Show Summary Details

Page of

 PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, COMMUNICATION ( (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

Subscriber: null; date: 22 September 2018

Online Talk About Mental Health

Summary and Keywords

There is a need to focus on research conducted on online talk about mental health in the domains of ethnomethodology, Conversation Analysis (CA), Discursive Psychology (DP), and Membership Categorization Analysis (MCA). We use the notion of “talk” in this article, as opposed to what could be considered a more common term such as “discourse,” to highlight that we approach computer-mediated discourse as inherently interactional. It is recipient designed and unfolds sequentially, responding to messages that have come before and building a context for messages that are constructed next. We will refer to the above domains that all share this view as CA(-related) approaches.

A characterizing feature of interactional approaches to online mental health talk is their focus on in-depth analyses of relatively small amounts of data. With this focus at the center of their attention, they sit in the wider field of Discourse Analysis (DA), or Computer-Mediated Discourse Analysis (CMDA) who use language as their lens to understand human interaction. DA and CMDA research include a much wider set of both micro- and macro-analytic language-focused approaches to capture online discourse. Of all the CA(-related) work on online materials, a disproportionally large number of studies appear to deal with (mental) health talk.

We aim to answer the question what the field of research on online mental health talk has yielded in terms of findings and methodologies. Centrally, CA (-related) studies of online mental health talk have aimed to grasp the actions people accomplish and the identities they invoke when they address their health concerns. Examples of actions in online mental health talk in particular are presenting oneself, describing a problem, or offering advice. Relevant questions for the above approaches that consider language-as-social-action are how these different actions are brought off and how they are received, by closely examining contributions such as e-mail and chat postings and their subsequent responses.

With a focus on talk about mental health, this article will cover studies of online support groups (OSGs, also called online communities), and interaction in online counseling programs, mainly via online chat sessions.

This article is organized as follows. In the historiography, we present an overview of CA(-related) work on online mental health talk. We discuss findings from studies of online support groups (OSGs) first and then move to results from studies on online counseling. The start of our historiography section, however, sets out to briefly highlight how the Internet may offer several particularly attractive features for those with mental health problems or a mental illness.

After the historiography, we discuss what an interactional approach of online mental health talk looks like and focuses on. We offer examples of empirical studies to illustrate how written contributions to a forum, and e-mails or chat posts that are part of online counseling sessions are examined as interaction and which types of findings this results in. We conclude with a review of methodological issues that pertain to the field, address the most important ethical considerations that come into play when examining online mental health talk, and will lastly highlight some areas for future research.

Keywords: mental health, conversation analysis, discursive psychology, membership categorization analysis, computer-mediated discourse analysis, interaction, action, online support groups, OSGs, online counseling, identity, accountability, forumability, advice


This historiography sets out to sketch the main areas of research on online mental health talk and provides a first broad overview of the research themes that have been addressed in the field. We will first discuss which features of the Internet might be particularly attractive for people with (mental) health problems.

Online Environments: Appeals and Concerns

Whereas anonymity and the ability to access the Internet irrespective of limitations posed by time and place are often generally mentioned as potentially attractive features of online environments (cf. Wallace, 1999), they may be particularly appealing for people who suffer from mental health conditions. The stigma and controversies often attached to these conditions are one possible reason for this. As mental conditions may often coincide with some degree of social isolation, accessing OSGs enables participants to obtain multiple perspectives on their illness, something which is often unavailable in their immediate environment for different reasons. Discussing their illness with friends and family may be difficult or simply undesirable (cf. Varga & Paulus, 2014, for how the latter is topicalized in online talk). For people who suffer from relatively rare, ill-defined, or taboo-laden mental conditions (e.g., people who self-harm, who promote anorexia as a life-style choice on “pro-ana” websites, or who suffer from “contested” conditions like chronic fatigue syndrome or CFS), the Internet may be the only environment where they can share their similar, sometimes controversial views. Oftentimes, sharing these views is geared toward normalizing behaviors or experiences tied to their particular condition (Giles & Newbold, 2013).

This article will deal less with fringe communities such as pro-ana forums (but see Giles, 2006) and instead focus more on the practices in “mainstream” OSGs, although the latter might sometimes very well be relatively small “niche” communities (i.e., catering to people with a specific mental health condition). While we will be referring to OSGs as online support groups, we have to keep in mind that this descriptive label by no means grasps all that goes on in these groups, nor that the social actions that take place in OSGs are necessarily best described as “offering support.”

For people who experience mental health problems, participating in OSGs is often viewed somewhat critically, as people with a mental condition are believed to constitute a group that is reluctant to seek professional help (cf. Thornicroft, Rose, & Kassam, 2007). The fact that they join an OSG to talk about their health concerns is positive in the sense that this online environment fulfills a need that would otherwise remain unmet, but it is also considered a reason for concern. On the one hand, fears exist that people with mental health concerns could start to overtly rely on OSGs to discuss their mental health problems, refrain from seeking professional help, be inclined to stop their regular treatment, or experience adverse effects from taking part in an OSG (see Dancy & Dooley, 2007, for a discussion). On the other hand, participating in OSGs has also been considered for its potential to increase demand for offline diagnoses of mental health conditions provided by health professionals (Giles & Newbold, 2013).

Empirical studies that have looked into the abovementioned fears show mixed results, depending on the type of conditions the OSG is for. A longitudinal study on people who joined an online support group for depression showed that people did not stop their regular treatment but reported the contrary. They experienced improved levels of care because they talked with their care providers about their participation in the OSG (Houston, Cooper, & Ford, 2002). Some also reported that taking part in the OSG encouraged them to become more active health care consumers, because their online peers stimulated them to ask their providers questions, for example, about medication or changes in medication. However, the fear that people with a mental illness might predominantly rely on OSGs still exists, for example, with conditions such as eating disorders (ED), where it is known that those who suffer from an ED are often not receiving nor actively seeking regular treatment (see Dancy & Dooley, 2007; also Hoek & Van Hoeken, 2003).

As for adverse effects, one area of discussion has focused on whether talking about mental health problems with fellow sufferers online would result in a negative spiraling effect, and therefore increase problems. This was considered a particular risk with people who suffer from depression (see Oudhoorn & Somers, 2006). In an OSG for troubled young people who engage in self-harming behavior, worries exist they might experience a “triggering effect” by participating in OSGs. Such a triggering effect could increase their self-harming behavior. A CA study of the interactions in this self-harm OSG showed that its participants adhered and reinforced forum rules that were set up to avoid triggering behavior or responding to triggers from other people on the forum. However, the study also showed that some participants struggled trying to follow these rules (Smithson et al., 2011a, 2011b).

Nowadays, and increasingly so, health professionals have also found their place on the Internet as online care providers. They have started to use the Internet for therapeutic purposes and provide different types of e-therapy and online counseling. As a result, people with mental health problems may not only join (expert-led or moderated) OSGs but also seek professional treatment online via counseling programs offered via chat or e-mail or mobile technologies (Oravec, 2009).

Where extensive research has focused on the demographics and the psychological profile of people who join OSGs to discuss their mental health condition, this article highlights the interactional or conversational practices people engage in when they have entered an OSG or online counseling session via chat and e-mail.

Areas of Mental Health Covered in Research on Online Talk

From the examples discussed above it may already become clear that people who suffer from a mental health condition are not a homogenous group, and that it is important to study each forum in its context. Empirical studies that have applied a CA(-related) approach to study online mental health talk have conducted studies focusing on the interactional practices of people with a large variety of conditions. They have examined in detail how interaction unfolds in groups for bipolar disorder (Vayreda & Antaki, 2009), eating disorders (Stommel, 2009), depression (Lamerichs, 2003; Morrow, 2006), emetophobia (i.e., fear of vomiting; Giles & Newbold, 2013), chronic fatigue syndrome (Guise, Widdicombe, & McKinlay, 2007), bullying (Osvaldsson, 2011), suicidal thoughts (Horne & Wiggins, 2009; also Kupferberg & Gilat, 2012), self-harming behavior (Smithson et al., 2011a, 2011b), and bereavement (Paulus & Varga, 2015; Varga & Paulus, 2014). Studies of online talk about other health conditions, cancer being one large field of study, are also conducted (see, for example, Winzelberg et al., 2003, on breast cancer; Seale, Charteris-Black, MacFarlane, & McPherson, 2010, on prostate cancer) as well as studies that have examined health in relation to food, for example, in discussions on veganism and dieting (Sneijder & Te Molder, 2004, 2005, 2009).

Studies that have examined the interaction between clients and health professionals in online counseling services have looked at counseling interactions about (minor) depression, and anxiety disorder (Lamerichs & Stommel, 2016), at suicidal help seekers who receive online counseling by trained volunteers (Gilat & Kupferberg, 2010; Kupferberg & Gilat, 2012), counseling related to addiction (Stommel & Te Molder, 2015) and counseling targeted at children and young people who are bullied (Danby, Butler, & Emmison, 2009; Harris, Danby, Butler, & Emmison, 2012).

Changes in Research Focus Over Time

The focus of the above studies has changed over the years, and studies of OSGs have preceded the work in the field of online counseling, which matured at a slightly later date. Early studies on OSGs have largely examined the actions people accomplish in their initial messages to a health forum (in analyses of newcomers’ posts) and the ways they present themselves and their health concerns in these posts (Antaki, Ardévol, Núňez, & Vayreda, 2005; Lamerichs, 2003) or the practices by which newcomers find their way into an OSG as an online community (Stommel & Koole, 2010). Attention has then shifted somewhat to studying larger interactional projects such as advice giving (cf. Locher, 2012; Morrow, 2006, 2012). However, for every study that addresses the online interactional dynamics of an OSG for a “new” condition not yet explored, the above themes are recurrently addressed (cf. Paulus & Varga, 2015, analyzing newcomer posts in an online grief support forum and Smithson et al., 2011a, for a similar focus in an OSG for youngsters who self-harm; see also Herring, 2012, for a discussion about the continuous need for such basic descriptive research).

A more recent development is that analyses of online talk are starting to explore how interaction unfolds after a first opening post to a forum and its first reply to examine responsive actions that occur over a larger number of posts in what constitutes a discussion thread (see Giles, Stommel, Paulus, Lester, & Reed, 2015, for an overview).

A recent review study of scholarly work using a CA approach to study online interaction showed that applying interactional approaches to examine the practices in OSGs seem steadily on the rise (Paulus, Warren, & Lester, 2016). Its growth also includes an increase in studies that are conducted by scholars outside the United Kingdom and the United States, which for long have been the two countries that were the most prolific contributors to the field. Overall, OSGs in languages other than English are being studied too, including Spanish, Swedish, Norwegian, and Dutch (i.e., they constitute 25 out of 89 studies that were identified in the review study, where first authors from the United States, the United Kingdom, and Australia contributed to the field with 24, 23, and 5 studies respectively; see Paulus et al., 2016, for a full account). This is important as it creates a broader cultural basis for CA studies of online mental health talk.

Studies that have examined online counseling talk have also showed a development over time, although this domain is much younger. An initial interest has been to examine how well face-to-face counseling translates to online environments. Studies have focused on how the relationship between counselor and client is constructed and managed in online environments (Ekberg, Barnes, Kessler, Malpass, & Shaw, 2013; also Stommel & Van der Houwen, 2013). Other studies have addressed particular phenomena, such as the characteristics of clients’ problem presentations and subsequent responses, or how an interactional phenomenon like a summary (also called a “formulation”) is conducted in chat services (Danby et al., 2009). A somewhat different area of research has focused on how health professionals manage proposals to change between different modalities, such as a transfer from telephone to e-mail counseling (Harris et al., 2012). It is important to note that the field of online counseling is younger and much smaller than the research conducted on OSGs.

Main Research Themes

This historiography aims to offer a first gloss on some of the more specific areas of research, starting with OSGs. These areas will be addressed in greater empirical detail in the section “OSG Talk as Interaction: Empirical Examples.” Research conducted on online talk in OSGs contains a number of well-researched topics, of which we want to present two main examples: (1) newcomers’ posts to OSGs and (2) participants’ interactional efforts at constructing and maintaining OSG norms.

First, the field has paid much attention to how newcomers enter an OSG and how they present their health concerns or problems (sometimes also referred to as “problem messages,” cf. Morrow, 2006; but it has to be noted that the actions that are fulfilled with these initial messages accomplish more interactional work than “presenting a problem”). In examining newcomers’ actions, CA(-related) approaches have provided a more nuanced view of OSGs as typically low-threshold services. They have done so by explicating the subtle interactional work writers accomplish in their first posts to acquire membership, for example, and how these posts are then received, offering a more granular view of these practices.

The analyses that address the actions of newcomers are often also captured under the header of “identity work,” a notion that reflects the sense in which CA(-related) research considers identity not a static but a performative category that is constructed in discourse and is therefore emergent and oriented toward action (Benwell & Stokoe, 2006; also Antaki & Widdicombe, 1998). Newcomers might, for example, actively construct an identity of someone who is competent in managing one’s condition and not necessarily in need of help or looking for advice (cf. Morrow, 2006).

A second research theme in studies on online talk about mental health has to do with the norms that participants invoke in OSGs. This has been topicalized in different ways, by exploring the local norms that seem to operate within different groups (Lamerichs & Te Molder, 2003) or by showing the ways in which participants guide newcomers to “appropriate” views of their illness (i.e., they may highlight a strict biomedical account that prescribes an official diagnosis as a condition for entering an OSG; see Vayreda & Antaki, 2009, for a full analysis of online talk in an OSG for people with bipolar disorder).

A helpful concept to consider how “group norms” are invoked through interaction is to perceive of an OSG as a “community of practice” (cf. Stommel, 2009; Stommel & Koole, 2010). As part of an interactional approach that explores online talk, this concept highlights the local interactional practices participants engage in, beyond their mere membership, and how with these practices they may uphold and safeguard the online community (i.e., through formulating strongly worded advice) as well as conduct “interactional boundary work,” vis-à-vis inappropriate posts (Smithson et al., 2011b).

Main areas of research in the domain of online counseling may be less visible yet, but it seems that there are two areas that have attracted most attention in this developing area of research so far: (1) studies that highlight the relationship between the counselor and the client; and (2) the way clients present their problems, and how these problem presentations and the interactional work they accomplish are taken up by the counselor. In the section “Online Counseling as Interaction: Empirical Examples,” further empirical examples of online counseling will be presented.

How clients in online counseling or participants in OSGs present themselves (as “experts” on their condition, for example) and the identity work thus established is a topic that surfaces in both areas of research on online mental health talk.

What Constitutes an Interactional (CA) Approach?

