Introduction

In this chapter we consider the effects of change over time on language use in the Anxiety Support forum. Bearing in mind that the corpus consists of almost nine complete years’ worth of data, we wanted to consider whether particular events in the outside world impacted on what posters discussed at various points. More generally, we wanted to gain a better understanding of how discussion relating to anxiety developed over time. Corpus studies of diachronic change are reasonably plentiful, including for example, research on language change using the Brown family, which consist of written reference corpora sampled from 1931, 1961, 1991/2 and 2006 (Reference Leech, Hundt, Mair and SmithLeech et al., 2009; Reference BakerBaker, 2017). Such studies have focused on orthographic, grammatical, semantic, lexical and pragmatic change over time, noting how written English has become more densified (more information packed into less space), informal (mirroring a colloquial style) and democratic (avoiding language that is seen as face-threatening or maintaining inequalities). Examples of these three trends would be, respectively, the decline of noun phrases containing of in favour of shorter genitive constructions (e.g., the hand of the king vs the king’s hand), increased use of first- and second-person pronouns such as you and I and avoidance of titles such as Mr, Mrs and Miss.

Other forms of diachronic corpus research have been focused more on specific registers of language. For example, Rey (Reference Rey, Conrad and Biber2001) built corpora of dialogue from scripts of the long-running television series Star Trek from 1966 to 1999, examining frequencies of lexical items used by male and female speakers. She found that over time women used less ‘involved’ speech, whereas the opposite was true for men. Reference Brookes and BakerBrookes and Baker (2021) examined a 10-year corpus of British news articles about obesity, finding that over time there was an increasing focus on personal responsibility and scientific research when discussing obesity, as opposed to articles about the role of politics, business and social inequality.

However, there are other pertinent ways of thinking about the passage of time in the forum which are linked to the identities of posters themselves. The first would be to consider that posts in the forum have been tagged for age (for posters who provided such information about themselves when they created a profile), in the same way that they were tagged for sex and location. This allows us to compare language use by different age groups to see whether there are age-graded differences. Research of this kind has been carried out on reference corpora such as the British National Corpus; for example, Reference McEneryMcEnery (2005) showed how adolescents and young adults used a higher proportion of swear words than other age groups. A corpus study of patient feedback by Reference Baker, Brookes and EvansBaker et al. (2019) found that younger people tended to complain more about staff rudeness and not being taken seriously, whereas older patients used negativisers such as never and not to construct themselves as ‘expert patients’; for example, ‘At 50 years old I have never experienced anything like it!’. In the Anxiety Support forum corpus, we can consider how posters in different age groups orient to anxiety, to other members and what they describe as causing or alleviating their anxiety.

A third way of conceptualising time is to consider individual posters and their participation in the forum over the collected data period. Some participants only posted one or two messages, whereas others posted hundreds. Using this kind of information, we consider the ‘journey’ that participants go on, from their first post to their last, and what such a journey could typically look like. For example, do posters assume different roles as they become a more established member of the forum? What kinds of language distinguish newer members from older ones? And are there distinctive forms of language which provide clues that someone is never intending to post again? This is an area where there is less existing work to draw on and we hope that the exploratory analysis in the penultimate section of this chapter inspires others to take these methods further. We begin, however, by considering the first of the preceding ways of considering time – by looking at changes across the whole forum over the nine-year period.

Comparing Years

The first message in the corpus, entitled ‘Agoraphobia’, was posted at 16.54 on 20 March 2012. We can not be sure that this is the first ever message posted to the Anxiety Support forum, since some of the replies included in the data correspond with earlier discussion threads. The last message in the corpus was posted approximately eight years and eight months later at 2.08am on 6 October 2020 as a response to another contributor who had posted about gagging while excited: ‘I’m so glad I’m not the only one to have this as well. Thanks for sharing this.’ The forum continues to operate and we are aware that this is not the last message posted to the forum, nor to many of the discussion threads included in our corpus. Figure 7.1 shows the number of posts per month over time. There was considerable fluctuation with an initial peak in April 2013, then a fall until 2014–15. This was followed by another rise in posts and a high peak in September 2017, then a reasonably steep fall-off, a slight recovery in May 2019 and then another decline.

Figure 7.1 Frequency of posts over time (months)

A pertinent question is whether external events in the world influenced people’s experience of anxiety, particularly events that took place in the UK or the USA, which is where 72.60% of the posters claimed to be from (also indicated in Figure 7.1). On 23 June 2016, the UK and Gibraltar voted on whether to remain in or leave the European Union, with the result, unexpectedly, being 51.89% to leave. Later that year on 8 November, the 58th presidential election in the USA took place, resulting in Republican Donald Trump defeating the Democratic candidate Hillary Clinton, despite losing the popular vote. Both events were seen as controversial and caused widespread consternation. However, they do not seem to have been the subject of much discussion on the Anxiety Support forum. Brexit (which refers to the British vote to leave the EU) only occurred three times across the whole forum, whereas Trump was mentioned 20 times.

