posts – or 1.45 times as often.
Table 3.1 Keywords and key terms relating to Health issues and symptoms
| Main posts | Replies | |
|---|---|---|
| Health issues and symptoms | anxiety (44,511), panic (10,836), panic attack (6,427), symptoms (7,936), pain (7,459), sick (4,409), attack (6,370), depression (4,011), attacks (4,690), dizzy (2,919), tired (3,167), pains (2,660), health anxiety (2,078), chest pain (1,780), dizziness (1,687), suffer (2,106), heart attack (1,516), side effect (1,470), anxiety attack (1,152), physical symptom (771), anxiety symptom (728), anxiety disorder (715), shortness of breath (603), bad anxiety (571), severe anxiety (540), social anxiety (533), heart palpitation (525), mental health (717), panic disorder (478), head pressure (410), brain tumor (406), acid reflux (377), symptom of anxiety (312), sharp pain (272), heart problem (267), brain tumour (258), racing heart (251), back pain (263), other symptom (248), stomach pain (233), blood clot (238), burning sensation (229), health issue (261), new symptom (205), dizzy spell (198), sore throat (204), first panic (198), brain fog (197), anxiety issue (192), arm pain (187), first panic attack (185), heart race (184), tension headache (177), fast heart (174), heart racing (173), shooting pain (172) | anxiety (109,907), symptoms (28,024), panic (20,396), pain (14,119), panic attack (10,723), depression (8,580), attacks (8,573), attack (10,544), suffer (6,086), sick (5,737), side effect (6,427), health anxiety (4,558), heart attack (2,689), chest pain (2,384), anxiety disorder (2,285), physical symptom (2,065), symptom of anxiety (1,701), anxiety symptom (1,685), anxiety attack (1,674), mental health (2,056), same symptom (1,074), acid reflux (907), heart palpitation (793), severe anxiety (782), panic disorder (775), health issue (888), other symptom (774), head pressure (726), shortness of breath (741), social anxiety (717), bad anxiety (664), anxiety issue (541), heart problem (540), racing heart (508), tension headache (504), health problem (609), brain tumor (487), mental illness (589) |
Comparing those key items in the Main posts with those in the Replies, we find that 35 were key for both, indicating a continued focus in the discussion threads on anxiety disorder, panic attack and physical symptom, for example, with commonly cited symptoms including chest pain, heart palpitation, head pressure and shortness of breath. A further 20 key items referring to symptoms appeared in the key lists for Main posts only, though these were often alternative expressions of similar experiences; for example, fast heart and heart racing, brain fog and dizziness. Ultimately, Main posts featured a greater range and specification of ways of expressing symptoms that were nevertheless acknowledged and referred to in Replies. When the same key terms describing symptoms appeared in Replies, members would often emphasise that these are widely recognised symptoms of anxiety and/or that they had personal experience of them. Furthermore, these were often a direct response to requests in the Main posts for indications that others in the forum could relate to what the original poster was experiencing. This is also reflected in the key term same symptom, which was key for Replies only, as contributors emphasised ‘sameness’ in their responses, reassuring the original poster that their experience is comparable with that of others.
Another key term, mental illness, also appeared in our key lists for Replies but not for Main posts, reflecting a range of positions regarding the validity and categorisation of anxiety as a mental illness, particularly in relation to physical illness. For instance, contributors lament the wider social recognition and acceptance of serious mental illness:
Would a person with pneumonia feel guilty, of course not. We have been brought up to believe that physical illness is fine but mental illness. Oh no!
im also very aware that mental illness is just not measurable in the same way as physical, there isnt enough knowledge, science if you will, in regards to mental health. I believe the public are not getting the support they require because the metal health profession just isnt there yet.
Furthermore, in response to the assertion that social anxiety is a very natural feeling and should not be classes [sic] as a mental illness with drugs to treat it, others in the forum were explicit in reiterating the legitimacy of such conditions as distinct from more ‘ordinary’ experiences of, for example, shyness:
Social anxiety disorder is, in fact, a recognised medical condition with diagnostic guidelines and treatment options. Whilst we take your point that perhaps for many in the past, they were merely described as being shy, if you speak to someone with Social Anxiety Disorder they will tell you it is so much more than that.
To some extent, this reflects the medicalising and normalising discourses we reported in Chapter 2, with evidence that in the Anxiety Support forum, there is more of an impetus for acknowledging the validity and seriousness of members’ experiences.
The next largest category, in terms of number of keywords and key terms, captured references to Time, as shown in Table 3.2.
Table 3.2 Keywords and key terms in the category Time
| Main posts | Replies | |
|---|---|---|
| Time | started (7,734), weeks (5,250), anymore (2,486), night (6,638), constantly (2,453), ago (5,417), morning (4,223), day (14,406), sometimes (4,607), months (5,549), today (7,230), stop (4,428), now (19,824), last night (1,676), few days (1,322), few months (876), few weeks (845), last week (800), long time (673), other day (438), couple of days (368), next day (441), next week (409), last year (905), first time (817), last few days (258), couple of weeks (277), day today (243), last couple (265), last time (340), few times (278), past week (245), few hours (275), few years (477), past few days (215), past couple (219), couple of months (211), middle of the night (193) | few days (2,188), last night (1,946), few weeks (1,719), long time (2,059), many times (1,464), few months (1,362), other day (853), many years (1,432), few times (722), next day (909), next week (855), couple of weeks (632), next time (739), couple of days (555), long term (862), sometimes (16,046) |
The use of these terms provides temporal structure to the ‘anxiety narratives’ that appeared in both Main posts and Replies, with 12 key items appearing in both sets of lists as ways of approximating time periods; for example, couple of days, few months, long time. There were 26 key items in this category, however, that appeared as key for Main posts but not Replies and which placed an emphasis on past events, both long-term (couple of months, few years, weeks) and more immediate (last few days, now, past week, today). These comparative time frames enable contributors (of Main posts) to contextualise a more acute and severe experience of anxiety within a more long-standing history, adding legitimacy to their perspective on anxiety disorders – because of their lived experience – as well as their assessments of the developed intensity of the more recent experience that means that they don’t know what to do anymore. This also accounts for why they are now reaching out to the forum. There were four items that appeared in the key lists for Replies but not Main posts, which similarly functioned to establish legitimacy, though more often in relation to advice-giving. For instance, references to many times and many years attest to contributors’ long-standing experiences navigating anxiety and exercising tried-and-tested coping strategies.