The methodologies that are drawn upon to study online health talk start from the generally accepted idea in CA that the sequential placement of messages can also be applied to the study of online talk. It is proposed that “. . . any turn at talk (or any post) both enacts a certain action (or more than one action) and also opens up a given space for the next turn to perform a corresponding action” (see Vayreda & Antaki, 2009, p. 932; also Sacks, Schegloff, & Jefferson, 1974). They also depart from the central notion of recipient design, which for the study of online talk can be described as “the multitude of respects in which electronic communication is constructed or designed in ways which display an orientation and sensitivity to the particular and more general other(s) who are receiving or co-participating in this electronic communication” (Meredith & Potter, 2013, p. 372).

Online talk can be understood for the ways in which actions are sequentially constructed. The analyst’s task of naming or labeling those actions in an appropriate way is based on an important methodological principle in CA, which is called the next turn proof procedure (Sacks et al., 1974). In line with basic assumptions in ethnomethodology (see Garfinkel, 1967), any description of actions by the analyst should reflect that one action makes another action conditionally relevant (a request for information makes a reply in which information is provided the expected next post). Applying the next turn proof procedure shows how the previous turn is taken up by the recipient (e.g., this might show that the request for information is not-only-received as a request to provide information but also as an invitation to express empathy). Similar to turns at talk, but often found in the relatively long messages written to an online forum is the fact that in one post multiple actions can be constituted. It is not uncommon for a response in a forum to express understanding, provide advice, ask a question for clarification, and invite the writer of the initial post to continue the conversation, all in one post.

To complicate matters a little more, we know that in many OSGs, interaction does not always present us with complete action pairs or sequences; sometimes a response to the initial post remains wanting or is delayed, or if a response occurs, a further reply by the writer of the first post remains absent. The particularities of online environments in terms of sequential placement, such as the disruption of adjacency and also the sequential delay of messages, can nevertheless sometimes still be observed, as contributors often explicitly orient to earlier posts when they respond (Reed, 2001; Severinson Eklundh, 2010). Some of the methodological challenges this may result in, are addressed in the later section, “Methodological Issues.”

Examining Mental Health Concerns in Online Talk as a Members’ Category

If we use an interactional approach to study online mental health talk, two relevant questions are these: (1) What counts as an interactionally relevant concern for someone with a mental illness and how is that observable in the discourse?; and (2) What range of social actions do participants accomplish with online mental health talk?

Focusing on these interactional concerns as members’ categories is an analytic starting point that is also inspired by ethnomethodology (Garfinkel, 1967). It is also central to the methods we have captured under the header of CA(-related) approaches (i.e., conversation analysis or CA, see Sacks, 1992; Sidnell & Stivers, 2013; discursive psychology or DP, see Edwards, 1997; also Edwards & Potter, 1992; and membership categorization analysis or MCA, see Sacks, 1972; Schegloff, 2007; also Stokoe, 2012).

Members’ interactional concerns or also sometimes called participants’ categories (as opposed to analysts’ categories) are not to be understood as participants’ individual “goals” or “intentions” for participating in OSGs, nor are they consciously aimed for. Rather, these interactional concerns become visible from a detailed examination of online talk and a focus on what participants themselves construct as relevant matters in their contributions when they respond to each other’s messages. As was mentioned before, the next turn proof procedure is an important analytic procedure to control that we as analysts maintain a focus on the understandings participants display to each other in their discourse, and use this analytical rule to come to appropriate action descriptions (see Guise et al., 2007, for a description of the steps they took to explore the action-oriented properties in CFS talk).

A concrete example of what an interactional concern might look live, can be derived from studying the interactional practices in an OSG for people who struggle with grief after having lost a loved one. One of the concerns participants orient to in their discourse in this support group is to establish their grief as “legitimately devastating.” One way in which this is established is by emphasizing the young age at which their partner has passed away as a cause of exceptional sadness (Varga & Paulus, 2014). Identifying interactional concerns and the discourse strategies used to address these concerns are important analytic tools for researchers to arrive at action descriptions that reflect the interactional work participants are attending to, rather than analyst’s informed guesses.

In what comes next, we present a data extract as an example to see how messages to an online forum can be studied as a sequence of actions and to illustrate what CA(-related) observations of this data extract may look like.

Applying an Interactional Approach to an Illustrative Data Fragment

To use a CA(-related) approach to online talk highlights the actions people accomplish with their messages to a forum (or a counselor), which can be demonstrated by considering how these messages are taken up in subsequent replies. We present an example consisting of three messages that have appeared in a sequence (a “thread”) of nine messages in total: (A) an initial opening post by Anonymous; (B) a response by someone whose name has been omitted for privacy reasons but who uses a name (possibly a pseudonym) and whom we will refer to as Name Omitted; and (C) a reply from Anonymous. The extracts are based on Morrow’s study of an OSG on depression (2006). Note that they are conveyed exactly as they appear online, that is, without correcting for any grammatical errors.

Extract (A)

Discussion: depression

From: Anonymous

Headline: Is there anyone out there?

i don’t realy know what to do. ive suffered from depression a lot in my life. Has anyone got any suggestions on how to relax, as I cannot seem to. I have been recently put on diazepam as I get a tight chest, and trouble sleeping, but the diazepam hasn’t seemed to help . . .


Discussion: depression

From: (Name Omitted)

Headline: advice

i think what you need to do is to get right to the heart of the problem. you are not going to be able to relax until yu doeal with the problem itself. Can you tell me what makes you depressed?

as for relaxation, join a yoga class! i did this and it really helps to relax you, though don’t expect it to make your problems go away. Obviously there is a deeper issue here and you have to address it before things get worse. Feel free to talk to me. i’ll always listen, ok?

love (Name Omitted)

p.s. you WILL get better


Discussion: depression

From: Anonymous


Hello again, thanks for the advice, one girl i know suggested i also write down any thing i end up worrying about, so that I can look at it later and deal with each issue individually.

Im off work at the moment, ad very unsure about going back, the thing is though, i know that the longer i leave, the worse im gonna be.

I don’t really know what to do…

If we want to make observations about these extracts from a CA perspective, we might start by mentioning that the headline of extract A is phrased as a question (“is there anyone out there?”), which seems to be an invitation to those present on the forum to respond. The action accomplished with this headline seems to differ somewhat from the start of the body of the message which seems to constitute a cry for help (“I don’t really know what to do”). This initial cry for help in the body of the message is then followed by a request for “suggestions,” (“has anyone got suggestions on how to relax”).

Morrow (2006) mentions that this opening post differs from the other data in his collection in that it contains a relative straightforward request for some “advice” (at least in the body of the message), although we already observe some ambiguity as to the different action that can be identified in this first post to the forum. Other CA studies confirm that requests for advice are not always formulated under that label, and may come in different forms (cf. Vayreda & Antaki, 2009; see also Locher, 2012).

The reply in extract (B) by Name Omitted, which appears one day after the posting from Anonymous contains the subject header “advice.” Note how this is a slightly different term from the words used by Anonymous, who asked for “any suggestions.” Name Omitted starts her message with a piece of advice (“i think what you need to do”) that stresses the need to engage in some self-reflective action, suggesting that Anonymous will otherwise find no lasting “cure.” It might be hearable as a mild critique too. She continues by inviting Anonymous to continue talking (“can you tell me . . .”) and inquires what the cause of his depression is.

In the second part of her reply, Name Omitted presents straightforward advice on the topic of relaxation tips, constructed as an imperative that establishes the order in which to take her “instruction” (“join a yoga class!”). She presents it as a course of action that “really” worked for her, which presents the advice as based on her own experience. This might add to the credibility of her advice, as in what she offers also proved to be an effective strategy. At the same time she specifies and possibly downplays the scope of what she has just recommended: If relaxation would happen as a result of the yoga, it would “obviously” not mean that the underlying problems are solved. Hence, she strongly claims not only that these problems are there but (again) that they need to be dealt with (“you have to address it”). In closing, Name Omitted offers an implicit invitation coupled with an intensified claim (“i’ll always listen, ok?”) and a postscript that offers strong (capitalized) encouragement.

Extract (C) by Anonymous was posted three days after the second post. It does not contain a header and Anonymous opens in a general way by thanking for “the advice.” This leaves open whether it is to be understood as a specific thank you to Name Omitted and may also demonstrate the lack of explicit uptake as advice, something we see more often in OSGs, as no reference is made to the particular suggestions presented by Name Omitted.

Anonymous continues by providing more background information about his current situation; he displays an awareness of the problematic consequences of his problems (“I know that . . . the worse im gonna be”) and provides an intensified description of his situation in the future. He closes with an even stronger, general statement of hopelessness.

These extracts demonstrate the two main areas of research in CA(-related) research on online mental health talk that were mentioned earlier: identity work and constructing group norms (see also sections “Category Ascriptions and Identity Work” and “(Re-)invoking Forum Rules”). If we begin with the latter we can see that the writer of the second post strongly orients to the need to take the underlying problems seriously, which can help to construct a “preferred view” of how to deal with the condition of depression and not settle for “easy solutions” that may address only the symptoms (cf. Vayreda & Antaki, 2009).

In terms of identity work, three further observations can be highlighted based on the full analysis of the dataset (Morrow, 2006), which resonate with other findings on OSG talk. A first point is that after offering a general description of their problem, writers go to great length to describe their symptoms. Often they elaborate with descriptions of (negative) feelings, which they convey by making use of metaphorical language. Metaphoric language can be used in order “to express the inexpressible” (Green & Kupferberg, 2000). As what one experiences might be highly individual, these metaphors work by adding to the vividness and also the seriousness of the writer’s current state of mind. An analysis of how patients in an OSG for chronic fatigue syndrome (CFS) use metaphoric language and extremized analogies to convey their experience of the illness, has also illustrated this (Guise et al., 2007). Participants may describe the seriousness of CFS in the following way: “remember your worst flu virus you had, now run a marathon, on top of that you have a migraine, feel dizzy, hurt from head to toe” (2007, p. 96).

Second, and building on the observation that requests for advice in OSGs occur in different linguistic shapes, Morrow identified the main forms in which requests for advice were done. Whereas requests for advice were often formulated as general request (“what can I do to help myself?”; “has anyone got any suggestions about how I can make myself feel better?”; Morrow, 2006, p. 541), he also noted that requests for advice, mostly related to a proposed self-diagnosis, were frequently formulated as “alternative questions.” These questions consist of two alternatives that allow the writer to suggest a sense of competence in considering them as relevant alternatives for the current problem (“can i really get it back to normal by forcing myself to ‘do’ stuff or do I have to wait till my mind heals,” 2006, p. 540).

A third strategy identified in these initial “problem messages,” as Morrow calls them, was the hesitant ways in which writers open their initial post, as if they incidentally stumbled upon this OSG (“I was just surfing when I came across this site . . .”). This makes available similar inferences about the writer’s competence, who might be seen to portray himself as “a person who was asking for advice simply because it was easily available and not as someone who was in trouble and in need of advice” (Morrow, 2006, p. 541). We could argue that our initial observations about the opening post displayed here, showed a slightly different type of ambivalence, that is between the general invitation present in the headline and the claim to helplessness presented in the body of the message. At the same time however, the post also illustrates ambiguity as to whether it is to be understood as requesting advice.

Our observations illustrate the subtle kind of identity work that writers of opening posts engage in, when requesting advice. Formulations whereby advice is given in OSGs are equally subtle but carry maybe even more interactional “risks.” We find that they are often coupled with hedges and displays of empathy; they may be formatted as indirect statements, questions that propose courses of action, “should-instructions” or “I would do x statements” (Locher, 2012). These lexical features show participants’ orientation to the sensitivity of providing advice in a way that is appropriate, relevant, and also nonthreatening for the receiver (Smithson et al., 2011a, 2011b; also Lamerichs, 2003;and Morrow, 2006, 2012).

Acceptance of advice by the person who requested it, whether it was explicitly sought or not, turns out to also be a sensitive matter. Straightforward acceptance might suggest that that the problem was (too) easy to solve (Smithson et al., 2011a, p. 494), which might carry a whole different set of potentially undesirable inferences (i.e., as to the seriousness of the problem, or its urgency). The section “Practices of Advice Giving in OSGS” presents more empirical examples of how requesting advice and advice giving in OSGs may take place and demonstrates how it may be influential for the ongoing interaction.

Mental Health Talk and Accountability

The interactional concerns that participants in online mental health talk orient to in their contributions might lead us to think that they are, at least at a first glance, not so different from what we know about “illness talk” more in general. It is convincingly shown how peoples’ descriptions of illness in face-to-face interaction are “hedged around with what have variously been called warrants, concerns with blame, or efforts at legitimation” (Radley & Billig, 1996, p. 224). CA research on spoken interaction in medical settings offers a useful set of sensitizing concepts that point in a similar direction, partially derived from studies of medical sociology, such as Parson’s influential work on the “sick role” (Parsons, 1951). This concept describes how people who are ill orient to rights and obligations to do with their illness (i.e., to treat the sickness not as the patient’s own fault and recovery of the illness not as a mere “an act of will”; while also viewing the sickness as undesirable). Conversational evidence of how these concerns surface in the interaction, can be found in CA studies of doctor–patient interaction, where patients present their health problem as legitimate or “doctorable” (Heritage & Clayman, 2010; see also the section “Talk As Interaction: Empirical Examples” for a more detailed exploration of parallel concerns in online mental health talk).

Participants in an online support group also seem to construct their posts to legitimize their condition, to avoid feeling stigmatized or judged, or to prevent from being seen as incapable of handling their problems. They may also work up descriptions that resist particular dominant and socially normed views of their condition. This may enable them to create a sense of normality and validation for their experience, especially when these experiences are different from generally accepted views, for example, about what is considered “normal” grief (Varga & Paulus, 2014; also Giles & Newbold, 2013).

CA(-related) approaches have begun to demonstrate the systematic ways in which these concerns are presented in OSGs. A comparative study that examined the descriptions people with CFS provide of their illness experiences in face-to-face talk and online talk, showed that these spoken and written descriptions were sensitive to similar concerns: to counter a particular version of the illness. When asked to describe their experience with the disease, CFS was similarly described as serious, enigmatic, and “not a psychological” condition in both settings (Guise et al., 2007).

The fact that much of the conversational practices in OSGs are oriented toward legitimizing health concerns can also be captured with the notion of “accountability.” Following earlier work (Garfinkel, 1967; Scott & Lyman, 1968) accounts are “verbal responses offered by individuals to others designed to protect the ongoing social interaction from the disruptive consequences of problematic events. Accounts attempt to explain apparently ‘untoward’ or ‘unusual’ events as understandable, or at least to minimize the actor’s responsibility for them” (Buttny & Morris, 2001). Accountability can be defined as the ways in which participants, “do things for which an account might, in principle, be called for and given” (Antaki et al., 2005, p.115).