In order to identify possible events that we might not have considered, we decided to compile a list of annual keywords. To achieve this, we compared each year against a sub-corpus consisting of all the other years; for example, we compared 2012 against 2013 to 2020 and repeated this for every year in succession. Unlike the analysis in Chapters 3, 5 and 6, we have not considered key multi-word terms in this chapter as we are splitting the corpus into a higher number of sub-corpora, meaning that frequencies are lower. The frequencies of multi-word terms are generally lower than the frequencies of words, which means that the former are less likely to reveal patterns that we can confidently generalise in this chapter.

Table 7.1 shows the top 50 keywords for each of the nine years of the corpus, with frequencies in brackets (usernames have been removed).

For this part of the analysis, we have not removed keywords which were due to a small number of prolific posters using a particular word often, as we wanted to gain an overall impression of the topics or forms of language that were encountered at different points across the timespan. In some cases, it appears that a small number of prolific posters influenced discussion in a particular year; for example, one poster made numerous references to Dr Claire Weekes’ self-help books in 2018 (Claire, Weekes) while another consistently referred to tapering off medication (taper, tapering) in 2019. Other keywords are indicative of stylistic features of particular posters. For example, in 2016, the keyword ah was mainly attributable to one poster who posted 1,969 messages between December 2015 and June 2017.

Some keywords do seem to be more linked to developments outside the forum. For example, the words blog and blogs were keywords in 2012, collectively mentioned 330 times. In the early days of the forum, posters were more likely to recommend anxiety-based blogs to one another, particularly those created by Health Unlocked members.

• Hello, thanks for the link to the site! I’m going to have a look too! Sounds like you’ve really been through it :(hope you find readin the blogs on here as useful as I have! I never knew so many people suffered from anxiety until I came on here! X (2012)

However, by 2020 there were only five mentions of blog(s), perhaps due to other forms of support becoming available. For example, in 2014, the keyword headspace refers to an online company who specialise in meditation. Founded in 2010, they initially operated through their website, but in 2012, a Headspace app was launched, becoming particularly popular among Anxiety Support forum members in 2014.

• I tend to do my headspace first thing in the morning set me up for the day ahead. (2014)

The final year, 2020, stands out as referring to a global event: the COVID-19 pandemic, which resulted in lockdowns in many countries around the world as well as high numbers of deaths and hospitalisations. Keywords specific to 2020 such as virus, fever and masks refer to COVID-19, with concordance analyses indicating that this was the only global-wide topic across the corpus which significantly impacted on the discussion in the forum. One reason for this is that in 2020, COVID-19 was a new disease which resulted in a range of symptoms, some of which were similar to the physical symptoms caused by anxiety.

• I think even more people are dealing with shortness of breath because of the virus it being listed as a symptom so we’re all thinking it and getting anxious and getting physical symptoms. It’s so annoying (2020)

Additionally, some people on the forum related how fear of COVID-19 meant that it was more difficult for them to seek medical help for their anxiety.

• It’s just hard to go see my gp as of this corona virus, yeh a scan of the heart would deffo put my mind at rest thanks a lot (2020)

Some people with anxiety on the forum had an intense fear of dying, so the possibility of being exposed to COVID-19 or having symptoms resulting from catching COVID-19 resulted in increased anxiety.

• just dont wanna die from this virus and get it. and i been waking up with the sore throat also late it goes away when i drink something but then it comes back again ugh. (2020)

While Brexit and Trump’s election caused anxiety to large numbers of people, these events did not immediately or directly impact on their lives in the same way as COVID-19, where lockdowns meant that businesses closed, livelihoods were lost, people had to stay indoors for long periods of time, children could not attend school and elderly family members could not be visited.

• I’m driving my family nuts. I don’t want anyone near my kids or myself. We literally stay in the house. I go to work and lock my office door to avoid contact. (2020)

In short, the impact of Brexit and of Trump’s election was not as direct and immediate as COVID-19. Additionally, while COVID-19 was a devastating event for the entire world, Brexit and Trump’s election were more contentious, with many people in the UK and the USA, respectively, voting for them. Writing about political events on a largely anonymous forum runs the risk of encountering disagreement from others, which is likely to exacerbate anxiety. As such, although members of the forum barely mentioned these two topics, we would hypothesise that this is because people avoided writing about them, rather than not being affected by them.