Even your monthly cycle can be upset by this condition. I have known it happen many times.
While authors of Main posts oriented towards the past to outline their history of anxiety, the prevalence of next time and long term reflect the future-oriented content of Replies, as contributors offer suggestions as to what the original poster could try on the next occasion of, for instance, a panic attack or a consultation with a health professional, as well as contemplating what are more long term solutions and benefits of, for example, treatment options.
Next time say “hello fear so your back to visit” ……note the word “visit”
Anti-depressants supposed to help in the long term but again can have side effects.
In Table 3.3, we present the composite key items of the next category, which includes references to various persons and things. It is important to note that the prevalence of the terms Im and Ive can, to some degree, be accounted for in the formatting of apostrophes, though the non-apostrophe forms do appear in both the Anxiety Support forum data and our reference corpus. Nevertheless, we checked the relative frequencies for the more conventional I’m and I’ve, finding that while both forms were more frequent in Main posts and Replies compared with the English Web 2020 corpus, they were particularly frequent in Main posts (with Simple maths scores of 15.19 for I’m and 10.31 for I’ve) compared with Replies (8.24 for I’m, 5.55 for I’ve).
Table 3.3 Keywords and key terms relating to Persons and things
| Main posts | Replies | |
|---|---|---|
| Persons and things | Im (9,501), my (114,255), I (341,223), anyone (14,574), myself (9,613), me (52,486), Ive (2,717), else (7,331), everyone (5,980), something (9,716), everything (5,047), anything (5,008), other people (583), only thing (365), same thing (392) | same thing (3,169), other people (1,947), lot of people (1,162), other thing (1,367), only thing (1,173), many people (1,781), anxiety sufferer (613), right thing (732), first thing (807), only way (788), Im (14,760), I (747,246), me (116,701), myself (20,577), yourself (20,049), u (14,244), my (181,030), you (447,708), your (152,732), something (27,752), thing (20,576), things (27,927), Ive (4,576) |
The frequency and range of first-person references (I, Im, Ive, me, my, myself) in both Main posts and Replies reflects the focus on individual perspectives, as authors of Main posts outline the specifics of their personal circumstances, for which they seek guidance, and respondents reciprocate with the sharing of personal experiences in Replies to demonstrate affiliation and legitimise advice-giving. The prominence of second-person forms (you, your, u, yourself) in Replies also demonstrates that respondents adopted what was described in the Main post as the topic of the discussion and sought to orient their perspective around the author of the Main post’s situation, specifically.
The term only thing was key for both Main posts and Replies and showed that members of the forum could empathise with the struggle of having limited treatment options for what was often a personalised experience of anxiety, that is, the only thing that helps me. What was more apparent in Replies, however, were references to where respondents had found motivation – for example, the only thing that keeps me going – offering encouragement to those who sought alternatives to their limited coping strategies. The items that were key only for Main posts also reflect contributors’ openness to unspecified possibilities, looking for anyone and anything else through which to relate their experiences, and in the Replies, members oblige by asserting that a lot of people/many people experience something comparable to what is described in the Main post, thereby contributing to the normalising representation of anxiety discussed in the previous chapter. Similarly, references to the same thing and other people tended to appear in queries in Main posts (Has anyone else had the same thing?), but more likely as affirmative responses in Replies (Yes I have had the same thing). Further encouragement is found in Replies through references to the right thing, supporting the actions of the author of the Main post – including coming to the forum – and reassuring them that they are the best judge of what is the right thing for you. Respondents are definitive in their encouragement, asserting, for example, that the only way to get the better of this thing is not to give up. Nevertheless, references to the other thing show that respondents strive to provide the options that original posters have sought, drawing from their own knowledge and experience (the other thing I would suggest) and prioritising what is likely to be the first thing in a series of actions: First thing I would do is go & see your GP.
The next category, which we have labelled Medical tests, remedies and coping strategies, attests to the involvement of health professionals in members’ experiences of anxiety, as they navigate diagnosis and treatment, including managing the side effects of medications. The key items categorised under this theme are shown in Table 3.4.
Table 3.4 Keywords and key terms relating to Medical tests, remedies and coping strategies
| Main posts | Replies | |
|---|---|---|
| Medical tests, remedies and coping strategies | doctor (4,907), meds (2,518), medications (2,845), doctors (2,764), help (13,692), advice (3,624), blood test (987), blood work (499), deep breath (327), beta blocker (258), anti depressant (233) | deep breathing (2,080), blood test (1,748), deep breath (1,599), self help (1,125), beta blocker (1,043), blood work (751), breathing exercise (742), anti depressant (732), low dose (708), CT scan (589), doing something (661), good advice (562), breathing technique (501), support group (537), doctor (23,563), meds (13,205), medication (12,401), helps (15,131), GP (8,779), help (44,336), doctors (7,976), relax (6,166), CBT (4,859), therapist (4,844), helped (8,931), therapy (7,025) |
There was a greater number of key items in Replies, compared with Main posts, reflecting an orientation towards suggesting coping strategies and remedies. Appearing in both Main posts and Replies were references to doctor(s) and blood tests/blood work, as well as meds, anti depressant, beta blocker and deep breath, providing an indication of some of the more routine aspects of how the physical symptoms associated with anxiety are tested and treated. As we observed in relation to references to persons (anyone else?), the authors of Main posts appear to be looking for indications of what is ‘normal’, with respect to prescribed medications and their side effects, for example, as well as generally seeking advice. In response, we found 17 key items that appeared in the lists for Replies only, which introduced further details relating to coping strategies (breathing exercise, breathing technique, deep breathing, self help, relax) as well as more formal treatment options (CBT, support group, therapist, therapy) including medication (medication, low dose) and testing (GP, CT scan). Members of the forum were open to a range of pharmacological, therapeutic and relaxation solutions, providing personal recommendations of what helps/has helped them and endorsing the good advice of others in the forum. This openness was also reflected in the encouragement to be pro-active and try doing something, even if that is only to distract the addressee from their anxiety.