Much work on how accountability plays out in online interaction has been done in discursive psychology or DP (Lamerichs & Te Molder, 2003), for example, by considering it in relation to identity work displayed in the opening contributions to an OSG for people with depression.

However, accounting for a mental condition in an OSG may be geared toward more aspects than justifying the condition in a broad sense; it may be a multilayered interactional project. In the section on “OSG Talk as Interaction: Empirical Examples,” we will unpack what accounting practices look like in OSGs with illustrative examples.

OSG Talk as Interaction: Empirical Examples

On one level, accountability in OSGs can be related to the structural properties of asynchronous online discussion forums. When writing a post to an online forum, participants can employ these structural features of the technological environment to initiate a post and start a new thread. In this new thread they can address the forum as a whole or someone in particular, but opening posts are generally addressed to the former.

Let us take as an example a Spanish online discussion forum (not specifically related to the topic of mental health talk) in which a user who self-identifies as Lourdes has posted a new, not immediately relevant message in the thread “Miscellaneous” (see Antaki et al., 2005, for a full analysis). Its subject line reads “‘For she who knows who she is,’ from Lourdes [written] 15.37 today[author] Guest.” The asynchronous, poly-logal1 features of an OSG in a technological environment that prescribes that visitors have already navigated to this thread, and are therefore (at least potentially) ready to be addressed as a reader, poses at least two implications for the accountability of the writer of this thread-initiating post:

  • it diminishes the accountability for the writer of an “unoccasioned” first post in a thread and renders it not particularly vulnerable for criticism of being the first post at that very moment in time;

  • sending a first post in a thread is not particularly risky in terms of nonresponse. The sender cannot be held accountable if a nonresponse occurs as it can always be explained away by attributing it to participants not having seen or read the message.

For the person who replies to this first message in the thread, the implications in terms of accountability work out somewhat differently. The first response to this opening post is, “what a way to make us green with envy, girlie.” The person who replies thus offers an account for taking that next turn; by expressing envy of the person Lourdes is referring to, the sender renders it explicit that she is not the intended addressee.

Accountability however does not relate to the structural properties of the medium alone. Moreover, CA(-related) studies of newcomers’ posts to OSGs on mental health talk have demonstrated that accounting for first posts often does occur, both when initiating first posts and also “further down the thread,” albeit in different ways. Based on a study of a German OSG for people with eating disorders, this multilayered sense of accountability is coined “forumability” (Stommel, 2009; see also Heritage & Clayman, 2010, for a link with “doctorabilty”). In OSGs, forumability sums up the practices participants establish to construct their current mental health concern as “legitimate for the service” (Stommel, 2009, p. 155).

In what comes next we want to highlight three ways in which forumability plays out in participants’ contributions to an OSG:

  1. 1) category ascriptions and identity work;

  2. 2) (re-)invoking forum rules;

  3. 3) managing bids for membership.

We will conclude by discussing a fourth dimension, that represents an interactional practice in OSGs that ties together all the above aspects:

  1. 4) practices of advice giving.

To suggest a clear distinction between the above three aspects is somewhat tricky as they constitute partially overlapping concerns that might also be oriented to simultaneously in one turn (here: in one message or post). But it is still useful to work with this distinction, because it enables us to see that the accounts accomplish different types of actions in online talk. Thus, it may become clear that forumability may be geared toward managing appropriate descriptions of the condition and presentations of the self, as well as constructing these descriptions in such a way that they correspond to and are constructive of particular forum norms. And it is precisely by attending to these forum norms that an initial post to a forum works to constitute a bid for membership, which can then be responded to accordingly. We will now present some illustrative examples of these three dimensions of accountability in OSGs.

Category Ascriptions and Identity Work

In their postings, participants may present themselves as “forumable” by describing themselves in terms of relevant categories and category ascriptions. They may use categories of diagnosis, duration of the condition, weight, gender, or age. To examine how these categories are drawn on and how their inferences work as interactional resources for participants in an OSG implies that we do not treat them as a cognitive schema in peoples’ heads that can bring order to the world (note that this treatment is similar to how CA and CA-related approaches “define” the notion of identity, as oriented to action, as was pointed out in the section “Main Research Themes”).

Thus, drawing on categories is considered a tool for social action in two ways: (1) by drawing on particular categories in their discourse (presenting oneself as a “single mother” or as a “psychologist who suffers from depression”) participants make relevant a collection of other related categories (e.g., professional or relational identities, particular knowledge claims); (2) every category carries “predicates,” which are the activities or attributes that are commonly associated with the category and which carry a set of different implications that can be drawn upon in interaction, both by the writer of a posting and the recipient (e.g., to do with “legitimate” causes of depression, or types of knowledge that can authoritatively be brought to bear in a discussion forum or not; cf. Sacks, 1972).

The so-called category work thus accomplished has been thoroughly explored in studies of newcomers’ posts to OSGs (see Giles & Newbold, 2013; Horne & Wiggins, 2009; Lamerichs & Te Molder, 2003; Varga & Paulus, 2014). In opening posts to an OSG for people with depression, new members often describe themselves in ways that counter the negative attributes of the identity category of “being depressed” (e.g., being “weak,” “unfit”). The use of these categories may then be embedded in descriptions in which they present themselves as “intellectually demanding,” as someone with a “heightened sensitivity” or as a “single mother who is able to manage a busy household and a full time job while faced with a long history of depression.”

A close examination of the category work participants accomplish in their online talk, allows us to see how they are drawn on to present oneself in particular way, as depressed but competent, for example, and how this is accomplished in a range of subtle ways (see also Guise et al., 2007).

(Re-)invoking Forum Rules

To present oneself as “depressed but competent” in an OSG for people with depression turns out to be only one side of the coin. We see that in this support group participants also orient to exposing a lack of competence, albeit temporarily. Hence, describing one’s illness is subtly managed to include orientations to showing one’s vulnerable side (or “incompetence”), as well as demonstrating a certain competence.

In extracts I and II we see how showing one’s vulnerable side (i.e., by addressing down feelings) is presented in ways that show how participants collaboratively create an “interactional space” to share these feelings and formulate some “conditions” for doing so. This is the way forum rules are invoked and get confirmed in the interaction. The contributions by Betty and Janice in extracts I and II illustrate the normative expectation to be able to display your vulnerable side, as well as how sharing emotions is treated as a matter to account for (based on Lamerichs, 2003).

Extract I [Lamerichs, 2003, p. 127]

Posted by Betty

  1. 1 I am sorry that I am upsetting everyone. I don’t want to depress

  2. 2 anyone or make them cry. Posting my honest feelings was my

  3. 3 therapists idea. He thought I needed to be able to talk honestly

  4. 4 about what I am feeling somewhere, but I don’t want to take my

  5. 5 friends down with me. Betty

Extract II [Lamerichs, 2003, p. 132]

Posted by Janice

  1. 1 Betty- you are NOT taking me down with you!! And you most

  2. 2 certainly should post your feelings here of all places. Here is

  3. 3 where you can be totally honest and there is no judgementalism or

  4. 4 censure. Please keep it up –it’s good for what ails you!!

  5. 5 [7 lines omitted]

  6. 6 This folder is for exactly what you are doing here. Don’t give up

  7. 7 on the ones who care the most for you

One element of what appears to constitute a forum “rule” here is referred to by Janice in extract II, in response to Betty’s message. Showing emotions is constructed as having two sides: First it is described as a moral obligation to the group; not showing your vulnerable side would mean to “give up” on other forum members and prevent them from being able to care for Betty (lines 2, 6–7). And second, the forum is also reinforced as the place where Betty can show her emotions without any repercussions (lines 3–4). Hence, Janice (re-)confirms and constructs the OSG as a place where one can legitimately share emotions.

A double orientation to what counts as a possible “forum rule” for sharing emotions is also attended to in Betty’s first post in extract I. She starts by offering an account for a previous message in which she apparently shared her down feelings (data not shown here). Her account includes attributing the decision to post her “honest feelings” to an expert third party (lines 2–3), which is also presented as the party who stressed the need to show those feelings. The claimed importance by this authoritative source is then downplayed by contrasting it with her own need to keep in mind the well-being of the other forum participants (referred to here as “my friends”) and to prevent any negative effects from her previous “outpourings.”

How participants orient to forum rules in their online talk can thus be made visible when we examine this sequence in extracts I and II. We have seen how forum rules may be collaboratively constructed as containing apparently opposing normative requirements (i.e., reporting one’s emotions to keep the community “working” while accounting for its adverse effects). The subtle management of forum rules that we can thus observe shows that what could be considered a low-threshold service does in fact constitute a highly regulated social practice.

Forum rules are often invoked in responses to opening messages, for example, when ratifying the sender’s displayed intentions to seek help or an official diagnosis, rather than denying the condition (Stommel, 2009, chap. 6, for a full account). In their opening posts to an OSG for people who are suicidal, participants also show a concern with what was established as a normative requirement in this group: to strike a balance between displays of being “too suicidal” versus “not being suicidal enough.” This has been described as “managing an authentic suicidal identity,” in which the members of this OSG collaboratively engage (Horne & Wiggins, 2009). Extracts III–V below show how the rule to present oneself as authentically suicidal is made visible in a sequence of posts, starting with Caroline’s announcement in extract III that she will be committing suicide.

Extract III [Horne & Wiggins, 2009, p. 176]

  1. 1 Caroline, Posts 628; 12:03 p.m.

  2. 2 Almost did it yesterday. Will do it today. Have nothing

  3. 3 left keeping me here. It’s worse than it’s ever been and

  4. 4 those i love have abandoned. It’s not worth it anymore.

  5. 5 i’ll miss you. i’ll always love you.

  6. 6 this is someone’s fault

Extract IV [Horne & Wiggins, 2009, p. 177]

  1. 1 Martina, Posts: 1313; 12:54 p.m.

  2. 2 (((((Caroline))))) NO hon- hang on! It will get better. We love

  3. 3 you and don’t want to see you go…we need you hon! Please get some

  4. 4 help, please!

  5. 5 Its someone’s fault you say- write it out here. Let ‘er loose-

  6. 6 maybe we can help.

  7. 7 Take care of yourself please hon- I don’t want to see you hurting

  8. 8 this badly.

  9. 9 (((((((((((hugs))))))))))

Extract V [Horne & Wiggins, 2009, p. 177]

  1. 1 Caroline, Posts: 628, 09:46 p.m.

  2. 2 Just here to say i appreciate all the replies, each and every one, 3 and that i’m trying as hard a i can to keep going. i’m trying

  3. 4 really hard. really hard.

In extract III, Caroline orients to being desperate and invokes a sense of immediate threat (“will do it today,” line 2), while also orienting to a “rational” weighing of options left (“have nothing left keeping me here”; “it’s worse than it’s ever been,” lines 2–3). Line 5 is reminiscent of the endings of suicide notes, and uses a type of formatting often used in such notes (“I’ll miss you,” “I’ll always love you,” line 5). In line 6, addressing the issue of blame is presented as an externalized account. However, it may also open up an opportunity for ongoing interaction rather than taking the final step, suggesting there may be a way out. The latter is precisely what is taken up in the reply she receives from Martina, in extract IV (line 5).

What is noteworthy with respect to creating forum rules is that Martina’s response treats Caroline’s post as a sincere notification that she is about to commit suicide (“hang on,” line 2 in extract IV) while also treating it as a phase (“it will get better,” line 2 in Extract IV). The latter is presented authoritatively, as coming from her own experience (see also Horne & Wiggins, 2009, for a full analysis of invoking similar experiences). In Martina’s reply, Caroline’s previous reference to “this is someone’s fault” is then taken up; not as something to argue against, but as an opportunity to invite Caroline to keep writing about this in the forum. Similar to what we’ve seen in the example in the depression forum, Martina’s reply also orients to Caroline’s role in the continuity of the forum’ (“we need you,” line 3). Beyond invoking this as a forum rule, it might also be an interactional strategy to deal with the immediate threat voiced in Caroline’s post to kill herself.

In extract V, posted on the evening of her initial post, and as the 10th post in the thread established after her first post, Caroline replies. She makes an explicit appreciative reference to (all) the replies she has received. She also attends to her accountability of not having committed suicide while she previously suggested that she would (“today”), and emphasizes her continuing struggle (see lines 3 and 4).

In this sequence of messages we can see how the forum rule to display an “authentic” suicidal identity is collaboratively established in the interaction. Forum rules may be complex and can involve a double normative orientation: presenting oneself as depressed and competent, but also showing one’s temporary bouts of “incompetence.” The latter also involves further subtle accounting practices, as we saw in extract I. The other examples showed how participants manage and confirm “authentic” suicidal identities, which encompass announcing the deed and the legitimate reasons for it, as well as presenting equally sincere claims for why they have not yet killed themselves.

In what follows we discuss how the responses offered to newcomers’ posts treat those posts as bids for membership, and how these are then interactionally dealt with.

Managing Bids for Membership

While constructing forum rules may happen at any point in the interaction in an OSG, bids for membership usually occur in participants’ initial posts and the subsequent replies they receive (Paulus & Varga, 2015; Smithson et al., 2011b; also Lamerichs, 2003). The way in which writers display eligibility for membership, and how these messages are responded to, will be illustrated in the next extracts.

Extracts VI and VII are two contributions to an online grief support group. In their original study, the authors show how participants in this OSG come together to validate their experiences of “extreme” grief in a discursive online space that facilitates resistance to dominant views of what constitutes “normal” grief (Paulus & Varga, 2015). Participants’ descriptions display this orientation by emphasizing the unusual and devastating circumstances of their loss and their extreme and uncontrollable emotional and physical states. They also display “troubles-telling” rather than requests for advice (Jefferson & Lee, 1981). Membership in this OSG for the bereaved is “granted” by validating the experience in the initial post as normal, and by receiving it with empathy. This is frequently done through the provision of second stories (Paulus & Varga, 2015; see also Arminen, 2004; Giles & Newbold, 2013). Extract VII is an example of such a second story, which “mirrors” the concerns that were presented in the previous post.

Extract VI [Paulus & Varga, 2015, p. 446]

  1. 1 I keep thinking this is all just a horrible nightmare and when I

  2. 2 wake up, he will be right here with me … but when I wake up. I

  3. 3 wander around it seems just waiting for him to appear … Of course

  4. 4 he doesn’t, and that is when I start howling like some kind of mad

  5. 5 woman … I feel so helpless so much of the time.

Extract VII [Paulus & Varga, 2015, p. 4]

  1. 1 We all get it. Like my family and some of my friends,

  2. 2 they just do not understand your loss and how it impacts

  3. 3 every minute of every day, every dream. No one can walk

  4. 4 in your shoes but the folks here come very close to it

  5. 5 as we are all walking this horrible path.