Changing Themes over Time

The analysis in the preceding section focuses on lexical items which were most typical of a particular year. A limitation of the keywords procedure is that it is not suited to identifying more gradual changes over time, particularly those that might involve combinations of words which together constitute a theme. In order to identify these kinds of changes, we considered the most frequent 1,000 words across the whole corpus. Many of these words were grammatical words such as of, the and at, so did not denote a particular theme. However, it was possible to assign 20% of the top 1,000 words into 12 categories by examining concordances of the words and using a bottom-up categorisation scheme to identify themes. These themes are shown in Table 7.2.

Many of the themes identify a topic (e.g., medication or symptoms), although others relate to functions of language (e.g., affiliative or evaluative language). We plotted the relative frequencies of the words in each category for each year, generating a trendline and calculating gradient and R2 scores for each category. The gradient indicates the slope of the chart: a high positive number indicates a strong increase in relative frequency over time whereas a high negative number indicates a strong decrease in relative frequency over time. A number close to zero indicates that the relative frequency has remained reasonably stable. The R2 score is always a number between 0 and 1 and it indicates the extent to which there is strong correlation (e.g., if we plot the points on a chart, do they fall in a straight line?). A score close to 1 indicates strong correlation whereas a score close to 0 indicates quite a bit of variation over time.

Table 7.2 shows that the categories Body, Symptom, Medication, Test and Doctor have all high positive gradients (above 100), indicating that, relatively speaking, they have increased quite markedly over time. These categories are shown in Figure 7.2 for comparison.

Figure 7.2 Relative frequency of categories of terms increasing over time

References to Symptom and Body were the most common and follow a similar trajectory: there is a slight drop from 2012 to 2013, then a rise until 2015 when the relative frequencies of words in these two categories stabilises, falling slightly for the remaining years. A hypothesis for these changes is that it is due to increasing numbers of American posters joining the forum in 2014. Considering the analysis in the previous chapter, we note that it tends to be American posters who are more likely to refer to medications and symptoms, particularly in relation to different body parts, and as Figure 7.1 indicates, there was a sharp increase in contributions from American posters in 2014. At the start of 2014 only 4.85% of posts were from Americans. This had risen to 27.67% by the end of 2014 and steadily increased in the following years, ranging between 43.93% and 57.96% across the months of 2017. However, when we examined non-US posts, we found that generally they also showed increases in the five categories shown in Figure 7.2. These increases appeared between 2012 and 2015, lagging slightly behind the American increases. This suggests that the American posters may have helped to focus the language of the whole forum in a particular direction. This appears to be one of the most significant changes to the forum and reflects the increased dominance over time of posts which discuss the relationship between anxiety and physical symptoms.

• My chest and heart tell me I’m having a heart attack, nearly every day. I know it’s not true, as those symptoms surfaced almost a year ago …. that is just my strongest, most persistent anxiety symptom. (2020)

Related to the topics of Body and Symptom are increases in discussion of the topics Medication, Doctor and Test between 2013 and 2015, although these are less dramatic. The use of terms in these categories is demonstrated in the following examples:

• I do still take klonopin for anxiety and panic attacks for years now with the best results of any of the anxiety type meds we’re familiar with. I’m prescribed 3, 1 mg tabs a day as necessary. (2020)

• The palpitations and the dizziness have made my anxiety worse, my doctor is checking me out but I have a feeling it will end up being anxiety related. (2020)

• I’ve had them for 6 years now. I’ve done every test and I am waiting for results for a cardiac MRI. (2020)

Taken together, the five categories in Figure 7.2 indicate an increased sense of medicalisation of anxiety in the Anxiety Support forum.

Other categories in Table 7.2 have either very small increases over time or decreases. Figure 7.3 shows the relative frequencies for five categories, of which four (Time, Evaluation, Social Actors, Other Coping) appear to be almost unchanging from year to year. The exception in Figure 7.3 is Affiliative, which shows a large increase between 2012 and 2013, then a decrease until 2015, followed by a more stable pattern. This appears to be largely due to the presence of one prolific poster who joined the group in December 2012 and posted a very high number of messages until she left in 2015, changing accounts several times. This poster had very high frequencies of affiliative words such as xxx, hun, lol, hi and thanks, which explains their spike in 2013 to 2014. The figure indicates how a single poster can make a major impact on the language of a forum.

Figure 7.3 Relative frequency of categories of terms decreasing or remaining stable over time

Two categories that are not shown in Figures 7.2 or 7.3 are Religion and Recovery. Both contain much lower relative frequencies than the other 10 categories and show slight increases over time although no clear linear pattern. The increase in the Religion category is again most likely due to the higher numbers of American posters who joined the forum in 2014 – as noted in the previous chapter, American keywords included God, pray, prayers and praying.

There is therefore evidence for an increased medicalisation discourse around anxiety in the forum – references to medications doubled between 2013 and 2019 and references to tests more than tripled between 2012 and 2020. The reasons for this increase in medicalisation terms are most likely correlated with the changing demographic profile of members who brought with them a focus on physical symptoms associated with anxiety.