A key term found in Replies but not Main posts was support group, reflecting the value that members place on being able to talk with others who can relate to the lived experience of anxiety disorders. Contributors referred to the forum itself as a support group, alongside references to in-person groups (is there a support group in your area?) and alternative online spaces that focused on more nuanced aspects of anxiety disorders, such as dealing with grief, or a support group specifically for people under the age of 18. The endorsement of these alternative spaces further indicates that members consider the opportunity to talk with others with comparable lived experience a valuable option for coping with anxiety.
Table 3.5 comprises keywords and key terms in the categories of Feelings & emotion and Cognition, demonstrating how participants expressed evaluation with respect to the many facets of their anxiety experiences and what they feel like.
Table 3.5 Keywords and key terms in the categories Feelings & emotion and Cognition
| Main posts | Replies | |
|---|---|---|
| Feelings & emotion | feel (34,805), feeling (18,565), like (36,189), panic (10,836), scared (7,039), anxious (6,173), felt (7,819), feels (5,168), fear (4,968), worried (3,460), worry (3,105), stress (2,942), worrying (1,910), feelings (2,402), hate (2,348), afraid (2,013), weird feeling (395), horrible feeling (178) | bad feeling (725), horrible feeling (555), feel (79,775), feeling (37,204), panic (20,396), anxious (12,605), fear (16,870), worry (12,160), scared (7,799), stress (10,209), feelings (8,324), calm (6,650), glad (7,105), felt (11,870), feels (6,898), like (76,821), scary (4,706) |
| Cognition | thoughts (4,564), thinking (5,095), know (17,591), negative thought (384), intrusive thought (263) | negative thought (1,438), peace of mind (812), intrusive thought (479), anxious thought (464), positive thought (465), thoughts (13,735), know (58,236), think (48,498), mind (20,576), thinking (12,168) |
In both Main posts and Replies there is explicit acknowledgement of the negative emotional states associated with anxiety, including feeling anxious, fear, panic, scared, stress and worry, as well more generally a horrible feeling. In addition, we observed the related terms afraid, hate, worried, worrying, negative thought, intrusive thought and weird feeling in Main posts, as contributors emphasised the unusualness and concerning aspects of what they have been experiencing. The expression bad feeling – which was more characteristic of Replies – also appeared to acknowledge the negative experiences described in Main posts, but rather than focus on the ‘weirdness’, respondents encouraged acceptance, thereby contributing to the normalising discourse we have observed through other features.
Nothing terrible is going to happen to you if you leave home. So just accept the bad feeling and just go outside any way. But you must accept the bad feeling calmly and with the minimum of fear you can. Say: “To Hell with you agoraphobia, you’re not going to stop me going out.
Furthermore, where authors of Main posts referred to an intrusive thought, there were more often references to anxious thought in Replies, as respondents were more definitive in attributing this part of the experience to an anxiety disorder. The prominence of the terms calm and peace of mind in Replies reflects how it is often positioned in contrast to – and the antidote to – anxiety, as original posters are encouraged to try to stay calm and offered suggestions as to what brings them peace of mind. In the Replies, we also find a contrastive emphasis from the negativity of anxiety disorders on positive thought and particularly the positive emotion of being glad. This appeared as advice, for members to Find a positive thought to counterbalance the negative and respondents themselves highlighted successes in other participants’ experiences (Glad you managed to go out with your friends today). Contributors also expressed comfort in finding others who can relate (glad to know I’m not the only one), again demonstrating the value of being able to find people who can relate to your own experience, which was also explicitly marked in the next category of key items, Forum relations.
The composite items of the category Forum relations are presented in Table 3.6, which shows that a greater number of these keywords and key terms were found in the Replies in the forum.
Table 3.6 Keywords and key terms relating to Forum relations
| Main posts | Replies | |
|---|---|---|
| Forum relations | hi (7,162), hello (2,596), hope everyone (330), can anyone (250), long story (223), thanks in advance (203), first post (187), good morning (190) | good luck (6,997), big hug (868), kind words (851), good idea (1,183), good night (706), thanks for the reply (509), hi (36,671), xx (18,876), xxx (13,761), hope (35,260), thank (25,929), sorry (14,609), lol (10,270), yes (20,717), luck (9,916), thanks (17,756), alone (12,560), maybe (14,584), hello (6,811), too (20,237, sounds (8,906), reply (6,254), wish (8,833) |
The only items that appeared in the key lists for both types of messages were hello and hi, demonstrating that greetings were a regular feature of the forum. We also found the term good morning as a key feature of Main posts and the term hope everyone would also tend to appear at the beginning of a message, as contributors set a positive and affiliative tone (hope everyone is having a good day). The prevalence of lol in Replies also indicates a light-hearted tone, even after disclosure, relating to personal struggle:
Every little pain in arms or chest I immediately think heart attack lol.
The Main post key term first post indicates how common it is in the forum to provide an introductory message and the inclusion of this term offers some justification for the subsequent structure and content, which typically includes a recap of the poster’s personal history with anxiety and an expression of their expectations. In these kinds of introductory Main posts, we also find the following terms: long story, as contributors apologise for the length of their message or look to summarise the key points (long story short); can anyone, which often precedes help, advise or relate, for example; and thanks in advance, as contributors anticipate responses to their request. The prevalence of maybe in Replies shows how members of the forum oblige in this request for advice, offering mitigated suggestions and thereby showing a preference for providing optionality, as opposed to being too prescriptive.