In extract VI the writer describes her situation in a vivid way that stresses the depth of her grief. She makes use of metaphoric language in lines 4–5 (see also section, “Examining Mental Health Concerns in Online Talk as a Members’ Category”) and points out how others do not understand or offer little support (“my family doesn’t really understand”; “some people at work have even commented that I just ‘need to get over it and move on’”). The response she receives demonstrates how these concerns count as legitimate reasons to enter this OSG. Note how the writer of the reply, by providing a second story, conveys empathic understanding as well as solutions, and shows to the newcomer “how to put their experiences into the proper context for the group” (Paulus & Varga, 2015, p. 639). Providing second stories might also be hearable as providing advice, without being formatted as such. Hence, it offers a way of managing the competence of the receiver by “merely” providing a parallel or similar experience (see Locher, 2012, for a discussion on how “informative sequences” or narratives of personal experiences are often interpreted as “advice” in OSGs).

This section has highlighted the myriad ways in which forumability is managed in mainly three ways: by drawing on the inferences of categories and category-bound predicates and inferences, by orienting to and confirming forum rules, and by treating newcomers’ posts as bids for membership of an OSG. In this way, participants in OSGs construct descriptions of their mental health condition as “legitimate,” “knowable,” “factual,” “rational,” “authentic,” or “normal.” These descriptions are particularly designed to deal with some of the negative inferences of suffering from a mental condition and the threat this may pose for the writer’s identity (Guise et al., 2007; see also Giles & Newbold, 2013; Horton-Salway, 2001).

Practices of Advice Giving in OSGs

The fourth set of empirical examples we discuss here are practices of requesting and providing advice that forum participants engage in, in which all three dimensions of forumability come together.

Where, as was mentioned before, most empirical studies of online mental health talk still explore opening posts to OSGs and their first replies, we see that increasingly also the additional responses that are produced further on in the discussion thread are considered in the analyses (Smithson et al., 2011a). As for the central practice of requesting and offering advice, this may be accomplished in the first post and the reply it receives, but it may also be taken up further down the thread, over a number of subsequent posts.

The data fragment in the section “Applying an Interactional Approach to an Illustrative Data Fragment” illustrated the basic sequence that could be identified in an OSG: (1) a first post in which a request for some type of advice is formulated, either explicitly formulated or worded more loosely; (2) a subsequent reply (or a series of replies by different participants further down the thread), which may include the offering of advice occurring in different shapes or linguistic forms (see also Locher, 2012); and (3) a response that shows some uptake of the replies. We know that the latter are often lacking in OSGs and we have also seen how explicit acknowledgment of advice is a sensitive matter (Morrow, 2006). The section “Advice Giving Versus Troubles Recipiency” explores the alternative ways in which a request for advice might be taken up, but we will first demonstrate how an advice sequence carries all the dimensions of forumability that were presented before.

Extract (A) already showed that requests for advice in opening posts are often worded rather ambiguously or implicitly, including a “vaguely specified request for help.” In an OSG for people with bipolar disorder, such a request for advice might be formulated as a want to hear about others’ experiences (“It’d be great if someone could share this first stage of acceptance with me, or tell me how they got through it”) or a request for basic information (“I’d like to learn all about this illness”). The former is how Mikokana phrases her request to the OSG in extract VIII below. Extract IX shows Lin’s reply (based on Vayreda & Antaki, 2009).

Extract VIII [Vayreda & Antaki, 2009, p. 935]

  1. 1 “Recently come to bipolarity” - mikokana

  2. 2 18:00 Just recently I’ve been diagnosed with it, which left me

  3. 3 gobsmacked, I hadn’t a clue what “bipolarity” meant it explains

  4. 4 lots of things in my life but it’s going to be hard getting used

  5. 5 to it… It’d be great if someone could share this first stage of

  6. 6 acceptance with me, or tell me how they got through it. Because my 7 initial reaction is a “shit – no way am I bipolar”.

Extract IX [Vayreda & Antaki, 2009, p. 935]

  1. 1 Lin

  2. 2 Welcome. Acceptance is a process. But the quicker you manage it,

  3. 3 the better you’ll cope with the illness, you have to learn to live

  4. 4 with it, follow the treatment religiously, appointments with your

  5. 5 psychiatrist, sleep well… - Lin

In extract VIII Mikokana presents herself as newly diagnosed with BD in a very explicit way (line 1) as a straightforward news announcement. She couples this “news” with descriptions of its impact, which are formatted with extremized emotions that describe her disbelief and surprise (lines 3–4). They index the depth of her feelings over this newly received diagnosis with BP. In the last part of her message, some form of help is asked, in a relatively undemanding way (“it’d be great if someone . . .,” lines 5–6). Based on these features of Mikokana’s first message to the forum, a response could, for example, entail a receipt of the particular feelings she describes, or a display of empathy. Extract IX is the actual reply that Mikokana receives. We see how Lin, the author of the reply, offers strong recommendations, formatted as an imperative (“you have to learn to live with it,” lines 3–4). This is then followed by a number of short and clearly formatted instructions how to behave and what to do (lines 4–5). Besides a welcome at the start of the message, it entails little uptake of the emotional state Mikokana describes, possibly some generically phrased words of encouragement aimed at enhancing coping with the illness (lines 2–3).

It is possible to view the action pair here (i.e., a vaguely specified request for help, and a strongly worded recommendation in response) as a “sequential mismatch,” as the new user is clearly (and unambiguously) told what to do when she seemed to request something that was formatted as much vaguer than that (see also Vayreda & Antaki, 2009, for a discussion). What this apparent “mismatch” demonstrates, however, is that providing advice might be used to do other kinds of interactional work in on OSG. The sequence presented here can be seen to illustrate the three dimensions of forumability: (1) in Mikokana’s opening we see category work at play, when she presents herself as diagnosed with bipolar but also opposing, maybe even resisting the category and its implications); (2) we see how Mikokana orients to the forum rules by presenting herself as recently diagnosed and formulating some request to the group; and (3) we see how Mikokana’s post receives a reply that works to “socialize” the newcomer. It shows how the bid for membership by Mikokana as a newcomer to this OSG is dealt with by providing strongly worded recommendations and also a more generic type of advice aimed at coping prior to that. The response guides the newcomer Mikokana not only how to live but also how to adhere to a particular medical model that is thus constructed as the norm in this group: one that emphasizes to seek an official diagnosis (and to accept it) and follow treatment (see Vayreda & Antaki, 2009, for a discussion; also Stommel, 2009).

The provision of advice can take different shapes, and is not always worded so strongly and explicitly. However, a close inspection of interactional practices in other OSGs shows that the social actions that are fulfilled by providing advice render them important “carriers” of forum rules. In an OSG for young people who self-harm (SH), requests for advice may be formulated as seeking very practical advice: “I SHed yesterday and now my cuts gone green i tried putting an antiseptic -sic- cream on it and then a proper dressing. is this enough?” (Smithson et al., 2011a, p. 491). Different from the strongly worded advice by Lin in the bipolar forum, these requests were not met with explicitly formatted advice but with responses that were typically hedged and coupled with disclaimers (e.g., “not sure hun” or “it’s hard to give advice via the internet”). At the same time, the hedged suggestions also contained “mundane” recommendations such as to visit the GP. Interestingly, the replies that contained these practical recommendations led to further troubles-telling from the person who wrote the opening post. Hence, talking about troubles might “piggyback” on other types of posts, such as first introducing a “solvable” or practical problem (Smithson et al., 2011a). This shows not only that advice can be formatted in different ways, but also that some types of advice are more “conducive” to further ongoing talk about troubles where others are not. The following section presents an empirical example of both instances.

Advice Giving Versus Troubles Recipiency

The ways in which participants engage in practices of advice giving in OSGs show that the interactional concerns that were demonstrated in some classic studies in CA (e.g., Jefferson & Lee, 1981) are also relevant online. For example, when it comes to adopting the conversational role of “troubles recipient,” as opposed to the role of “advice giver” when responding to a “troubles-telling” or a “problem message.” Adopting the former role where the latter is sought or vice versa, may matter in ways similar to how it matters in face-to-face interaction: Offering (preliminary) advice might terminate a “troubles-telling,” or lead to rejection of the advice that was proposed. A study of online talk in an OSG for online grief found that participants preferred to engage in troubles-tellings and refrained from requesting any advice (Varga & Paulus, 2014).

How the differences between these conversational roles and the actions they constitute work out in online mental health talk can be illustrated by looking at how the interaction unfolds in response to extracts X and XI below (based on Stommel & Lamerichs, 2014).

The extracts constitute two opening posts that were posted in different threads to an OSG for people with eating disorders: a thread named “fear” and a thread called “I’m new.” They were written by the same participant who we will refer to as Laura. In extract X Laura inquires whether her situation might mark the onset of bulimia.

Extract X [Stommel & Lamerichs, 2014, p. 203]

  1. 1 07.03.2005 10:13 fear –Laura

  2. 2 hi guys!

  3. 3 Somehow i am afraid that bulimia could also hit me.

  4. 4 i have binging addiction you see and often think of

  5. 5 hanging myself over the loo. But have never done it though.

  6. 6 is that the beginning?

  7. 7 love Laura

Note how the question in line 6 might be considered a rather difficult one to answer but not uncommon in OSGs. And moreover, these are the types of questions that were shown to render advice a relevant next turn (Stommel, 2009). Laura’s question presupposes two things: (1) that the beginning of bulimia can be identified; and (2) that others in this OSG are able to judge whether this is the case, based on the short description of Laura’s situation that precedes her question in lines 3–5.

Extract XI shows Laura’s post in the other thread, in which she poses a somewhat different request. The request follows a description of her continuous craving for food (lines 7–8) and also, her fondness of nibbling (lines 5–6). Laura concludes her post with a question that might be equally difficult to answer: “Is that normal?” in line 9 (see also Giles & Newbold, 2013).

Extract XI [Stommel & Lamerichs, 2014, p. 205]

  1. 1 03.02.2005 18:56 I’m new -Laura

  2. 2 Hi People!

  3. 3 Came across this site by accident and I hope I can find help here.

  4. 4 I am just 17 and have had eating disorders for as long as i can remember.

  5. 5 I constantly have the desire to eat something.

  6. 6 What I like most is to nibble at things

  7. 7 I’ve hardly finished eating something when I need something else.

  8. 8 When shopping an extra bag of crisps or chocolate always has to come along.

  9. 9 Is that normal?

  10. 10 Hope to get answers soon.

  11. 11 Love Laura

A detailed analysis of how the interaction unfolds in both sequences shows that Angela who replies in extract Xb below in the thread “Fear,” treats Laura’s opening post as a request for advice. The provision of advice itself is heavily accounted for by drawing on personal experience as an insufficient basis for advice (lines 3–7). This shows that drawing upon personal experience as an “appropriate” basis for advice in OSGs is a matter of negotiation for participants, rather than a straightforward and “safe” basis to ground advice (see Lamerichs, 2003, for a discussion; also Locher, 2012). After this hedged beginning however, the advice is formatted in an explicit fashion (lines 8–13).

Extract Xb [Stommel & Lamerichs, 2014, p. 203]

  1. 1 07.02.2005 10:30 Re: fear - Angela

  2. 2 Hello Laura,

  3. 3 i can’t tell you where bulimia begins. In my case it gradually

  4. 4 developed from AN to B.

  5. 5 before though i never thought about how it would be to vomit . . .

  6. 6 in my case in the beginning it went just like that. everything

  7. 7 just came out again . . .

  8. 8 if i were you i wouldn’t start thinking deeper along these lines,

  9. 9 who knows what you will

  10. 10 trigger with that.

  11. 11 B is disgusting . . . i know that . . . it is extremely costly, smells

  12. 12 disgusting and is just obnoxious . . . don’t go

  13. 13 there . . .

We will not provide a full analysis of her reply here (but see Stommel & Lamerichs, 2014), but Laura’s subsequent response in Xc below shows initial qualified receipt that is followed by resistance to Angela’s response. In the light of the advice Angela provided, which suggested to make an active decision not to contemplate those matters (“wouldn’t start thinking deeper along those lines,” line 8; and “don’t go there,” extract Xb, lines 12–13). Laura addresses the unavoidability of the thoughts that keep creeping up (lines 3–5).

Extract Xc [Stommel & Lamerichs, 2014, p. 205]

  1. 1 07.02.2005 10:35 Re: fear – Laura

  2. 2 thanks for your response!

  3. 3 you are certainly right.

  4. 4 but it is really quite tough when the thoughts to do it is there

  5. 5 and it doesn’t go away anymore.

  6. 6 because often i just eat and eat everything in, afterwards i feel

  7. 7 nauseous and think, when you vomit, you

  8. 8 will certainly feel better again.

In the thread “I’m new,” the interaction proceeds in a somewhat different way. Mandy, who replies to Laura’s post in this thread in extract XIb below, treats Laura’s post differently; she takes it as the initiation of a topic for further conversation rather than a request for advice.

Extract XIb [Stommel & Lamerichs, 2014, p. 205]

  1. 1 03.02.2005 19:20 Re: I’m new – Mandy

  2. 2 Hello laura

  3. 3 I am also “just” 17.

  4. 4 In my case the thing is that I often have the desire to eat

  5. 5 something although I am totally not hungry.

  6. 6 Do you also have real binging attacks or not? (sorry, if you

  7. 7 don’t want to answer that question that’s fine

  8. 8 too. I am always so curious )

  9. 9 Would you like to tell a bit more about yourself?

  10. 10 Love,

  11. 11 Mandy

We will not focus on how the particularities of Laura’s initial post may contribute to this response by Mandy (but see Laura’s acknowledgment of her eating disordered past, in line 4). What we want to emphasize here instead is how in her reply to Laura, Mandy provides a description of a similar experience, or a second story (Arminen, 2004). The way it is formatted enables her to enact solidarity without providing advice (“In my case the thing is that I often have the desire to eat something although I am totally not hungry,” lines 4–5). She also presents an invitation to Laura to say a little more about herself. Laura’s subsequent response (not shown here) displays appreciation and expressed a wish to continue communicating with Mandy (“I would be pleased to receive another response from you”).

The analysis of these two threads that allowed us to examine the opening posts, the subsequent replies by Angela and Mandy, and the uptake by Laura that followed these replies, shows that participants in OSGs find ways to continue the interaction in different directions. It also shows, as we have seen before, that what may count as “advice” strongly differs and may also be taken up in different ways. Responses that contain “advice” may be subtly formatted in ways that maintain room for ongoing interaction rather than an early closure of a thread.

We may also begin to see how the different implications of providing some types of advice, as they can constitute an interactionally risky undertaking. Here, Angela’s reply to Laura “implied a demand on the new user to look more critically at her situation and to actively prevent the development of bulimia” (Stommel & Lamerichs, 2014, p. 207). In displaying advice resistance in extract Xc, Laura treats it as advice that she cannot apply. It might be because of this that the interaction in this thread ended after Laura’s message, but we can’t be sure.