Comparing Age

Numerous studies have pointed to age as an important factor in the experience of anxiety in terms of the degree of anxiety people experience and what they tend to be anxious about. Reference Powers, Wisocki and WhitbournePowers et al. (1992) note that undergraduates tend to have anxiety about social events and financial situations while Reference Hunt, Wisocki and YankoHunt et al. (2003) found that older adults worry more about world issues and familial concerns. Reference Gould and EdelsteinGould and Edelstein (2010) found that younger adults tend to report significantly higher amounts of trait and state anxiety symptoms than older adults. Similarly, in a comparison of two groups of adults, one consisting of students with a mean age of 20.1 years and the other of older people with a mean age 71.8 years, Reference Mahoney, Segal and CoolidgeMahoney et al. (2015) found that the younger group reported significantly higher levels of anxiety sensitivity (i.e., the tendency to fear body sensations associated with anxious arousal) and experiential avoidance (i.e., excessive negative evaluation of unwanted thoughts and feelings and efforts to control or escape them). On the other hand, the older group reported significantly higher levels of trait mindfulness, that is, being fully cognisant of thoughts and feelings without judgement or avoidance, thereby decreasing their distress levels. Reference Gerolimatos and EdelsteinGerolimatos and Edelstein (2012) have suggested that such a difference is likely to be due to older adults having better emotional regulation skills so they do not devote anxious attention to sensations that might exacerbate anxiety. Reference Mahoney, Segal and CoolidgeMahoney et al. (2015) add that this may be the result of life experience of the older group and higher likelihood of consultations with medical professionals.

Table 7.3 shows the number of posts by age across the forum. Just over 44% of people did not give their age, so were not included in the analysis that follows. Only a small number of people claimed to be under 20 (1.02%) so there is not enough data for us to compare this group against other groups. Similarly, only 0.64% of people claimed to be 80 or over, and this included 1,061 posts from people who said they were 100 or over (including someone who said they were 244 years old). Therefore, we have not included people aged 80+ in the analysis. The analysis instead focuses on people aged between 20 and 79, which represents 53.92% of posts.

We obtained top 100 keyword lists for each age group by comparing each group against sub-corpora consisting of the remaining age groups; for example, we compared posts made of people aged 20 to 29 against those aged 30 to 79. After removing usernames and keywords which were used at least 50% of the time by a single poster, this resulted in the keywords shown in Table 7.4.

The very small number of keywords for people in their 70s is the outcome of one prolific poster who wrote 23.4% of the messages in this group. This poster tended to use language in a highly repetitive way, which meant that the majority of the top 100 keywords for this age group were the result of that single poster. While this poster’s language would have provided numerous interesting keywords to examine (such as their tendency to refer to blips and glitches in people’s nervous systems), we cannot generalise their language use more widely to people in their 70s. As a result, the majority of the analysis that follows will focus on the 20s to 60s groups.

How Do Posters of Different Age Groups Characterise Their Anxiety?

Posters in their 20s use a set of keywords which suggest an all-or-nothing perspective. Some of these keywords are more suggestive of stylistic choice (e.g., densification realised by combining two words into one – everyday, anymore), others are more clearly indicative of distinct worldview: literally, really, constantly, everything, extremely, seriously, completely, super, whole.

• i just remember being in his car and it cold outside and just feeling like a little mouse in a trap. thats literally what i am, a little mouse and the world is the trap and i want out (20s)

• I constantly check my body in a mirror I have no clue how many times a day (20s)

• Anyway I guess my biggest complaint is that I am afraid of EVERYTHING right now, and I mean everything. (20s)

• Right now my stomach is completely destroyed because of anxiety and I can only eat bland food. (20s)

• My whole body hurts and I become totally incapable of having any sort of thought process within 10 minutes of getting up. I feel completely numb. (20s)

A question arises as to whether these kinds of words are typical of how people in their 20s use language generally, or whether they are more distinctive of situations where they talk about anxiety. To address this, we examined two other corpora containing spoken British English. The Spoken BNC2014 contains 11 million words of informal speech recorded in 2014 which is tagged for speaker information, including age. In this corpus, collective relative frequencies of the preceding nine keywords were obtained. The age groups used in this corpus are delineated differently (0–14, 15–24, 25–34 and so on), so we cannot isolate people in their 20s, although the group with the highest frequency of these words is the 15–24 group, with use of the terms decreasing as age categories increase.

In order to consider whether older speakers have other ways of expressing hyperbole, we examined words with the part-of-speech tag JJT, indicating superlative adjectives such as biggest in the BNC2014. It was the 25–34 group that had the highest frequency of these terms, with the 15–24 group using them second most. We also looked at the top 100 words in the BNC2014 which received the tag RR (general adverb) and identified 10 which tended to be used to express all-or-nothing perspectives: always, never, definitely, absolutely, especially, totally, particularly, certainly, clearly and perfectly. Again, it was the 25–34 age group which used them the most, so we can more widely link these kinds of words as being more strongly associated with the language use of people in their 20s.