The key terms found towards the end of Main posts contrast with the routine closings indicated for Replies, with examples including big hug, good luck, good night and the conventions for sign-offs, which include xx and xxx:
hope you had a good day love big hug :) xx
In this example, we can also see how hope and big hug are used to show affiliation and the keywords alone, sorry and wish function in similar ways, with contributors reassuring other members you are not alone, expressing that they are sorry to hear you’re having such a bad time and wish you well. With respect to the keyword too, we distinguished its meaning as ‘also’ from other functions (e.g., indicating degree and included in the Maximal category, see Table 3.9) in recognition that this was another way in which members demonstrated shared experience (me too). The use of sounds in it sounds like was a very simple way in which participants could show receipt and acknowledgement of what the original poster had discussed, which could be reformulated through the alternative perspective of the respondent (Sounds like you’re doing brilliantly), or even offer confirmation: Yes this sounds like anxiety. Some of the key items for Replies reflect additional contributions, including the original poster’s response to Replies as demonstrated in the terms thanks for the reply/kind words and contributors generally evaluating suggestions as a good idea.
Relatability was also expressed, in Replies in particular, through the key items we categorised as relating to Personal circumstances, which are presented in Table 3.7. For authors of Main posts, this reflected participants’ reconsideration of what is ‘normal’. For instance, they express a desire to get back to normal, that they are unable to do normal things and wonder if what they are experiencing is normal. This break from the ‘normal’ is also indicated in that anxiety has affected my whole life and one routine behaviour that symbolises this normal life and which has been disrupted is going to or getting out of bed. In response, other members of the forum emphasise normality without being dismissive (It’s totally normal but very unpleasant) and by providing sameness (I feel the same way, I have the same problem, I’ve been in the same boat).
Table 3.7 Keywords and key terms relating to Personal circumstances
| Main posts | Replies | |
|---|---|---|
| Personal circumstances | normal (5,535), bed (2,855), new job (365), whole life (196) | same way (1,622), same problem (713), new job (602), same boat (510), normal (11,584) |
References to a new job in Main posts show that this is often used as an indicator of wellness and recovery (I see my new job as a breakthrough on the road to recovery), which anxiety threatens to jeopardise (I’m scared I will get dizzy and have anxiety and just not have enough energy for my new job), although the disruption or pursuit of a new job can itself be a source of anxiety. In Replies, members express understanding (Starting a new job can be a stressful time for most people) and offer a more celebratory perspective (Congratulations on your new job!), further demonstrating the sympathy and positivity of Replies that we have observed in relation to other categories.
Related to the discussion of physical symptoms, there were a number of keywords and key terms that related to Body parts and processes, which are presented in Table 3.8.
Table 3.8 Keywords and key terms in the category Body parts and processes
| Main posts | Replies | |
|---|---|---|
| Body parts & processes | chest (6,263), sleep (7,423), heart (11,540), stomach (3,198), head (8,063), breath (2,534), throat (2,201), breathing (2,253), woke (2,000), wake (2,301), brain (3,095), blood (3,909), heart rate (1,098), blood pressure (966), left side (592), left arm (380), right side (430), heart beat (342), whole body (269) | blood pressure (2,531), heart rate (2,196), nervous system (1,974), heart beat (639), left side (589), sleep (15,799), breathing (9,742), heart (21,399), chest (7,807), stomach (5,223), brain (7,888) |
The key items that appeared for both Main posts and Replies show that anxiety is consistently reported as manifesting in the brain, stomach, chest, heart and blood pressure, disrupting the ordinary bodily processes of sleep and breathing. There are a greater number of key items relating to this category in Main posts, though they refer to approximately the same body parts or systems (blood, breath, head, left arm, right side, throat), with the keywords wake and woke reiterating the disruptive effect of anxiety on sleep. Key for Replies were references to the nervous system, which present as a kind of elaboration on how anxiety disorders manifest in physical symptoms throughout the body, providing an underlying structure through which digestive (stomach, throat), cardiovascular (heart, blood), pulmonary (breathing), cerebral (brain) and somnial (sleep) processes can be affected. This is also the basis on which certain treatments – such as vitamins – are recommended in that if they help the nervous system, they will in turn help with symptoms affecting different parts of the body.
In Chapter 2 we discussed our observations of Catastrophising and Minimising discourses in relation to anxiety and these are, to some extent, reflected in the key item categories Negative evaluation, Positive evaluation, Minimal and Maximal, which are presented in Table 3.9.
Table 3.9 Keywords and key terms in the categories Negative evaluation, Positive evaluation, Minimal and Maximal
| Main posts | Replies | |
|---|---|---|
| Negative evaluation | bad (9,881), worse (5,387), weird (3,443), horrible (2,055), crazy (2,348), wrong (3,729), bad day (540), hard time (371) | bad day (1,538), hard time (785), vicious circle (609), bad time (469), worse (14,633), bad (22,261), horrible (6,215), hard (17,158), awful (4,860) |
| Positive evaluation | ok (3,017), good day (905) | good day (2,499), good thing (1,243), ok (11,586), better (35,107), good (45,245) |
| Minimal | just (33,777), little bit (422), little thing (244) | little bit (1,092), small step (464), just (87,446) |
| Maximal | really (16,985), so (40,750) | really (39,843), too (13,491), so (113,684) |
Comparing the keywords and key terms that appear in Main posts and Replies, we find matching expressions of negativity (bad, bad day, horrible, worse, bad time, awful) and difficulty (hard time, hard), as well as the term vicious circle in Replies, demonstrating an understanding of the accumulative effects of feeling anxious about having anxiety, for example. However, our key lists indicate that authors of Main posts more explicitly express the abnormality of their experience, as shown in the keywords weird, crazy and wrong, which are not represented in the key items for Replies. Indeed, in our discussion of the Personal circumstances category, we showed how respondents more readily describe experiences of anxiety as normal. As with other categories, we also find more evidence of positivity in Replies, compared with Main posts, with more explicit acknowledgement of what is considered a good thing and of when other contributors report feeling better. What we find in both Main posts and Replies are hopeful expressions on behalf of other members that they are ok, feel better and have a good day.
Positivity was subject to minimisation in Main posts, as contributors refer to a little bit of positivity or inspiration, or feeling a little bit more relaxed, better or stronger, for example. This contrasts with the maximal (really, so) negative effects that even the slightest prompt can have:
Every little thing pushes me off edge.