The implications this might carry for newcomers like Laura who have had the courage to post in an OSG and see their opening post met with such explicit advice, may actually work to discourage newcomers to visit again. Or to put it differently, when participants engage in advice giving that is too far removed from showing troubles recipiency, the appeal of an OSG as a unique and safe space to interact with people “who are in the same boat” (cf. Sillence, 2013) may be threatened (also Stommel & Meijman, 2011). Participants may actually demonstrate this in their interaction, as the above examples illustrate.

Online Counseling as Interaction: Empirical Examples

In this section we discuss the second major field of online mental health talk addressed in this article, next to OSGs. We present examples of research conducted on the interactional practices in online counseling. In the field of online counseling, various studies have examined the effects and experiences of counselors and clients using online counseling or telehealth (e.g., Beattie, Shaw, Kaur, & Kessler, 2009; Fukkink & Hermanns, 2009a, 2009b; King, Bambling, Reid, & Thomas, 2006; King et al., 2006; Yuen, Goetter, Herbert, & Forman, 2012).

Where a substantial amount of research has been undertaken on counseling and therapy in face-to-face interaction (e.g., Davis, 1984; Peräkylä, Antaki, Vehviläinen, & Leudar, 2008; Vehviläinen, 2003; Voutilainen, Peräkylä, & Ruusuvuori, 2011), CA(-related) studies of online counseling are less abundant, but increasing (e.g., Danby et al., 2009; Ekberg et al., 2013; Ekberg, Barnes, Kessler, Malpass, & Shaw, 2016; Ekberg, Shaw, Kessler, Malpass, & Barnes, 2016; Harris et al., 2012; Stommel & Van der Houwen, 2013, 2014, 2015; Stommel & Te Molder, 2015, 2016; Thurnherr, von Rohr, & Locher, 2016).

These studies are all more or less driven by the question whether counseling online works as well as face-to-face counseling and more specifically what the online variants look like. For example, by addressing how clients phrase issues and questions, how counselors respond to these messages and also how the interaction between clients and counselors is linked to the communication medium. They resulted in detailed insight on issues such as how counselors and clients interactionally manage the counseling relationship (Ekberg et al., 2013, Stommel, 2012), explore the clients’ emotions (Ekberg, Shaw, et al., 2016), ask for and provide advice (Thurnherr et al., 2016), propose a different medium (Harris et al., 2012) or discuss the particularities and appropriateness of the medium (Stommel & Van der Houwen, 2014). Rather than attempting to sketch a complete overview of studies of online counseling, this section focuses on one key practice in the interaction between counselors and clients in online counseling. This allows us to showcase what an interactional approach to online counseling talk looks like.

In doing so, we concentrate on chat (a frequently used medium for counseling) and on the ways in which counselors formulate clients’ problems and what social actions are accomplished when doing so. We will discuss two examples from different chat counseling settings in depth.

How Counselors Use Formulations in Chat Counseling

Counselors do their work by accomplishing particular social actions such as asking questions, complimenting, and summarizing. Some of these actions may be part of a larger interactional project called a “formulation.” In CA, the term is used to refer to turns at talk “which characterize states of affairs already described or negotiated (in whole or in part) in the preceding talk” (Heritage & Watson, 1979, p. 126). In counseling, formulating is a way of presenting an alternative, counseling relevant view of the client’s account (see Antaki, 2008), and it is because of this that it is considered an important interactional instrument for counselors. For instance, a formulation can steer a client’s account into a modified version of that account with a focus on feelings/emotions (“hmm, yes a kind of feeling that you are being blown over?,” Stommel & Van der Houwen, 2013). At the same time, formulating another person’s personal concern is an interactionally sensitive thing to do as it is related to the ownership of experiences and feelings (Heritage, 2011).

Formulations are not only used in spoken forms of counseling, but they also, and quite often, occur in online chat counseling (Danby et al., 2009; Stommel, 2016; Stommel & Van der Houwen, 2013). In chat counseling formulations are generally used to organize the interaction, for example, to link the current session to a previous one, to summarize and close the session, or to solve misunderstandings in case of an apparent ambiguity or contradiction. In the latter case, the counselor may formulate an interpretation of what the client posted (Stommel & Van der Houwen, 2013).

Some of the uses of formulations in online counseling seem to solve particular interactional challenges to do with the medium of chat, but their use is not confined to the medium: parallels with face-to-face settings have also been drawn (see Sliedrecht, 2014 for a discussion of the interactional uses of formulations in different institutional settings; also Sliedrecht, Houwen, & Schasfoort, 2016).

We want to highlight a number of recurrent uses of formulations in chat counseling. First, formulations are employed to clarify an ambiguity in the immediately prior utterance as the means to deal with a discontinuity or misunderstanding in the chat interaction (also called “disrupted turn adjacency,” see Herring, 1999; cf. Nilsen & Mäkitalo, 2010; Schönfeldt & Golato, 2003).

Despite their capacity to solve misunderstandings, counselors’ formulations in chat counseling have also been found to be sequentially ambiguous: “it may not be as evident [as in spoken counselling] whether the formulation is being used to summarize the client’s problem, to prompt a continuation, or to invite the client to offer their perspective on the problem” (Danby et al., 2009, p. 109). The lack of intonation seems to explain this ambiguity (but see Sliedrecht, 2014 on the possible ambiguous nature of formulations as an interactional resource in face-to-face settings as well). In an attempt to deal with the lack of prosodic cue in chat counseling, formulations are extended with a question mark, for instance, “So meditating does you good?” (Stommel & Van der Houwen, 2013). The question mark disambiguates how to respond as it is shown to be a strong signal for the client to respond with a confirmation (Stommel & Van der Houwen, 2013).

Formulations in online chat counseling are also used to display active listening and/or to achieve a transformation in a direction relevant to the counseling goal or the particular counseling type (Danby et al., 2009; see also Antaki, 2008 for how formulations may be employed for doing similar interactional work in face-to-face psychotherapy).

We now turn to examples of formulations in online counseling that are related to specific counseling goals: exploring the question (section “How Counselors Use Formulations in Chat Counseling”) and articulating a solution-focused view on the issue (section “Formulations to Articulate a Particular View on the Problem”).

Formulations to Explore the Client’s Question

In chat counseling where clients can ask questions having to do with alcohol or drugs, formulations appear to be used to explore early in the session if the client is searching for a factual answer or for psychological help or support. See, for example, Excerpt XII, which includes two formulations from the counselor, transmitted in multiple chat posts. The formulations both convey an understanding of the essence of the client’s problem: having difficulties in restricting unhealthy behavior such as overeating, smoking, drinking, but the second one (lines 29–31) also articulates the emotional aspects of the problem.

Excerpt XII (Chat 9)

Name: Cl. Age: 32. Female. City: X. Question: Advice about self-restraint. Counsellor: Co. Next to the line numbers is the time stamp provided for the chat entries during this session.









Welkom bij de chat

Welcome to the chat




Waar kan ik je mee helpen?

How can I help you?





ik ben even aan het nadenken hoe ik het precies moet verwoorden

I am just thinking how to word it exactly




ik heb moeite af en toe met matigen

I have trouble restraining now and then




niet alleen met drank, maar ook bijvoorbeeld met eten

not just with alcohol, but also for instance with food





ik geloof dat ik af en toe de neiging heb om me te


I believe I have a tendency now and then to “misbehave”




en ik merk dat ik vaak overspring van het ene gedrag naar het andere als ik probeer dat te minderen

and I realise that I frequently jump from one behavior to the other once I try to lessen that






is het dan handiger om met één ding tegelijk te beginnen

of kan ik beter voor de allestegelijk aanpak


is it then more useful to start with one thing at a time or should I better start with the allatonce approach?




ik weet niet of ik het zo helder uitleg

I am not sure if my explanation is clear



Co: →

Ik denk dat ik snap wat je bedoelt.

I think I know what you mean.




Co: →

Je hebt moeite met de maat houden, zoals je aan het

begin zei.

You have trouble doing things in moderation, like you said in the beginning.




Co: →

En dat is niet alleen met alcohol, maar ook met

andere dingen.

And that is not just with alcohol, but also with other things.






als ik gezond ga eten ga ik opeens een fles wijn leegdrinken, als ik weinig drink ga ik opeens roken, als ik daarmee stop eet ik mijn koelkast leeg

when I eat healthy I suddenly empty a bottle of wine, when I drink little I suddenly start to smoke, when I quit with that I eat everything from my fridge until its empty









en dat is onhandig

and that is unpractical




Dat kan ik me voorstellen.

I can imagine that.





en ik doe dat niet elke dag natuurlijk maar wel als ik gestresst ben of me verveel

and of course I don’t do that every day but will when I’m stressed or when I’m bored



Co: →

Ok, dus er zijn “special” momenten

Ok, so there are “special” moments




Co: →

Uit verveling ga je iets doen, maar ook asl je je gespannen voelt.

Out of boredom you do something, but also when you feel stressed.




met name die twee ja

mainly those two yes





Zou je willen leren hoe je met je emoties om moet


Would you like to learn how to deal with your emotions?

After the opening greetings, the counselor invites the client to present her reason for joining the chat (line 3). In response, the client states the difficulty of phrasing her problem accurately (lines 4–5), which prepares for a lengthy problem presentation, including the mention of a lack of self-restraint concerning alcohol, food, and cigarettes and “jumping behavior” (lines 6–11). She works toward a question for advice: “is it then more useful to start with one thing at a time or should I better start with the allatonce approach?” (lines 12–14; see also Morrow, 2006, on alternative questions in OSGs). As the counselor has not posted anything to display “listening” yet (cf. Danby et al., 2009), the client evaluates her previous posts as potentially unclear. The counselor then posts a response that formulates the client’s problem (lines 16–20), indirectly ensuring the client’s account was insightful. The client first confirms the formulation indirectly by exemplifying her jumping behavior from drinking, to smoking, to eating excessively (lines 21–23) and then explicitly (“exactly,” line 24). Thus, mutual understanding of the problem has been achieved and now an answer to the client’s question in lines 12–14 could become relevant. However, this question is not returned to; instead, “talking” about the problem has become the main activity: The client evaluates her problem with an understatement (“and that is unpractical,” line 25) and specifies when it occurs (lines 26–27). Now, the counselor uses a second formulation highlighting the moments the problem occurs (“special moments”) and framing them in terms of emotions (boredom, stress). In design (“ok so”) the formulation works toward a conclusion about the core of the problem. The client’s confirms the formulation (line 32), which leads the counselor to propose an emotion-focused question as the reason for the chat session (cf. Danby et al., 2009): “Would you like to learn how to deal with your emotions?” (lines 33–34). This question is derived from the co-constructed problem description and is notably different from the original question in lines 12–14 (advice about restricting behavior).

To conclude, formulations of the problem by the counselor in the opening phase of the session both clarifies and co-construct the client’s problem. Formulations may also be used to highlight the emotional aspects of the problem, which leads up to emotion-focused re-specification of the client’s question. This shows how formulating the problem may enhance the “depth” of the counseling from more practical advice to advice on emotional issues (cf. Madill, 2015). Hence, formulations are used as a strategy to deal with the professional dilemma of chat counselors whether to provide an answer immediately or to explore if there is a more complex, emotional question “at the root” of the initial request for information.

Formulations to Articulate a Particular View on the Problem

Formulations in chat counseling can also carry an imprint of a particular counseling approach (e.g., cognitive behavioral therapy, solution-focused therapy) in their use and specifically their design. An example of an “optimistic” formulation that can be associated with solution-focused therapy is presented below in extract XIII.

Extract XIII [Stommel & Van der Houwen, 2013, n.p.]





okay fine, is there maybe a topic that you would like to discuss with me?










well i am quite complicated and just about everything bothers me and often ramble on and coincidentally i just meditated and have less a feeling that i feel bad








So meditating does you good?




well actually not really




sometimes it does sometimes it doesn’t





Note: *unrelated utterance, referring to previous topic.

In extract XIII, the client describes his problem (lines 3–5) and makes a distinction between his problems more generally and his actual feelings during the counseling session. In other words, the client has presented his current state as relatively positive thanks to meditation (“i just meditated and have less a feeling that i feel bad”). In line 7, the coach formulates this positive aspect: “So meditating does you good?” The formulation ignores the negative general state of the client (“often ramble on”). Therefore, this formulation is a so-called optimistic formulation, comparable to “optimistic questions” (MacMartin, 2008). By formulating it, she invites the client to confirm that meditating not just coincides with “feeling less bad” but actually can be considered the cause of it. Thus, a potential solution or coping strategy to the client’s depressive feelings is put on record.

However, the client disconfirms the formulation with extensive hedging (“well,” “actually,” “not really”) (line 8), denying meditation is an effective coping strategy. He continues to nuance this denial by explaining that meditating merely has an occasional, that is, partial or temporary positive effect. Then, the coach uses an acknowledgment token (“hmhm”), aligning as a recipient and inviting a more extensive account (Jefferson, 1985). By displaying resistance to the optimistic formulation, the client denies his problems could be easily solved. In other words, the optimistic formulation was launched as a therapeutically relevant pathway (a discussion of what works for the client), but in this case, it did not work out that way.

In sum, online talk about mental health in a counseling setting centers on discussing the client’s problem. In these discussions, counselors formulate the problem in specific counseling directions, such as focusing on emotions or on the positive aspects. Client responses indicate the precariousness of these formulations, despite anonymity and what is generally described as the low-threshold character related to the online nature of the interaction.

After we have presented some key empirical examples to illustrate the interactional approach to online mental health talk, we now turn to a discussion of the most important methodological issues that are discussed in the field of CA(-related) approaches to online mental health talk.

Methodological Issues

A number of methodological issues are recurrently touched upon in CA(-related) studies on online talk and Computer-Mediated Discourse Analysis (CMDA). We will discuss their relevance for studying OSGs and online counseling, knowing that the data of most of the studies that have focused on the interactional properties of online mental health talk in OSGs and chat counseling still mainly consist of (asynchronous) text-based forum contributions and (synchronous) chat posts. We think that a discussion of methodological issues is significant given the continuous development of social network and mobile technologies (WhatsApp communities, for example) and the ways in which OSGs are increasingly embedded on larger platforms hosted by health organizations. Oftentimes, these platforms, combine text-based discussion groups with other modalities such as profile pages with avatars. Registered members may also have access to blogging functionalities and can create elaborate member’s pages that include text and images. These technological developments clearly have methodological consequences: not only for data collection but also when we want to study how identity construction takes place with the help of these different modalities (see also Hunt, 2015, who explored the affordances posed by Facebook).