Another question to consider is whether this is a current phenomenon – is it something distinctive about people in their 20s who are living in the 2010s, or is it something that people in their 20s have always done? We would need to consider a much larger set of corpora to fully do justice to this question, although one corpus we can consult is the Spoken BNC1994, which contains recorded conversations from 1994. Again, the results look similar (see Figure 7.4), with the 15–24 group using the hyperbolic keywords the most, with decreasing use in older people. We should also note that for all age groups, the equivalent figures are around double in 2014 compared with 1994. The results indicate, therefore, that there is evidence both for younger people using more hyperbolic language and for language use becoming more hyperbolic for everyone in the 20-year period between 1994 and 2014.

Figure 7.4 Combined relative frequencies per million words of literally, really, constantly, everything, extremely, seriously, completely, super and whole in spoken corpora

Figure 7.5 The number of contributors who made between 1 and 20 posts

Returning to the Anxiety Support forum data, people in their 20s also use a high number of keywords which indicate negative evaluation or reference to negative concepts: sucks, crazy, weird, freaking, hate, pains, sick, worried, hurts, wrong, attack, die, dying, scared, scares and the sad face emoticon :(.

• And the dreams make no sense at all or they’re about what gives me anxiety then I’ll wake up in the middle of the night scared and thinking I’m gonna die. (20s)

• Ugh I hate It I cant even enjoy going out i HATE anxiety (20s)

• I am just so scared … Uhhhhhhh : (: (: (: (Does anybody know if you can get some form of tranquilizer from the doctors or MH professionals? I seriously need something! (20s)

With respect to pains, Reference BakerBaker et al. (2017, p. 184) found that in a British corpus of patient feedback to the National Health Service, the word pains was used significantly more by men than women. In the Anxiety Support forum data, male posters were also more likely to use the term, with a relative frequency of 335.17 times per million words, compared with 239.7 times per million words by female posters. When we focus specifically on the 20–29 age group however, the rates of use are much more comparable, with 8,342.29 occurrences per million words for female speakers and 8,896.28 times per million words for male speakers. This suggests that we can associate the term more with the age group, rather than aspects of gender. As pain can be a count or non-count noun, the choice of referring to the plural form pains as opposed to pain suggests a desire for emphasis, highlighting the transient intensity in contrast to a more enduring state.

• On boxing day I went out with a few mates to town and all of a sudden I got off the bus walked around for 5 mins and then I felt really bad chest pains that lingered for about 15 mins (20s)

In contrast, older age groups use a more moderated form of language. For example, the next age group up (30s) use the more euphemistic-sounding words issue and issues to refer to a range of phenomena, including anxiety.

• Knowing your past focus on health issues, it sounds like an anxiety issue. (30s)

The 30s age group also offers explanations for anxiety; for example, by referring to it as consisting of a cycle, whereby anxiety is seen as producing physical symptoms which result in increased anxiety.

• The thinking and worrying are reinforcing the cycle. Anxiety can cause so many symptoms – this is definitely one of the many I’ve experienced. (30s)

People in their 30s write more often about controlling their anxiety (or it controlling them), anthropomorphising it through use of the keyword allow.

• 4 months ago I felt like my heart stopped and started again. This made me panic and spin my anxiety out of control. (30s)

• As I see it the drug im taking helps control the anxiety just as a drug would for any illness/condition. (30s)

• Anxiety will suck u dry if you allow it you have got to want control over your life your so young embrace your youth and do n’t let anxiety destroy it. (30s)

The 40s group used the word fight relatively more frequently than other age groups, although an examination of its uses in context suggests a range of positions. Some posters conceptualise dealing with anxiety as a fight.

• Keep the fight, we are all behind you, learning to deal with anxiety is the key, we will always be in recovery, that is part of us. (40s)

Others talk about the difficulty of trying to fight anxiety or argue that irrational fears shouldn’t be fought.

• I feel like iam fighting a losing war with this anxiety I fight evey day to have a good day but it just don’t happen (40s)

• It is a very scary thing when it happens as all the symptoms are very real but it’s our body fighting something … a fear …. but as it’s an irrational fear there’s nothing to fight. (40s)

The range of uses of fight suggests that the concept of fighting anxiety is both important and a source of contestation for this age group, whereas other groups do not have the same tendency to refer to whether anxiety should be fought (or not).

People in their 60s favour the word cope, which suggests a different set of expectations around resolving anxiety.