In this way, the problem of anxiety is maximised, in terms of both severity and how many potential triggers there are, since while the triggers themselves are minimal, they are ubiquitous. In the Replies, there is a sense that members have established provisional thresholds for the extent or amount of negativity they can tolerate in that there are frequent references to too, through which they refer to worrying too much, to dosages being too much and to general behaviours which can become problematic when practised too much. Subsequently, they encourage others to adopt a critical view and to intervene when something becomes ‘too much’. This also contributed to the solution-oriented focus of Replies, in contrast with a problem-oriented representation in Main posts. Minimisation is also evident in how respondents present recommendations for such solutions as a small step towards living well with anxiety, which can serve to manage expectations about the potential for recovery. We can also consider this a kind of face-saving, as respondents avoid overstating the utility of their advice and we find the obverse in Main posts, as contributors minimise the imposition of what they are asking of other members and present their desire as reasonable. For example:
Just wondering how everyone copes the best with anxiety
I just want to feel normal again
The final categories of key items capture Auxiliary & modal verbs and Lexical verbs, which are shown in Table 3.10. Generally, terms in these categories are highly multifunctional and cannot readily be summarised, given the range of uses that they have. Nevertheless, their keyness attests to their use at a higher-than-usual rate and we can see how they contribute to some of the topical patterns discussed earlier. The prevalence of having in Main posts, for example, captures references to posters presently having or historically been having panic attacks, breathing problems, palpitations and so on as they outline their personal experience of anxiety.
Table 3.10 Keywords and key terms in the categories Auxiliary & modal verbs and Lexical verbs
| Main posts | Replies | |
|---|---|---|
| Auxiliary & modal verbs | am (24,009), can’t (10,361), dont (3,531), cant (2,827), having (9,367) | I dont (1,514), I cant (732), am (48,004), dont (10,054), do (118,867), doing (16,126) |
| Lexical verbs | going (15,987), getting (7,619), got (9,571), go (13,953), get (22,649), went (6,979), trying (4,649) | try (36,614), get (81,586), going (40,239), go (43,515), take (46,033), keep (21,162), taking (14,939) |
In the key items (I) dont and (I) cant we observe the same issue of formatting apostrophes as we discussed in relation to Im and Ive, though the standard form of can’t also appeared in our keyword lists for Main posts. Looking at the use of these terms in context, we once again find a focus on problems in Main posts, as contributors express the ways in which they are inhibited: I dont know, I dont sleep, I cant do it anymore. In the Replies, we also find personal disclosures of limitations as a result of anxiety alongside expressions of sympathy (I cant begin to imagine what you’re going through), as well as ‘solutions’ in terms of the adjustments that members have made to minimise the effects of anxiety: I dont drink caffeine anymore and if I go to the sauna then I dont get panic attacks anymore. This focus on solutions is also evident in the forms try/trying. In Main posts, the present continuous and perfect progressive use of I am trying/I have been trying aligns with the poster’s narrative descriptions of the struggles that have led them to their present moment, reaching out to the forum. In Replies, the use of the present habitual I try suggests a course of action that the poster returns to, as a proven coping strategy, thereby implying advice:
I try and breathe and distract myself as best I can
The imperative is also more often found in Replies, such as Try and stay positive n take deep breaths, as well as in instances of keep – that is, calm, in touch, thinking positive – and take (care, your time, a deep breath).
A keyness analysis that highlighted prevalent words and terms in Main posts and Replies, respectively, has enabled us to observe prominent themes in the data and the categories we have discussed broadly represent the foremost topics of the Anxiety Support forum. Members of the forum showed an impetus, particularly in Replies, to emphasise the legitimacy and normalcy of anxiety, offering reassurance and advice to those who have come to the forum detailing their struggles with anxiety. We have shown that members of the forum seek and offer affiliation and understanding, and that while those creating Main posts appear to be preoccupied with the negative and problematic aspects of their lived experience, they are met with positivity and guidance from other members. The function of querying key terms was particularly useful for identifying concepts such as panic attack and expressions such as good idea, same problem and right thing, which would have been more difficult to discern based on individual words. It was this observation that encouraged us to look at longer phrases in the data and to consider regular articulations of topics and communicative functions in the forum, which we did by examining key n-grams.
Key N-Grams
Studies of n-grams (alternatively labelled ‘lexical bundles’) have shown that they usually do not represent a complete structural unit and often bridge two phrases or clauses (Reference BiberBiber, 2009). Furthermore, there is typically overlap between shorter n-grams (3-grams, 4-grams), which combine to form longer strings. As such, in Main posts we find the key 4-grams I was going to (681) and felt like I was (459), which were occasionally combined to form felt like I was going to (136). Subsequently, while our investigation also considered key 3-grams, 4-grams and 5-grams, we have focused our attention on key 6-grams in this discussion, since they generally incorporated what we observed in shorter strings. In comparing key n-grams from Main posts and Replies, we established a relative frequency threshold of 15 words-per-million (wpm), on the basis that it would generate a manageable number of examples to discuss. In the Main posts data, this was approximate to a raw frequency of 88 and the resultant key n-grams are provided in Table 3.11.
Table 3.11 Key 6-grams in Main posts
| Rank | 6-gram | Occurrences in Main posts |
|---|---|---|
| 1. | I don’t know what to do | 593 |
| 2. | in the middle of the night | 178 |
| 3. | felt like I was going to | 136 |
| 4. | as long as I can remember | 120 |
| 5. | going to have a heart attack | 118 |
| 6. | I was having a heart attack | 113 |
| 7. | for as long as I can | 102 |
| 8. | up in the middle of the | 101 |
| 9. | all I want to do is | 96 |
| 10. | I was going to pass out | 95 |
| 11. | I thought I was going to | 95 |
| 12. | I felt like I was going | 89 |
We can see from Table 3.11 that some 6-grams overlapped (e.g., felt like I was going to/going to have a heart attack) and n-grams were also combined within posts. In the abridged version of the following Main post, we see that the participant summarises their concern in the heading of the post (I don’t know what to do) before describing their experience (I woke up in the middle of the light [sic]) in terms of physical symptoms and how they tried to make sense of it (I thought I was going to pass out). In the rest of their post, they report their experience of seeking professional help and treatment, leading up to this message to reach out to the forum:
Title: I don’t know what to do anymore. Anxiety, Stress, Depression. Very Lightheaded 24/7
I suffer from extreme anxiety, stress & depression. When I was 16 I woke up in the middle of the light brain fogged, dizzy, lightheaded, very anxious and panicky and had alot of outter body experiences.