To a large extent, the discussion about methodological issues represents older debates around the nature of CMC as a “new” medium (Lievrouw, 2004) in which the dichotomous view is adopted that online and spoken interaction are a contrastive pair. Recently, attempts have been made to review the methodological issues that have arisen over time. This work provides some directions for thinking (see Herring, 2012; also Giles et al., 2015; Meredith & Potter, 2013).

In what comes next, we present four methodological issues that have recently been considered. They are presented as separate sections for the sake of clarity, but the arguments are in fact closely interlinked.

A Persistent Lack of Structured Methods

Already in 2004, when computer-mediated discourse analysis (CMDA) was first proposed as an approach to analyze online discourse and sought to incorporate the role of “technological variables” (e.g., channels of communication like text, audio, video; synchronicity) and “situational variables” (e.g., participant characteristic; purpose, tone), the need for structured methods (and theoretical frameworks) was acknowledged (see Herring, 2004b for a full account). Today, it is still put forward as an important methodological concern in the field (Herring, 2012).

Possible reasons for why few attempts to develop these structured methods were undertaken might be that within the disciplines that considered discourse analysis their core business (e.g., linguistics, psychology, and sociology), scholars refrained from theorizing on online discourse because it was considered a “popular” cultural phenomenon (Herring, 2004b). Also today online communities are frequently considered “weak simulacra of offline communities” (Giles et al., 2015, p. 47). As a consequence researchers that employed the analytic principles of CA to study online talk were considered not doing real CA” (Giles et al., 2015, p. 45; also see Paulus, Warren, & Lester, 2016). Hence, this attitude adopted within the broader scholarly community may have prevented researchers with an interest in studying online interaction from disseminating their work as widely as they would have wanted, or hindered an atmosphere in which systematic and ongoing discussions with fellow researchers on methodological issues could take place.

Apart from noting a lack of structured methods, the question is simultaneously posed whether CA(-related) approaches and CMDA form a sufficient basis to examine online talk or whether they are in need of adaptation (Giles et al., 2015; also Herring, 2012). This brings us to the second methodological issue.

Do Existing Methodologies Suffice to Study Online Interaction?

An answer to this question has been provided in two ways mainly: While a disaffirmative answer poses that new methods or adaptations of existing methods are needed, an affirmative answer poses that existing conceptual tools are adequate for the study of online talk. We explore both possible answers to this question here in greater detail.

The former line of arguing basically puts forward the claim that the conceptual apparatus that underlies interactional approaches to online talk (initially developed to analyze face-to-face interaction and telephone conversation), needs some form of adaptation to better cater for collecting and analyzing what are considered “new” types of online data. As a result, a more “natively digital” research method is proposed (Giles et al., 2015). It is noted that the new research method should embrace the central tenets in CA that are employed to study spoken interaction, but also include adaptations that render the method more specifically applicable to the study of different environments (Twitter, as well as Instagram, or a discussion board).

Transcription is the first topic that is addressed in this respect and it highlights the question what counts as transcribable data in different technological environments? Where this might be relatively easy to establish for text-based environments, some research projects (e.g., those extracting data from visual media) are faced with technical challenges to do so. Apart from the modalities available, there are also some other challenges to do with what counts as data and what needs therefore to be transcribed in the first place (see section “Determining the Influence of Technological Affordances in Studies of Online Talk”).

A second central tenet concerns the sequential organization of online talk. We have already seen how sequences might operate differently in OSGs as a result of disrupted or “virtual” adjacency or delays. We have demonstrated also how this carries implications for adopting the next turn proof procedure, which is crucial for providing action descriptions in members’ terms (see section “Applying an Interactional Approach to an Illustrative Data Fragment”). Hence, we may ask whether the central notion in ethnomethodology and CA that participants locally construct interactional coherence offers us enough analytical leverage to deal with the issue of sequential organization (Reed, 2001; see also Herring, 2012, for a discussion).

Continuous systematic studies that address this question are called for. It has however, also been pointed out that matters of sequential orientation are virtually impossible to grasp by using other research methods, as they are “performable skills, but not necessarily expressible knowledge” (Antaki et al., 2005, p. 128; italics added). Or to put it differently, as a member’s practice, creating “coherence” (i.e., demonstrating that one is responding to a particular forum member, or to a particular part of a post) constitutes something that people do in online talk. As a result, it would most likely not be something that participants are able to describe or verbalize in any way better in interviews or surveys. Thus, CA’s analytical procedures may be “as good as it gets” (Meredith & Potter, 2013).

The second, affirmative answer to the above question of whether existing methods suffice to study online interaction adopts a more “reluctant” view on whether methodological changes are necessary, and stems from the field of CMDA (Herring, 2012). It addresses the question of whether empirical studies can still be undertaken with the same conceptual tools that were developed in computer-mediated discourse analysis some time ago; most notably the distinction between technological and situational variables that can be used to study online talk (see Herring, 2004a; also Locher, 2012, for a recent application). This “answer” from the field CMDA research does not so much investigate the implications for the field of CA(-related) approaches but it has resulted in a set of broader but possible very relevant observations nevertheless. The observations have resulted in a three-part categorization scheme that builds on the two sets of variables that were mentioned above and distinguishes between “familiar,” “reconfigured,” and “emergent” online discourse phenomena (Herring, 2012). More importantly, one of the aims of proposing this scheme is to challenge the assumption that all discourse phenomena that appear in new online environments are in fact new (and are because of that, in need of new methods of study).

Two thoughts that are put forward in this classification scheme in CMDA are particularly relevant for this article. The first is that the field of studying online discourse has a tendency toward “ahistoricity” (i.e., it overlooks the fact that new phenomena resemble already existing “relevant antecedents,” 2012, p. 10; also Finn & Light, 2016). The second is that because some discourse phenomena seem different on the surface, they may be mistaken for being new (e.g., retweets in Twitter that may seem new but are in fact comparable with quoting practices in e-mail; Herring, 2012, p. 12). As a result of focusing on online discourse phenomena as novelties, a “loss of comparative insight” might occur (2012, p. 15).

Studies that adopt CA(-related) approaches are also vulnerable to a loss of such comparative insight. These insights might be hampered precisely because the studies seek to implicitly or explicitly juxtapose online and spoken interaction. Or because it starts from the view that because we are dealing with a new environment, this also implies that people use it to accomplish “new” (or different) actions (see also section “Determining the Influence of Technological Affordances in Studies of Online Talk”).

A recent study that analyzed how self-repair in Facebook chat is being done, as opposed to spoken interaction might illustrate this (Meredith & Stokoe, 2013). Although concluding that the difference between online and offline communication is premature, it is implicitly used as a frame of reference. Self-repair, a phenomenon well researched in CA (Schegloff, Jefferson, & Sacks, 1977), was found to manifest itself in Facebook chat in two ways: online and offline. People may engage in self-repair online, when a chat entry is posted and “seeable” by all parties taking part in the interaction. And people may engage in self-repair offline, during message construction, for example, by deleting, or rephrasing parts of their contribution/these are repairs not “seeable” by other parties. When we examine spoken interaction, we can use the production details and receipt by another speaker at the same time, to examine how repair occurs. It turns out that if we want to examine “offline” practices of self-repair in Facebook chat we also need production data (screen capture data) of how those chats are constructed and repaired and how it matters for users what happens during live production of the chats. In terms of data collection this may pose far from a new demand; production data was already used in a very early study on chat interaction (Garcia & Jacobs, 1999). Examining offline production practices, as in the case of offline self-repair in chat (i.e., what people do when they edit their messages) does beg the question whether such actions are best studied with an interactional approach or with other methods. To put it differently, it makes it relevant to engage in a methodological discussion on how to apply important analytical principles such as the next turn proof procedure in an environment of self-repair that exists outside of an interactional context (see also Meredith & Potter, 2013, for a discussion of different types of screen capture data, to study chat interaction).

Adhering to a framework in which spoken interaction is often implicitly (or more explicitly) considered a frame of reference or even the golden standard to study online talk, brings us to the third methodological point.

From a Dominant Dichotomous View to More “Nested Analyses”

The continuous development of information technologies and the ways in which they are becoming integral parts of peoples’ social lives, has led to the critique that a research focus on online talk alone fails to reflect that entwined nature. It has resulted in a plea for research on “situated” or “nested” media use (Baym, 2011), which addresses how offline and online modes of talking are increasingly interwoven (e.g., having dinner with a friend in a restaurant while tweeting about your food with your smartphone, checking in on the restaurant’s Facebook site, and placing a photo of your main course on Instagram).

While there might be a need for this type of “nested analysis” when it comes to a study of “food talk,” like in the example above, it is arguable whether it is as important for studying the interactional practices of those with mental health concerns. This might be because online talk about mental health will not likely be taken offline in the same way as “food talk” or via other communication technologies. For people with a mental illness, particular online spaces may very well remain the “safe” environments where they can communicate with their peers, and precisely so because they purposefully prevent any crossovers to other practices or platforms (see Giles et al., 2015, for a discussion).

A plea for more nested analysis when it comes to online counseling studies may look a bit different. For online counseling practitioners there is an increasing demand to work in both online and offline settings and to combine different Internet-based ways of communicating with their clients (chat, e-mail, and increasingly also video-mediated sessions) while they also see their clients in face-to-face sessions. These “blended” working practices make a nested approach interesting. Such a nested approach could address how the different modes of communication are chosen, and for which purposes. It enables a further exploration of how professionals and clients sometimes flexibly move between these different media and how this may result in particular interactional challenges (Harris et al., 2012; Stommel, 2012; see also Stommel & Te Molder, 2016, for a discussion). Such a nested analysis can then also consist of comparative studies, which might further theory building.

Determining the Influence of Technological Affordances in Studies of Online Talk

Communicating via Twitter, Facebook chat, or a discussion forum differs in the types of structural features available through these technologies. How people orient to these features in their interaction may be easily overemphasized (see Herring, 2012, for a discussion). How the importance of technological features matters, and how researchers have attempted to structurally examine their role is the fourth methodological concern for CA researchers we want to address.

Apart from the critique that CA-(related) approaches tend to overlook relevant antecedents, it is still a methodologically worthwhile project to explore how technological mediation relates to online talk, without assuming a priori that it matters in certain pre-specified ways. An example might illustrate some of the complexities involved in coming to terms with such a project, derived from a study that has explored practices by which a professional service offers online advice (cf. Locher, 2012).

A close examination of the interaction on a professional advice column on the Internet, called “Lucy Answers,” where participants can ask questions and receive an answer from a professional called Lucy, showed that advice was largely constructed in ways that toned down the face-threatening character of the recommendations, by using indirect, hedged, or “qualified” types of recommendations (cf. Morrow, 2006, on similar practices in an OSG). These practices can be systematically explored, based on the set of technological and situational variables identified by Herring (2004b). The main conclusion reached in this study of “Lucy Answers” was that the purpose of the professional advice column (reflected in the situational variable “purpose”) has a stronger impact on the choice of language than the fact it is computer mediated: The mission of the health program to which the advice column belongs, promotes non-directiveness in providing advice. The public dimension of the advice column is also considered important, and this is shown in how “Lucy Answers” addressed problems that are both relevant to the general readership and the individual seeking advice (see Meredith & Potter, 2013, on recipient design). Thus, the study showed how the relevance of technological factors may not be taken for granted.

The same argument that stresses a more cautious view on whether technology matters in a straightforward way for the study of actions in online talk, is demonstrated by examining a new technological feature on Facebook (Koteyko & Hunt, 2016). Here, Facebook users with diabetes mellitus use “status updates” to address their illness experience. The status update, which offers an opportunity for micro-blogging (i.e., to post a short self-report on what one is doing) is shown to enable people with diabetes mellitus to construct relevant illness identities and narrate about ways of managing their illness on a day-to-day basis (Koteyko & Hunt, 2016). Similar to what Herring puts forward, these new technological features are used to accomplish similar actions in terms of identity work as we might observe in OSGs (Hunt, 2015).

What is useful in coming to terms with the relative weight of technological variables is that the conceptual models such as the ones proposed by Herring (2004b, 2012) allow us to unpack the relevant characteristics of communication technology in a structured way. Coupled with an analysis that builds on the central CA tenets of sequential analysis and recipient design, this can be a most useful conceptual scheme for robust, basic descriptive research.

We are beginning to see that these technological factors do not operate in isolation from situational factors and maybe this integrated view most closely resembles the practice of online talk in different online environments. Thus, the particular importance or influence of (novel) technological affordances may very well be studied best in a case-study approach. This was shown for example in a study of the interactional consequences of working with clients providing prestructured information in chat counseling, and the adverse effects this had on the start of the interaction (Stommel & Te Molder, 2016). Based on descriptive single case studies, we can show how particular technological affordances are oriented to locally, in interaction; be that a chat environment, Facebook, Twitter, or mobile technologies such as WhatsApp and then broaden the empirical base for our findings.

Avenues for Future Research

Future research in the field of online talk about mental health may develop in different areas. We want to highlight four possible avenues we think are fruitful routes for future research. We also point to a more general remark about the types of research that are desirable; to conduct basic descriptive explorations but also to strive for studies that inform theory.

Our suggestions for future research are aimed at the interactional practices in OSGs and online counseling, while taking into account new online environments where people are already starting to discuss matters to do with mental health (cf. Hunt, 2015, for an example of how people manage diabetes with the help of the modalities their Facebook page provides). These are the areas in which basic descriptive studies are called for.

A first area has to do with opening up the disciplinary boundaries between the field of CA and CA-related approaches and to become more aware of research conducted in other areas. It seems that CA and CA-related work sometimes overlooks relevant work in CMDA for example (but see Giles et al., 2015). Given some of the complexities in examining online talk as interaction, it is worthwhile to systematically attempt to combine insights from CA and category membership analysis (MCA) and work conducted on CMDA (Stokoe, 2012). Such combined approaches may be able to address not only the complicated matters of sequential coherence in evolving discussion threads but may also further the analytic focus in much of the research carried out to examine identity construction and the related category-bound work that participants in OSGs (but also in online counseling) heavily engage in (see Stommel & Koole, 2010). The work conducted on an interactional practice such as advice giving may count as a relevant example (see Locher, 2012, for a discussion).

Second, more basic descriptive studies are needed that move beyond what happens in newcomers’ opening posts and their first reply (cf. Paulus & Varga, 2015; Stommel & Lamerichs, 2014). The same holds true for the much smaller field of online counseling. In the latter field, studies have just begun to explore how the interaction unfolds when the client’s problem presentation is discussed over a series of e-mails (Lamerichs & Stommel, 2016). Although there are little “missing replies” in these one-on-one encounters, the analysis of these relatively large e-mails still poses some challenges from an interactional point of view: They almost represent a practice of exchanging letters and with each e-mail multiple actions are fulfilled in response to a previous message. It would be very interesting to collect production data to capture how editing of turns takes place here too. Exploring the different actions that participants attend to in message construction would be informative for the field of online counseling. At the same time data collection would probably be very complex, not in the least because of ethical implications.