• I have to go and see my GP in the next day or so, but how do I cope with the fear and anxiety? (60s)

This group, along with people in their 70s, also use problem and problems relatively more frequently than others. While problem is indicative of something negative, its strongest collocates in the English Web 2020 corpus include solve, solving, solved, solution, fix, address, resolve and facing. The word problem therefore has a discourse prosody which indicates something that has the capacity to be solved. This is unlike the word issue (which as noted earlier, was key for people in their 30s), whose collocates have a discourse prosody for importance (important, key, major, serious, critical) and acknowledgement (address, raise, discuss, facing) rather than resolution.

• You are already staring from a good place – you know the problem and you are actively seeking solutions. (60s)

The 60s group is the only one to have age as a keyword, a word which sometimes occurs in autobiographical statements such as ‘I left school many moons ago at the age of 15.’ However, some posters in their 60s use age to refer to expectations that older people should not experience anxiety.

• I’m being told is grow up and act my age now I’m 61 and don’t seem to able to get out of this hole I’m in (60s)

The Journey from First to Last Post

There were 17,770 unique posters who contributed to the corpus. However, the majority of people only contributed a small amount of text. A third of people (5,894) only made a single posting, and 78.9% of all people who posted only made between 1 and 10 posts. This figure goes up to 87.2% for people who made between 1 and 20 posts. The distribution of users according to how many messages they posted (up to 20 posts) is shown in Figure 7.5.

We end the analysis in this chapter by considering the ‘journey’ that people go on, from making an initial post, to becoming a more established member of a community, to then leaving the forum. During this journey, in what ways does their language use change?

To answer this question, we compared the language of posters at different points across their journey. First, we collected the 17,770 first posts that people made, and compared them against the rest of the corpus to identify keywords. We then did this for the 20th, 40th, 60th and final posts. For final posts we only considered posters who had made at least 60 posts, and we did not consider final posts that occurred in 2020 (the final year in the corpus), as it is possible that people would continue to post to the group past the point where our data collection ended. To rule out idiosyncratic or unrepresentative cases, we removed keywords that had a frequency below 10.

The keywords are shown in Table 7.6. It should be borne in mind that the majority of people who post an initial message to the board will not go on to post a 20th message, and as such, the keywords are based on considerably less data as we work down the table. Correspondingly, while first posts amount to 2,212,827 words, the 20th posts consist of just 168,725 words. Nevertheless, we should not devalue the impact that long-term posters have on the group – they may be in the minority but the sheer volume of posts that they make can mean that they disproportionately affect the kinds of language and views that other posters encounter on the forum. Owing to the lower frequencies, in comparison to the previous keywords analyses in this book, we can be less confident about generalising the findings.

First posts tend to contain elements of personal history, with people describing themselves as new or having just joined the forum and giving information about their name, age (old, years) and sex (male, female).

• Hi everyone i am new here my name is x i am male from egypt 28 years and i think i have anxiety since i was 16. (1st post)

These posters are likely to provide details about how long they have experienced anxiety or related issues (recently, year, started, years, since, months, ago, month, currently, past, began). Posters are also likely to report a diagnosis (diagnosed) or medical prescription (prescribed), or visits to emergency services (ER) as well as labelling different kinds of mental health conditions (panic attacks, depression, social anxiety) and characterising their symptoms (dizzy, sick, fear of dying). They describe their anxiety or related symptoms in very negative terms (suffered, suffering, constant, constantly, extreme, extremely, attacks, severe), which in part legitimises their participation in the forum, as well as their search for advice or confirmation (anyone, appreciated, similar).

• Does anyone know how I can get my life back? Be greatly appreciated. (1st post)

At the 20th post, the amount of negative language has reduced somewhat, although posters use the word awful, to either refer to their own state, or to show sympathy towards others.

• I’m currently going through this, going on 5 months. It’s absolutely awful. (20th post)

• Hi that’s awful that your psychiatrist just told you to stop taking them. (20th post)

The keyword positive indicates another way that posters try to provide support, advice or encouragement, as well as suggesting a different stance towards anxiety.

• Keep on talking on this blog, its surprising how much positive you can feel after writing or even answering peoples questions. (20th post)

• Please think positive thoughts.. you can do it (20th post)

These posters are more likely to refer to wider aspects of their lives which go beyond their own symptoms (sister, girlfriend, dog).

• My friends are more sympathetic than my own sister is and it bothers me to know that! (20th post)

They are also more likely to recommend calming methods such as mindfulness, meditation, meditate and techniques to others.

• Mindfulness meditation is the best way to go because it teaches you to look at your thoughts from an outside perspective without judging them. (20th post)

These posters are more likely to view themselves as being on a journey, which involves research or learning how to improve. Other verbs related to this are realize and train.

• I am still learning to improve what I use and add in a new technique occasionally as well. (20th post)

• Learning what works for you is a very personal journey but you sound like you know exactly what I mean (20th post)

• mornings are the hardest I have found, but am beginning to realize it’s just a false thought. (20th post)

• Read up on brain research, and read up on ways you can train your mind to work In your favor. (20th post)

The keywords reassure and assured are used as emphasis by these posters to be supportive of others.