[…]
I started panicking every single night even having to call my wife home cause I thought I was going to pass out on a nightly basis. this went on for months and I tried controlling them with NO LUCK at all.
This example reiterates how key items in the forum data highlight what participants seek through their messages, the ways of reporting experiences of anxiety and their motivations for reaching out at this moment. This corresponds with Reference BatesBates’ (2021) observations of advice-seeking in online forums in which participants express having lower epistemic status (i.e., in need of knowledge or ideas from other members), perform relational work by introducing themselves and providing autobiographical information, which in turn provides justification, often by way of a present crisis, for requests for participants’ input. Our observations of key n-grams indicate how these aspects are formulated in the Anxiety Support forum.
The n-gram I don’t know what to do expresses a need for help that indirectly encourages advice-giving from other members of the forum and was often followed by an explicit request (Please help me). This shows that it is common for participants to come to the forum when they are uncertain: when they are lacking ideas or the information to navigate their experiences with anxiety. Furthermore, 28.13% of instances of the n-gram came from participants aged 20 to 29, suggesting a ‘junior’ status in relation to anxiety. In 119 instances, the n-gram was followed by anymore/any more, suggesting that the poster had previously been working their way through their experience but had just recently encountered a problem that tested the limits of their own understanding or capacity to find a solution. Indeed, other key n-grams establish that participants have had some kind of experience of anxiety for as long as I can/as long as I can remember and there are often details of long-term problems and efforts to combat issues of anxiety, attesting to the poster’s efforts in trying to deal with the problem themselves.
The severity of such experiences is reflected in the various n-grams that describe the physical sensation and impact: I felt like I was going to/going to have a heart attack/I was having a heart attack/I was going to pass out. Contributors would also describe feeling such as they were going crazy, mad or to lose my mind. As such, they describe experiences that corroborate posters’ accounts of a ‘crisis’, for which they are seeking support or explanation. These anxiety encounters could mark the beginning of a narrative that subsequently involved medical tests, diagnosis, treatment and so on or a problematic turn in an otherwise consistent series of events, of which the contributor has better comprehension.
The key n-grams up in the middle of the/in the middle of the night echo our observations of anxiety reported in terms of sleep disturbances and these depictions minimise the agency of the person experiencing them, since they are at rest. In other words, the individual is not engaging in any particular anxiety-inducing activity, rather anxiety can be activated even when the individual is asleep. Subsequently, authors of Main posts spoke of their desperation, expressed in the key n-gram all I want to do is, which prefaced desires such as go back to work and normalise my life a bit, or get on and do my job.
Our investigation of n-grams in Replies similarly focused on 6-grams, with the same relative frequency threshold of 15 wpm. We observed similar overlap for key n-grams (including 3-, 4- and 5-grams), in that the key 6-grams shown in Table 3.12 can be summarised according to a handful of phrases. Furthermore, Replies tended to be shorter than Main posts and the n-grams we observed were more likely to offer ‘complete’ expressions that could be taken at face value. In other words, they expressed unambiguous sentiments, the meaning of which was not subject to change based on the context of the forum post in which it appeared.
Table 3.12 Key 6-grams in Replies
| Rank | 6-gram | Occurrences in Reply posts |
|---|---|---|
| 1. | thank you so much for your | 679 |
| 2. | i hope you feel better soon | 503 |
| 3. | us know how you get on | 500 |
| 4. | let us know how you get | 498 |
| 5. | i know exactly how you feel | 411 |
| 6. | i wish you all the best | 318 |
| 7. | at the end of the tunnel | 281 |
| 8. | light at the end of the | 276 |
| 9. | us no how you go on | 260 |
| 10. | let us no how you go | 258 |
| 11. | Hope and Help For Your Nerves | 258 |
| 12. | in the middle of the night | 256 |
| 13. | at the end of the day | 256 |
| 14. | let me know how you get | 237 |
| 15. | me know how you get on | 235 |
| 16. | you so much for your reply | 230 |
| 17. | hope you have a good day | 215 |
| 18. | you are not on your own | 205 |
| 19. | thank you for your kind words | 205 |
| 20. | i was having a heart attack | 200 |
| 21. | and let us know how you | 200 |
Based on key 6-grams, Replies can be characterised in terms of showing affiliation and reciprocity, suggesting that members of the Anxiety Support forum respond in kind when authors of Main posts share personal anxiety narratives and seek support. In the first instance, the key n-grams i know exactly how you feel and you are not on your own explicitly denote shared experience, which we have seen is often directly requested in Main posts. In 54.19% of cases, this was written by someone identifying as Female (in 12.56% of cases, Male).
The key 6-grams light at the end of the/at the end of the tunnel served a similar function in that authors of Main posts often implored other members for signs of hope – there were 87 instances of light at the end of the tunnel in Main posts. Whether prompted by the matching phrases or otherwise, respondents are unequivocal in affirming the light at the end of the tunnel, alongside empathetic disclosures that they too had struggled to be optimistic about coping with anxiety. In addition, we see reciprocal personal narratives as respondents describe their own ‘crisis’ moment using the key n-gram (I thought/felt like) I was having a heart attack. As with Main posts, these could equally mark the onset of their experiences with anxiety, or a more recent episode that indicates that they are still facing those types of challenges.
Similarly, respondents would refer to waking up in the middle of the night to show their shared experience of sleep disturbances, however, in Replies, this more often preceded suggestions for how to deal with that aspect of anxiety disorders.
when you wake up in the middle of the night feeling anxious just drink a cup of green tea
This happens to me in the middle of the night so I’ve been keeping a heat pad in my bed.
There are further indications of advice-giving in Replies, in the key n-gram at the end of the day, for example, marking the articulation of a considered point of view that spoke of the poster’s wisdom.
At the end of the day worrying about it isn’t going to make it go away
At the end of the day, it should be your call about what medication you are on
As a summative statement, it often appeared after a more detailed description of personal experience, coping strategies and treatment options. Again, we could say that this type of response is requested in Main posts by people who otherwise don’t know what to do. Another recommendation is indicated in the key 6-gram Hope and Help For Your Nerves, which refers to the title of a popular self-help book by Dr Claire Weekes and its alternative title, Self Help For Your Nerves, also appeared 196 times in Replies. This specific book was recommended by 72 different users (although one specific user was responsible for 117 (44.49%) mentions), among other resources and coping strategies.
There is clear evidence that members of the forum look to facilitate dialogue, beyond a simple problem-solution exchange. The key n-grams thank you so much for your/you so much for your reply and thank you for your kind words predominantly indicate that authors of Main posts return to acknowledge other participants’ replies, which are described as advice, help, lovely comments and support and so on. Less often, this would be a comment on the original post itself, indicating that advice is more commonly found in Replies than Main posts. Seven of the key 6-grams represent some part or version of the phrase let us know how you get on, a clear indication that respondents in the forum are receptive to the personal stories of other members and encouraging of further contributions. The anticipation of further developments reflects the narrative structure observed in Main posts, by which posters position themselves at some kind of crisis moment that has prompted them to seek advice. Subsequently, respondents express a vested interest in knowing that the poster was able to navigate this challenge and/or that their advice was successful:
Let us know how you get on with the further tests.
If you do try it let us know how you get on with it.
Even when a respondent might not relate to the specific experiences described in a Main post, there were other expressions of empathy, notably I hope you feel better soon, I wish you all the best and hope you have a good day. Like the phrase let us know how you get on, these typically appeared as a kind of sign-off to a Reply, punctuating the message with a strong affiliative statement.
Emoticons and Emoji
The final part of our analysis in this chapter concerns the use of emoticons and emoji. We found 59,757 occurrences of precisely 200 different characters in the Anxiety Support forum corpus, appearing in 14.25% of posts and used by 24.51% of contributors. Compared with the English Web 2020 corpus, emoticons and emoji were underused in the Anxiety Support forum and none of the individual characters reached our threshold for keyness. The relative frequency of emoticons and emoji (2.53 per 1,000 words) was also lower than that observed by Reference FrehnerFrehner (2008) in SMS texts (4.88 per 1,000 words) and emails (3.18 per 1,000 words). While their general use was not notably frequent in the Anxiety Support forum, we observed that emoticons and emoji were a key feature of Replies (3,116 occurrences per million words) compared with Main posts (1,481 occurrences per million words). Based on the observations we have made so far of keywords and key n-grams and taking the use of emoticons at its broadest function (Reference Dresner and HerringDresner and Herring, 2010), this finding would attest to the affiliative nature of Replies.
The most frequently used characters for Main posts and Replies, respectively, are shown in Table 3.13. The prevalence of smiley emoticons – having a resemblance to facial expressions – would suggest that users are largely using these characters to provide some indication of affective states, strengthening the valence of the message as it is communicated in text (Reference Jaeger, Roigard, Jin, Vidal and AresJaeger et al., 2019), for example, or softening the impact of potentially negative messages (Reference Skovholt, Grønning and KankaanrantaSkovholt et al., 2014).
Table 3.13 Frequent emoticons and emoji in Main posts and Replies
| Main posts | Replies | ||||||
|---|---|---|---|---|---|---|---|
| Character | Occurrences | Posts | Users | Character | Occurrences | Posts | Users |
| :( | 2,648 | 2,053 | 855 | :) | 21,134 | 17,462 | 2,088 |
| :) | 2,013 | 1,463 | 542 | :-) | 9,625 | 6,525 | 399 |
| :-) | 612 | 418 | 101 | :( | 4,125 | 3,752 | 1,043 |
| :-( | 483 | 390 | 137 | :D | 2,237 | 1,756 | 127 |
| :/ | 415 | 366 | 165 | ❤ | 1,421 | 1,103 | 316 |
| ❤ | 157 | 135 | 78 | ;) | 1,315 | 1,158 | 211 |
| :D | 109 | 83 | 28 |  | 1,145 | 1,046 | 250 |
 | 100 | 79 | 47 | :-D | 978 | 777 | 34 |
| ;) | 91 | 84 | 41 | :/ | 875 | 827 | 307 |
| :’( | 87 | 74 | 49 | ;-) | 763 | 693 | 90 |
 | 82 | 72 | 40 | :-( | 1,313 | 1,106 | 248 |
Our keyness analysis pointed to more prevalent use of negative emotion characters in Main posts and more positive characters in Replies. The emoticons :(, :-(, :/ and :’ (were key for Main posts and the following were key for Replies: :), :-), :D, ;), :-D, ❤, 😊 and ;-). These results are consistent with our observations of the ‘problem’ nature of Main posts and more positive tone of Replies in the forum.
In 390 (19.00%) Main posts that featured the emoticon :(, this punctuated the Title for the message, establishing a negative tone for the post. For example:
Worst iv been for ages :(
Why is this happening to me :(
Great … :(
These instances demonstrate that contributors have been prompted to post on the basis of a recent (re)emergence of a particularly bad experience associated with anxiety, looking for explanations and advice, as well as the use of emoticons to indicate, in this case, the ironic use of ‘great’. If we consider the full content of Main posts, there are frequent references to help (325 instances), please (185) and advice (64) alongside this emoticon, as well as explicit references to corresponding emotional states such as scared (162), panic (89) and hate (54). In comparison, in Replies :( more often appears alongside expressions of sympathy; for example, sorry to hear you’re feeling crappy too :(. We observed similar patterns for the comparable characters :-(, :/, :’ (and ☹.