• Hi, I can reassure that you definetly are not alone. (20th post)

• Please don’t worry and be assured it is only built up anxiety that is doing this and it takes a while for you nervous system to calm down only AFTER you stop thinking and worrying. (20th post)

For the 40th posts, the keyword Google involves discussions which evaluate (usually negatively) the effectiveness of using the internet search engine Google to diagnose symptoms.

• We’ve all been where you are. Convinced we have deadly diseases that we really don’t, no matter what the doctor says. That’s why Dr Google is the number one thing NOT to do. The internet is not a doctor and can’t diagnose anything. (40th post)

Other keywords involve specific pieces of advice which are not based on medication or therapy; for example, drinking water but not caffeinated drinks, using heat pads for sore muscles or taking a bath. The keywords listen and voice refer to advice to use music, podcasts or guided meditations, but compared to the 20th posts, there are references to a much wider range of coping mechanisms.

• I listen to mind calming music too which helps me massively to relax. (40th post)

• I find comfort in doctor weekes voice and advice. (40th post)

• Going out is my coping mechanism even if I do not feel like it or nervous, or apprehensive. (40th post)

In contrast with the 1st posts which talk about suffering or 20th posts which conceptualise anxiety as awful, by the 40th post, people write about being frustrated with their anxiety.

• I get anxiety over anything and everything. I’ll be feeling fine and it’s as if my body’s not happy that way do it makes me freak. I’m just frustrated and it makes me mad at myself. (40th post)

Also, while content in 20th posts describes the experience of anxiety as journey, in the 40th posts we start to see references to what the end of the journey might look like: peace or being free of anxiety.

• I have been free of anxiety for the last year now an off medication lexapro since February. (40th post)

• Hope you can find a little peace today. (40th post)

The small set of keywords used by people making their 60th post contain some words that have already been observed at earlier stages, such as mindfulness, exercises, peace, which suggests that posts start to become less distinctive after a certain point.

The keyword slowly often involves quite specific advice regarding breathing.

• Breathe in for a count of 2 pause then breathe out slowly for 4. Do this about 10 times. (60th post)

This set of people are more likely to use the word group, which either refers to advice to join a group or to the Anxiety Support group itself.

• If there is a buddhist meditation group then that is a good start but things like yoga are also related to practicing mindfulness (60th post)

• Talking is the best medication invented and it is free. Sharing our issues here in this group has been so wonderful for me. (60th post)

The word ended does not refer to people’s anxiety resolving as might be expected but instead tends to be used in personal narratives which involve the phrase ended up.

• I worked somewhere where this lady was really awful and nasty to me just because she was older and worked there longer. She wasnt really nice to anyone.. I ended up leaving my job because she was so rude. (60th post)

It was initially expected that the final posts would contain messages where posters said goodbye to others or provided reasons why they were leaving. Such posts did exist but they were in the minority.

• Well, I am off! It is time to bid you all goodbye in this forum and for me to leave. It is a good forum but I find it causing me more stress than easing it. Thanks everyone for your time and I declare freedom over all of you! (final post)

A handful of final posters noted that it had been a long time since their previous post, suggesting that to all intents and purposes they had already left the group. Their final post therefore seems to signal a nostalgic revisit, despite intentions to resume interaction with the group.

• Haven’t posted for months,time goes by so quickly.been difficult to cope with new job/transition (final post)

One final post was from a friend of the original poster who revealed that she had died.

• Thank you for your condolences and kind words about X. She loved this forum gave her the knowledge she was not alone on her struggle with anxiety and depression. Many on here gave her advice and I thank you for that (final post)

However, the majority of final posts contained no explicit indication that the poster was leaving, suggesting that even after at least 60 posts, when people stopped (regularly) posting to the forum, they did not announce their intention to others or formally say goodbye. This is perhaps due to the somewhat fluid and asynchronous nature of forum interactions. Unlike a face-to-face conversation, which has a distinct end, perhaps some posters did not realise that they were writing their final post, or it is possible that they wanted to leave the door open to return at a later point. This goes some way to explaining why three of the final posts contained reassurances to other posters that the poster was always here.

• I had really bad health anxiety, I also suffered with a bit off depression but this was off the medication I was taking, I couldn’t leave my house … I’m out everyday enjoying life. If you need to talk I’m always here :) (final post)

The small number of keywords used by final posters revealed a couple of possible reasons why some posters decide to stop posting. For example, the keyword simply occasionally occurred in posts which expressed annoyance at others.

• Do … NOT … shout. You’ve just made my eyes bleed with all that uppercase nonsense. Apart from anything else, there is quite simply NO need for that type of ill-mannered silly behaviour. None. Period. Thanks. (final post)

• But I think I’m done with this site, simply because I’ve noticed the negativity I feel after it. (final post)

Similarly, the keyword yours was sometimes used by posters to refer to other people’s opinions. In the following examples both posters thank another person for their opinion but also indicate that they are not in full agreement while engaging in mitigation work (e.g., saying they respect the poster’s views or that they may well be right).

• Hi X, thank you for taking the time to explain things much clearer for me, its my problem to be honest, not yours. i dont mean offence at all, im just always interested in opinions and different topics. we all have to have Different ones dont we. im certainly not picking holes in what you believe, i would not do that, i completly respect your views. (final post)

• you have made yourself much clearer for me, thank you, you have provided me with a very clear opinion of yours. i dont think my mind will ever be settled, thus, you may well be right stde, as i do hold anger, (final post)

In a similar way, the keyword experience was used by final posters as a way of validating an opinion that they suspect may not be welcome.

• From my personal experience, your stubborness could be positive, or, self-defeating. (final post)

• Please get help to talk to your boyfriend about his problem and I know it is tough, but you have to let him know that it is make or break time because it is not fair on you or any future kids if he doesn’t want to get help. This is my opinion and experience, but as they say hindsight is a wonderful thing. (final post)

On the other hand, learned tended to be used in last posts in order to indicate that the poster had made progress (and was therefore less likely to need the support of the group).

• I’m back to my old self again, thank God! It was a rough 3 months for me. I’ve been going to therapy since January and have learned to just accept the anxiety and the symptoms. (final post)

• Some of the things I learned from schema therapy have really changed my outlook. (final post)

Earlier, we saw how learning (a verb which indicates the continuous or progressive tense) was a keyword for 20th posts, so it is interesting to see how the past tense form learned is a keyword for the final posts. At least for some long-time forum members, their final post indicates that their journey has ended positively.

The analysis in the final section of this chapter indicates distinctive forms of language use by contributors at various stages of their involvement in the forum. Of course, the majority of posters do not engage with the forum by making 60 or even 20 posts, so the analysis here focuses more on a small set of people who do remain for a longer period and thereby can potentially have more influence on how the forum operates. The typical first poster provides autobiographical information, writes about their symptoms, often using catastrophising language and asks for help. For the 12% of posters who remain with the group long enough to make a 20th post, by this point they have assumed an advice-giving rather than advice-seeking role, providing sympathy, reassurance and advice to others. They talk more about wider aspects of their own lives and characterise their experience with anxiety as a journey. They are likely to suggest meditation or mindfulness techniques to others. For the 7% of posters who make a 40th post, advice is based on engaging in a wider range of lifestyle practices such as listening to music, taking hot baths and drinking the right fluids. Only 1 in 20 people contribute 60 or more posts to the forum. By this point they provide advice in terms of relating their own experiences as narratives or give specific details about different kinds of exercises. As long-term members, they are also more likely to provide positive evaluations of the forum. For those who leave the group, in their final posts a small number express annoyance at other posters or indicate that they have made sufficient progress to not need the group anymore (a case of having learned as opposed to still learning), but generally there are few clues that a poster’s journey is over. The door is usually left open, even if the majority of posters will never return.

Conclusion

This chapter has outlined three ways in which the concept of time can be used to analyse the language in the Anxiety Support forum, each providing a different picture. Change over time revealed that posters tended not to use the forum to express anxiety over worldwide events – with the exception being COVID-19 – although trends in technology (e.g., the declining popularity of blogs) did impact on the ways that forum members sought ways to address their anxiety.

The comparison of the age of forum members revealed that younger posters (those in their 20s) had a clearly distinct way of writing, which tended to involve extremely negative representations of their anxiety while asking for advice. Older members tended to take advice-giving roles, using more affiliative and supportive language, although even here there were more subtle differences; for example, around viewing anxiety as an issue to be addressed (people in their 30s), to be fought (40s) or a problem to be solved (50s and 60s). Different ways of managing anxiety also appeared to be age-related, with younger posters favouring medication and therapy, while middle-aged contributors were more likely to suggest meditation or techniques involving distraction or relaxation.

And finally, the analysis of the number of posts made by forum members showed a progression from a troubled first post, which typically sought advice, to longer-term members taking on supportive roles, characterising their experience as a journey which involves learning. While only a small number of people engaged with the forum long enough to make 60 posts or more, it is notable even among the long-standing members that few of them made formal goodbyes to the group. As the analysis throughout this book has indicated, the Anxiety Support forum has a range of social norms, some of which are clearly different from those which occur in face-to-face contexts. One point that is worth reaffirming is how the longer-term posts referred to the value of the forum itself in helping them with their anxiety. It is with this in mind that we turn now to our concluding chapter.