We found that preferences for form of comparable emoticons/emoji were associated with different types of contributor. For example, in 32.10% of messages featuring :-(, the contributor reported as being from the UK and in 5.61% they reported being from the USA, compared with instances of :( where 25.40% of messages came from the UK and 15.99% came from members based in the USA. Furthermore, users aged 20 to 34 contributed 39.28% of messages featuring :( but only 15.64% of messages featuring :-(. The :-( form was potentially favoured by older participants, with members aged 40 to 49 contributing 14.91% of messages with this emoticon but only 5.17% of posts featuring :(. The proportion of messages contributed by female and male posters was comparable between :( and :-(.
The more explicitly positive characters :), :-), :D, ;), :-D, 😊, (:, :) and ;-) were key for Replies and often appeared alongside the kinds of supportive statements we have observed as more strongly associated with Replies:
Good idea! Let me know how it goes :)
Im glad you had a great day!! Lets keep going !! 😊 😊
hope you can see that light at the end of the tunnel ;)
Indeed, here we can see how the characters combine with key items we have discussed already in this chapter, as well as how the repetition of emoji can more emphatically convey the sentiment being expressed. Similarly in Main posts, where the use of such positive smileys was relatively less frequent, it appeared alongside the kinds of affiliative statement we have seen at the start of such messages (Hope you’re all okay 😊) and the pre-emptive thanks that tend to appear at the end (Thank you for your advice :)).
Certain characters appeared to be more adaptable to a range of functions. For example, ❤ was used to express positivity, empathy, as an abstract representation of love or an iconic representation of the human heart:
❤❤❤❤❤❤❤❤❤❤❤❤❤❤ his post makes me happy!! It brightened my day!! Thank you!!!!!
I get you. I understand. You are not alone. ❤
I ❤ my boyfriend
It’s the adrenaline! Don’t worry it won’t be your ❤ your heart rate goes up with the anxiety.take care don’t panic xx
Only when the icon stood in for a lexical item (‘love’, ‘heart’) would it appear in the middle of a clause, that is, in the appropriate syntactical position. When performing other functions, it would more typically occur at phrasal or sentence boundaries much like punctuation marks, as has been observed across various CMC contexts (Reference Dresner and HerringDresner and Herring, 2010; Reference Skovholt, Grønning and KankaanrantaSkovholt et al., 2014).
When we looked at the distribution of emoticons and emoji in relation to different kinds of contributor, we found that, collectively, these characters were key for posts written by female members (3,089 wpm) compared with male posters (1,072 wpm) and that of the 4,356 users who posted messages with at least one emoticon or emoji character, 48.71% identified as female (13.66% identified as male). Since emoticons were also more frequent in Replies, the prevalence of female members among posts with emoticons may in part be a result of a higher proportion of female members posting Replies. Of the participants who used emoticons and emoji and disclosed their age, 35.67% were in their 20s and 39.42% were aged between 30 and 39, with significantly lower frequencies for participants aged 40+. These findings, indicating higher usage among women and by younger people, are consistent with previous findings (Reference Prada, Rodrigues, Garrido, Lopes, Cavalheiro and GasparPrada et al., 2018). This raises questions about the different roles that participants carry out in the forum and how they manifest in different communicative strategies. For instance, we have reported findings that suggest that younger participants express not knowing what to do, so are more senior members of the forum offering advice? If they are not in a position to impart wisdom, are less senior members opting to express affiliation and empathy (with or without emoticons)? In the chapters that follow, we focus more on the interactive dynamics of discussion threads, including the range of response types documented in the forum and the contributions of different types of participants.
Conclusion
We set out in this chapter to conduct an exploratory analysis of the Anxiety Support forum, to identify prominent themes and to consider the manner of expression used in this online space, as a form of CMC. The themes we identified based on keywords and key terms pointed to a focus on symptoms and personal anxiety narratives, indicating that in Main posts members set out to introduce themselves and give other members sufficient information about their circumstances in order to offer reassurance and advice. The key items for Replies also reflected a focus on symptoms and personal stories, but more significantly demonstrated some of the coping strategies recommended by other members of the forum, alongside expressions of affiliation and sympathy. The use of emoticons reflected a relatively informal tone, particularly in Replies, and a heightened sense of affective expression.
Consistent with the findings of Reference Stommel, Lamerichs, Hamilton and ChouStommel and Lamerichs (2014), the agenda-setting (Main) posts in this online forum were characterised by ‘problems’, which was evident not only in keywords and terms but also key n-grams and emoticons associated with negative emotions. Replies were notably more positive, expressing optimism and empathy and encouraging further dialogue (let us know how you get on). While the Main posts sometimes evidenced features linked to catastrophisation, the Replies tended more towards a normalisation representation of anxiety. That these patterns were shown to be consistent across the forum and evidenced in different units of analysis attests to the shared understanding among participants of why people come to the forum. We saw a pattern of behaviours by which new members – or members experiencing ‘crisis’ moments – outlined their situation to the forum and respondents were accommodating in their focus on that individual’s case, drawing on their own individual stories as a show of solidarity and as the basis for advice-giving. There was an imperative among authors of Main posts to show that they had made attempts to navigate their own challenges, but that the current crisis has pushed them over the limit of what they had the capacity to deal with on their own. As such, they positioned other members of the forum as more knowledgeable and potentially able to offer the wisdom of their experience.
In addition to highlighting the different sentiments and communicative functions of Main posts versus Replies, our key n-gram analysis has indicated that there is a general structure to forum posts, with introductory Main posts often beginning with well wishes for other members (hope everyone is having a good day) and closing with an invitation (Has anyone else had the same thing?) and/or pre-emptive thanks (thanks in advance); and in Replies, expressions of sympathy are foregrounded (sorry to hear you’re having such a bad time) and respondents often close on an optimistic tone (I hope you feel better soon), inviting further dialogue (let us know how you get on). In the next chapter, we turn our attention to these structural aspects, as well as the sequencing of posts as discussions unfold in the forum.