Third, there is also room for comparative studies that shed more light on practices such as advice giving in chat counseling and OSGs, and how people construct “expert” identities on their disease in these different technological environments (see also Locher, 2012, for a discussion).

Fourth, whereas interactional studies of online counseling have until now mostly explored text-based interaction, developments such as video-mediated counseling (cf. Pappas & Seale, 2010) make more integrated and also multimodal analyses necessary.

To conclude, given the methodological issues addressed in the section “Some Methodological Issues,” pertaining to the challenges of collecting data (i.e., production data, visual data), and analyzing them as interactional data, it would be important for the field to keep pushing for more than basic descriptive research alone. There is a clear need to conduct more comparative studies across technological environments and generally strive for studies that seek to inform theory building.

Primary Sources

Online discussion forums or OSGs are to some extent publicly available on the Internet and constitute an enormous source of data that is often easy to collect. For researchers who study online counseling settings there are often other steps that need to be taken. We want to say a few words here about what studying both types of online data means from an ethical point of view.

Ethical Considerations

In their working paper on ethical considerations (2012), the Association for Internet Researchers (AoIR) stresses the need for new models of ethical evaluation and consideration. They emphasize the need for such systematic attention given the wealth and the variety of research that is conducted on the Internet in which data is collected online. The ethical considerations may even hold a greater importance when it concerns collecting naturalistic data on online mental health talk.

What is considered first in the six key guiding principles proposed in the report is an assessment of the vulnerability of the community/author/participant involved. This estimate should correspond with the responsibility of the researcher to ensure their protection. Moreover, it is proposed that ethical considerations should be addressed in all stages of the research process, starting with the planning, continuing throughout the research, and including publication and dissemination of the research findings at the end of the project (i.e., a “process approach”). The AoIR thus clearly counters the view that ethical considerations are an “obligatory hurdle” that only needs to be taken at the start of a research project. A third key principle is to consult as many sources as possible to collect insights that are informative for the ethical decision making process. These sources might be fellow researchers; published scholarship both within one’s discipline, and also outside the own disciplinary boundaries, people participating in or familiar with the context/site that is being studied, and also research review boards and ethics guidelines (2012, pp. 4–5). Importantly, the working committee offers a warning about the complexity of ethical decision making, which includes lots of gray areas. Moreover, the proposed process approach requires continuous and often dilemmatic judgments about ethical concepts such as harm, vulnerability, respect for persons, and beneficence in each stage of the research.

One of the main tensions discussed in the working paper, which also applies to online mental health talk in OSGs, is the public/private division when it comes to definitions and expectations of privacy. The discussion is a little different for the “restricted” environment of online counseling (we will say more about this context for research), but it is highly relevant for OSG research in at least three ways: (1) people may interact in (semi-)public online spaces such as OSGs while maintaining strong perceptions or expectations about the privacy of their interactions; (2) people might be aware that the content of their interactions is largely public but might have strong ideas about the way in which these contents should (or should not) be put to use by third parties; and (3) Internet search tools can make interactions accessible to a wider audience while this was not intended (see Antaki et al., 2005, for a discussion on how to deal with retrievability of data for OSG fragments).

It is clear that these considerations go further than only considering the practical reality of whether a forum is accessible for the public as a whole or whether it has some restrictive layers in place (i.e., registering as a member) as some health platforms do. It also becomes apparent that the ethical decision making process when it comes to weighing the rights of subjects versus the weight of research benefits, is complicated when it involves people with mental health conditions. They are often highly “vulnerable” populations who experience taboos, shame, and stigma related to their condition. This might even be more the case when it concerns young people (see Aronsson, 2011, for a discussion). While far from homogenous, people who suffer from a mental condition often go online to communicate with peers where there is little opportunity to discuss these topics with friends and family. We have already seen how the unique opportunity of being able to do so online is often topicalized in their talk (Varga & Paulus, 2014) as well as how they seek normalization of their particular experiences (Giles & Newbold, 2013). These aspects should also be taken into account in the process of weighing research benefits.

A very useful and well-developed starting point for appropriate decision making is a checklist that contains specific questions for researchers to consider throughout their project (AoIR Working Committee on Ethics, 2012, pp. 8–11). It includes questions that pertain to: establishing the primary object of study (“what are the ethical expectations commonly associated with these types of data?”); data management and storage (“what are the potential ethical consequences of stripping data of personally identifiable information?”; but also “how might removal of selected information from a dataset distort it such that it no longer represents what it was intended to represent?”); the way findings are presented (“what immediate or future risk might occur by using exact-quoted material in published reports?”); and what the study’s potential harms or risks are (“would a mismatch between researcher and community/participant/author definitions of ‘harm’ or ‘vulnerability’ create an ethical dilemma?”). The list of question is extensive and goes beyond some of the points we have outlined above. It offers both novice and experienced researchers an invaluable set of well-thought-out pointers that are best considered before starting out on a research project (see also Buchanan, Markham, & Ess, 2010; Flicker, Haan, & Skinner, 2004; and Markham & Baym, 2009, for a discussion).

Although CA researchers might argue that they are not interested in participant demographics but in their language use, we must not forget that this distinction is not necessarily a relevant nor a recognizable division for participants whose contributions we study. Participants whose data we collect may also be unaware of what such methodological differences mean in practice. Or, more importantly often for them, whether it prevents examples of their talk to “show up at other places on the web.” And, as we have seen from the AoIR working paper (2012), it might also counter their perceived sense of privacy or the preferred way in which they would see their interactions reported on, let alone analyzed (see Pomerantz, 2012, for a discussion).

With the research examples discussed in this contribution, we can begin to see that CA(-related) approaches to online mental health talk, precisely because they work with naturalistic data often included as “exact quotes” in their articles, come along with important ethical considerations for researchers to pay attention to.

For online counseling the same guidelines hold true as for working with online data from OSGs. However, these data are often collected through collaborations with other health organizations, as access to the data would otherwise not be possible. CA research on online counseling are mostly secondary analyses of data collected for different research purposes (examining the effectiveness of online counseling as a health intervention, for example, see Oravec, 2009). As these data have their own consent procedures in place before they become available for a secondary analysis, researchers often work with consent regulations that were part of the research design as it was initially developed (see Stommel & Van der Houwen, 2013, for a discussion). Setting up a counseling service and making the data available for further research and secondary analysis means that researchers will have had to gain approval from the human subjects ethics review committee and seek informed consent. The checklist that was discussed above nevertheless poses important questions for CA researchers in the field of online counseling too, aside from the formal procedures for consent that were already taken. For example, participants’ sense of perceived privacy will likely be higher in an institutional setting in which a trusted relationship with a health professional is built. And also the sense of harm for participants (to life, career, reputation) is very much present when seeking help for addictive behavior or severe depression. While participants may not think this information is sensitive at the time consent is given, this might change over time and become relevant when studies that work with naturalistic data become available as published work.

Additional Resources

Further Reading

AoIR Ethics Working Committee (2012). Ethical decision-making and Internet research: Recommendations from the AoIR Ethics Committee. Chicago: AoIR.Find this resource:

    Baym, N. (2011). Personal connections in the digital age. Cambridge, UK: Polity Press.Find this resource:

      Edwards, E. (1997). Discourse and cognition. London: SAGE.Find this resource:

        Ekberg, S., Shaw, A. R. G., Kessler, D. S., Malpass, A., & Barnes, R. K. (2016). Orienting to emotion in computer-mediated cognitive behavioral therapy. Research on Language and Social Interaction, 49(4), 310–324.Find this resource:

          Giles, D. C., Stommel, W., Paulus, T., Lester, J. N., & Reed, D. (2015). The microanalysis of online data: Methodological developments. Discourse, Context & Media, 7, 45–51.Find this resource:

            Heritage, J., & Clayman, S. (2010). Talk in action: Interactions, identities and institutions. Hoboken, NJ: John Wiley & Sons.Find this resource:

              Heritage, J., & Robinson, J. (2006). Accounting for the visit: Giving reasons for seeking medical care. In J. Heritage & D. Maynard (Eds.), Communication in medical care: Interaction between primary physicians and patients (pp.48–85). Cambridge, UK: Cambridge University Press.Find this resource:

                Herring, S. C., Stein, D., & Virtanen, T. (Eds.). (2013). Handbook of pragmatics of computer-mediated communication. Berlin: Mouton De Gruyter.Find this resource:

                  Lester, J. N., & O’Reilly, M. (2016). The social construction of normality and pathology. In J. N. Lester & M. O’Reilly (Eds.), Palgrave handbook of adult mental health (pp. 1–19). London: Palgrave MacMillan.Find this resource:

                    Meredith, J., & Potter, J. (2013). Conversation analysis and electronic interactions: Methodological, analytic and technical considerations. In H. L. Lim & F. Sudweeks (Eds.), Innovative methods and technologies for electronic discourse analysis (pp. 370–393). Hershey, PA: IGI Global.Find this resource:

                      Paulus, T., Warren, A., & Lester, J. N. (2016). Applying conversation analysis methods to online talk: A literature review. Discourse, Context & Media. 12. Advance online publication.Find this resource:

                        Sacks, H. (1992). Lectures on conversation (Vols. I and II). Oxford: Blackwell.Find this resource:

                          Sidnell, J., & Stivers, T. (2013). The handbook of conversation analysis. Chicester, UK: Wiley-Blackwell.Find this resource:

                            Stommel, W., & Koole, T. (2010). The online support group as a community. A micro-analysis of the interaction with a new member. Discourse Studies, 12(3), 357–378.Find this resource:


                              Antaki, C. (2008). Formulations in psychotherapy. In A. Peräkylä, C. Antaki, S. Vehviläinen, & I. Leudar (Eds.), Conversation analysis and psychotherapy (pp. 26–42). Cambridge: Cambridge University Press.Find this resource:

                                Antaki, C., Ardévol, E., Núňez, F., & Vayreda, A. (2005). “For she who knows who she is”: Managing accountability in online forum messages. Journal of Computer-Mediated Communication, 11(1), 114–132.Find this resource:

                                  Antaki, C., Barnes, R., & Leudar, I. (2005). Diagnostic formulations in psychotherapy . Discourse Studies, 7(6), 627–647.Find this resource:

                                    Antaki, C., & Widdicombe, S. (Eds.). (1998). Identities in talk. London: SAGE.Find this resource:

                                      AoIR Ethics Working Committee. (2012). Ethical decision-making and Internet research: Recommendations from the AoIR Ethics Committee. Chicago: AoIR.Find this resource:

                                        Arminen, I. (2004). Second stories: The salience of interpersonal communication for mutual help in AlcoholicsAnonymous. Journal of Pragmatics, 36, 319–347.Find this resource:

                                          Aronsson, K. (2011). Bullying in context: Stories of bullying on an Internet discussion board. Children & Society, 25, 317–327.Find this resource:

                                            Baym, N. (2011). Personal connections in the digital age. Cambridge, UK: Polity Press.Find this resource:

                                              Beattie, A., Shaw, A., Kaur, S., & Kessler, D. (2009). Primary-care patients’ expectations and experiences of online cognitive behavioural therapy for depression: A qualitative study. Health Expectations, 12(1), 45–59.Find this resource:

                                                Benwell, B., & Stokoe, E. (2006). Discourse and identity. Edinburgh, UK: Edinburgh University Press.Find this resource:

                                                  Bou-Franch, P., Lorenzo-Dus, N., & Garcès-Conejos Blitvich, P. (2012). Social interaction in YouTube text-based polylogues: A study of coherence. Journal of Computer-Mediated Communication, 17, 501–521.Find this resource:

                                                    Buchanan, E., Markham, A., & Ess, C. (2010). Ethics and Internet research commons: Building a sustainable future. Association of Internet Researchers 11th annual workshop conference. Gotenburgh, Sweden.Find this resource:

                                                      Buttny, R., & Morris, G. H. (2001). Accounting. In W. P. Robinson & H. Giles (Eds.), The new handbook on language and social psychology (pp. 285–302). New York: John Wiley & Sons.Find this resource:

                                                        Danby, S., Butler, C. W., & Emmison, M. (2009). When “listeners can’t talk.” Comparing active listening in opening sequences of telephone and online counselling. Australian Journal of Communication, 36(3), 91–113.Find this resource:

                                                          Dancy, A. M., & Dooley, B. (2007). A clinical profile of participants in an online support group. European Eating Disorders Review, 15, 185–195.Find this resource:

                                                            Davis, K. (1984). The process of problem (re)formulation in psychotherapy. Sociology of Health and Illness, 8, 44–74.Find this resource:

                                                              Edwards, E. (1997). Discourse and cognition. London: SAGE.Find this resource:

                                                                Edwards, E., & Potter, J. (1992). Discursive psychology. London: SAGE.Find this resource:

                                                                  Ekberg, S., Barnes, R. K., Kessler, D. S., Malpass, A., & Shaw, A. R. G. (2013). Managing the therapeutic relationship in online cognitive behavioural therapy for depression: Therapists’ treatment of clients’ contributions. Language @ Internet, 10, article 4, urn:nbn:de:0009-0007-36986. Retrieved from this resource:

                                                                    Ekberg, S., Barnes, R. B., Kessler, D. S., Malpass, D. A., & Shaw, A. R. G. (2016). Managing clients’ expectations at the outset of online Cognitive Behavioural Therapy (CBT) for depression. Health Expectations, 19(3), 557–569.Find this resource:

                                                                      Ekberg, S., Shaw, A. R. G., Kessler, D. S., Malpass, A., & Barnes, R. K. (2016). Orienting to emotion in computer-mediated cognitive behavioral therapy. Research on Language and Social Interaction, 49(4), 310–324.Find this resource:

                                                                        Finn, M., & Light, B. (2016). Imagining the Internet. Information Communication & Society, 19(6), 709–714.Find this resource:

                                                                          Flicker, S., Haan, D., & Skinner, H. (2004). Ethical dilemmas in research on Internet communities. Qualitative Health Research, 14(1), 124–134.Find this resource:

                                                                            Fukkink, R., & Hermanns, J. (2009a). Children’s experiences with chat support and telephone support. Journal of Child Psychology and Psychiatry, 50, 759–766.Find this resource:

                                                                              Fukkink, R., & Hermanns, J. (2009b). Counseling children at a helpline: Chatting or calling? Journal of Community Psychology, 37, 939–948.Find this resource:

                                                                                Garcia, A. C., & Jacobs, J. R. (1999). The eyes of the beholder: Understanding the turn-taking system in quasi-synchronous computer-mediated communication. Research on Language and Social Interaction, 32(4), 337–367.Find this resource:

                                                                                  Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice Hall.Find this resource:

                                                                                    Gilat, I., & Kupferberg, I. (2010). A journey into the future: A computer-mediated conversation between a suicidal help seeker and a group of trained volunteers. Israel Studies in Language and Society, 2, 124–143.Find this resource:

                                                                                      Giles, D. C. (2006). Constructing identities in cyberspace: The case of eating disorders. British Journal of Social Psychology, 45, 463–477.Find this resource:

                                                                                        Giles, D. C., & Newbold, J. (2013). “Is this normal?” The role of category predicates in constructing mental illness online. Journal of Computer-Mediated Communication, 18(4), 476–490.Find this resource:

                                                                                          Giles, D. C., Stommel, W., Paulus, T., Lester, J. N., & Reed, D. (2015). The microanalysis of online data: Methodological developments. Discourse, Context & Media, 7, 45–51.Find this resource:

                                                                                            Green, D., & Kupferberg, I. (2000). Detailed and succinct self-portraits of addicts in broadcast stories. Discourse Studies, 2, 305–322.Find this resource:

                                                                                              Guise, J., Widdicombe, S., & McKinlay, A. (2007). “What is it like to have ME?”: The discursive construction of ME in computer-mediated communication and face-to-face interaction. Health, 11(1), 87–108.Find this resource:

                                                                                                Harris, J., Danby, S., Butler, C. W., & Emmison, M. (2012). Extending client-centered support: Counselors’ proposals to shift from e-mail to telephone counselling. Text & Talk, 32(1), 21–37.Find this resource:

                                                                                                  Heritage, J. (2011). Territories of knowledge, territories of experience: Empathic moments in interaction. In T. Stivers, L. Mondada, & J. Steensig (Eds.), The morality of knowledge in conversation (pp. 159–183). Cambridge, UK: Cambridge University Press.Find this resource:

                                                                                                    Heritage, J., & Clayman, S. (2010). Talk in action: Interactions, identities and institutions. Hoboken, NJ: John Wiley & Sons.Find this resource:

                                                                                                      Heritage, J., & Robinson, J. (2006). Accounting for the visit: Giving reasons for seeking medical care. In J. Heritage & D. Maynard (Eds.), Communication in medical care: Interaction between primary physicians and patients (pp. 48–85). Cambridge, U.K.: Cambridge University Press.Find this resource:

                                                                                                        Heritage, J., & Watson, D. R. (1979). Formulations as conversational objects. In G. Psathas (Ed.), Everyday language: Studies in ethnomethodology (pp. 123–162). New York: Irvington.Find this resource:

                                                                                                          Herring, S. (1999). Interactional coherence in CMC. Journal of Computer-Mediated Communication, 4(4). Available from this resource:

                                                                                                            Herring, S. (2004a). Computer-mediated discourse analysis. In S. A. Barab, R. Kling, & J. H. Gray (Eds.), Designing for virtual communities in the service of learning (pp. 338–376). Cambridge, UK: Cambridge University Press.Find this resource:

                                                                                                              Herring, S. (2004b). Online communication: Through the lens of discourse. In M. Consalvo, N. Baym, J. Hunsinger, K. B. Jensen, J. Logie, M. Murero, & L. Shade (Eds.), Internet research annual (Vol. 1, pp. 231–240). Selected papers from the Association of Internet Researchers Conferences 2000–2002. New York: Peter LangFind this resource:

                                                                                                                Herring, S. C. (2012). Discourse in Web 2.0: Familiar, reconfigured, and emergent. In D. Tannen & A. M. Tester (Eds.), Georgetown University round table on languages and linguistics 2011: Discourse 2.0: Language and new media (pp. 1–25). Washington, DC: Georgetown University PressFind this resource:

                                                                                                                  Herring, S. C., Stein, D., & Virtanen, T. (Eds.). (2013). Introduction to the pragmatics of computer-mediated communication. In Handbook of pragmatics of computer-mediated communication (pp. 3–31). Berlin: Mouton De Gruyter.Find this resource:

                                                                                                                    Hoek, H. W., & Van Hoeken, D. (2003). Review of the prevalence and incidence of eating disorders. International Journal of Eating Disorders, 34, 383–396.Find this resource:

                                                                                                                      Horne, J., & Wiggins, S. (2009). Doing being “on the edge”: Managing the dilemma of being authentically suicidal in an online forum. Sociology of Health and Illness, 31(2), 170–184.Find this resource:

                                                                                                                        Horton-Salway, M. (2001). Narrative identities and the management of personal accountability in talk about ME: A discursive approach to illness narrative. Journal of Health Psychology, 6(2), 247–259.Find this resource:

                                                                                                                          Houston, T. K., Cooper, L. A., & Ford, D. E. (2002). Internet support groups for depression: A 1-year prospective cohort study. American Journal of Psychiatry, 159, 2062–2068.Find this resource:

                                                                                                                            Hunt, D. (2015). The many faces of diabetes: A critical multimodal analysis of diabetes pages on Facebook. Language & Communication, 43, 72–86.Find this resource:

                                                                                                                              Jefferson, G. (1985). Notes on a systematic deployment of the acknowledgement tokens “Yeah” and “Mm hm.” Papers in Linguistics, 17(2), 197–216.Find this resource:

                                                                                                                                Jefferson, G., & Lee, J. R. E. (1981). The rejection of advice: Managing the problematic convergence of a “troubles-telling” and a “service encounter.” Journal of Pragmatics, 5, 399–422.Find this resource:

                                                                                                                                  King, R., Bambling, M., Lloyd, C., Gomurra, R., Smith, S., Reid, W., & Wegner, K. (2006). Online counseling: The motives and experiences of young people who choose the Internet instead of face to face or telephone counseling. Counseling and Psychotherapy Research, 6(3), 169–174.Find this resource:

                                                                                                                                    King, R., Bambling, M., Reid, W., & Thomas, I. (2006). Telephone and online counseling for young people: A naturalistic comparison of session outcome, session impact and therapeutic alliance. Counseling and Psychotherapy Research, 6(3), 175–181.Find this resource:

                                                                                                                                      Koteyko, N., & Hunt, D. (2016). Performing health identities on social media: An online observation of Facebook profiles. Discourse, Context and Media, 12, 59–67.Find this resource:

                                                                                                                                        Kupferberg, I., & Gilat, I. (2012). The discursive self-construction of suicidal help seekers in computer-mediated discourse. Communication & Medicine, 9, 1–13.Find this resource:

                                                                                                                                          Kupferberg, I., & Green, D. (2005). Troubles talk: Metaphorical negotiation in problem discourse. Berlin: Mouton de Gruyter.Find this resource:

                                                                                                                                            Lamerichs, J. (2003). Discourse of support. Exploring online discussions on depression. (PhD diss., Wageningen University & Research Centre). Retrieved from this resource:

                                                                                                                                              Lamerichs, J., & Stommel, W. (2016). “But how often does this happen?” Problem reducing responses by coaches in email counselling. In J. N. Lester & M. O’Reilly (Eds.), Palgrave handbook of adult mental health (pp. 287–307). London: Palgrave MacMillan.Find this resource:

                                                                                                                                                Lamerichs, J., & Te Molder, H. (2003). Computer-mediated communication: From a cognitive to a discursive model. New Media & Society, 5(4), 451–473.Find this resource:

                                                                                                                                                  Lievrouw, L. A. (2004). What’s changed about new media? Introduction to the fifth anniversary issue of new media & society. New Media & Society, 6(1), 9–15.Find this resource:

                                                                                                                                                    Limberg, H., & Locher, M. A. (Eds.). (2012). Advice in discourse. Amsterdam: John Benjamins.Find this resource:

                                                                                                                                                      Locher, M. A. (2012). Internet advice. In H. Limberg & M. A. Locher (Eds.), Advice in discourse (pp. 255–279). Amsterdam: John Benjamins.Find this resource:

                                                                                                                                                        MacMartin, C. (2008). Resisting optimistic questions in narrative and solution focused therapies. In A. Peräkylä, C. Antaki, S. Vehviläinen, & I. Leudar (Eds.), Conversation analysis and psychotherapy (pp. 80–99). Cambridge, UK: Cambridge University Press.Find this resource:

                                                                                                                                                          Madill, A. (2015). Conversation analysis and psychotherapy process research. In O. Gelo (Ed.), Psychotherapy research: General issues, outcomes and process (pp. 502–514). New York: Springer.Find this resource:

                                                                                                                                                            Markham, A., & Baym, N. (2009). Internet inquiry: Conversation about method. Thousand Oaks, CA: SAGE.Find this resource:

                                                                                                                                                              Meredith, J., & Potter, J. (2013). Conversation analysis and electronic interactions: Methodological, analytic and technical considerations. In H. L. Lim & F. Sudweeks (Eds.), Innovative methods and technologies for electronic discourse analysis (pp. 370–393). Hershey, PA: IGI Global.Find this resource:

                                                                                                                                                                Meredith, J., & Stokoe, E. (2013). Repair: Comparing Facebook “chat” with spoken interaction. Discourse & Communication, 8(2), 181–207.Find this resource:

                                                                                                                                                                  Morrow, P. R. (2006). Telling about problems and giving advice in an Internet discussion forum: Some discourse features. Discourse Studies, 8, 531–548.Find this resource:

                                                                                                                                                                    Morrow, P. R. (2012). Online advice in Japanese: Giving advice in an Internet discussion forum. In H. Limberg & M. A. Locher (Eds.), Advice in discourse (pp. 255–279). Amsterdam: John Benjamins.Find this resource:

                                                                                                                                                                      Nilsen, M., & Mäkitalo, A. (2010). Towards a conversational culture? How participants establish strategies for co-ordinating chat postings in the context of in-service training. Discourse Studies, 12(1), 90–105.Find this resource:

                                                                                                                                                                        Oravec, J. A. (2009). Online counselling and the Internet: Perspectives for mental health care supervision and education. Journal of Mental Health, 9(2), 121–135.Find this resource:

                                                                                                                                                                          Osvaldsson, K. (2011). Bullying in context: Stories of bullying on an Internet discussion board. Children & Society, 25(4), 317–327.Find this resource:

                                                                                                                                                                            Oudhoorn, N., & Somers, A. (2006). Patient organizations and the development of health websites. Information, Communication & Society, 9(5), 657–675.Find this resource:

                                                                                                                                                                              Pappas, Y., & Seale, C. (2010). The physical examination in telecardiology and televascular consultations: A study using conversation analysis. Patient Education and Counselling, 81, 113–118.Find this resource:

                                                                                                                                                                                Parsons, T. (1951). The structure of social action. New York: McGraw-Hill.Find this resource:

                                                                                                                                                                                  Paulus, T., & Varga, M. A. (2015). “Please know that you are not alone with your pain”: Responses to newcomer posts in an online grief support forum. Death Studies, 9(10), 633–640.Find this resource:

                                                                                                                                                                                    Paulus, T., Warren, A., & Lester, J. N. (2016). Applying conversation analysis methods to online talk: A literature review. Discourse, Context & Media, 12, 1–10.Find this resource:

                                                                                                                                                                                      Perakyla, A., Antaki, C., Vehviläinen, S., & Leudar, I. (2008). Conversation analysis and psychotherapy. Cambridge, U.K.: Cambridge University Press.Find this resource:

                                                                                                                                                                                        Pomerantz, A. (2012). Do participants’ reports enhance conversation analytic claims? Explanations of one sort or another. Discourse Studies, 14(4), 499–505.Find this resource:

                                                                                                                                                                                          Radley, A., & Billig, M. (1996). Accounts of health and illness: Dilemmas and representations. Sociology of Health and Illness, 18(2), 220–240.Find this resource:

                                                                                                                                                                                            Reed, D. (2001). “Making conversation”: Sequential integrity and the local management of interaction on Internet newsgroups. Proceedings of the 34th Hawaii International Conference on System Sciences. Retrieved from this resource:

                                                                                                                                                                                              Sacks, H. (1972). On the analyzability of stories by children. In J. J. Gumperz & D. Hymes (Eds.), Directions in sociolinguistics: The ethnography of communication (pp. 325–345). New York: Rinehart & Winston.Find this resource:

                                                                                                                                                                                                Sacks, H. (1992). Lectures on conversation (Vols. I & II). Oxford: Blackwell.Find this resource:

                                                                                                                                                                                                  Sacks, H., Schegloff, E., & Jefferson, G. (1974). A simplest systematics for the organization of turn-taking for conversation. Language, 50(4), 696–735.Find this resource:

                                                                                                                                                                                                    Sandaunet, A. (2008). A space for suffering? Communicating breast cancer in an online self-help context. Qualitative Health Research, 18(12), 1631–1641.Find this resource:

                                                                                                                                                                                                      Schegloff, E. A. (2007). A tutorial on membership categorization. Journal of Pragmatics, 39, 462–482.Find this resource:

                                                                                                                                                                                                        Schegloff, E. A., Jefferson, G., & Sacks, H. (1977). The preference for self-correction in the organization of repair in conversation. Language, 53(2), 361–382.Find this resource:

                                                                                                                                                                                                          Schönfeldt, J., & Golato, A. (2003). Repair in chats: A conversation analytic approach. Research on Language and Social Interaction, 36(3), 241–284.Find this resource:

                                                                                                                                                                                                            Scott, M. L., & Lyman, S. M. (1968). Accounts. American Sociological Review, 33, 46–62.Find this resource:

                                                                                                                                                                                                              Seale, C., Charteris-Black, J., MacFarlane, A., & McPherson, A. (2010). Interviews and Internet forums: A comparison of two sources of qualitative data. Qualitative Health Research, 20(5), 595–606.Find this resource:

                                                                                                                                                                                                                Severinson Eklundh, K. (2010). To quote or not to quote: Setting the context for computer-mediated dialogues. Language@Internet 7, art. 5. Retrieved from this resource:

                                                                                                                                                                                                                  Sidnell, J., & Stivers, T. (2013). The handbook of conversation analysis. Chicester, UK: Wiley-Blackwell.Find this resource:

                                                                                                                                                                                                                    Sillence, E. (2013). Giving and receiving peer advice in an online breast cancer support group. Cyberpsychology, Behaviour and Social Networking, 16(6), 480–486.Find this resource:

                                                                                                                                                                                                                      Sliedrecht, K. Y. (2014). Formulations in institutionele interactie: de praktijk van samenvatten in het politieverhoor, sollicitatiegesprek en journalistiek interview. Utrecht: LOT.Find this resource:

                                                                                                                                                                                                                        Sliedrecht, K. Y., Van der Houwen, F., & Schasfoort, M. (2016). Challenging formulations in police interrogations and job interviews: a comparative study. Journal of Pragmatics, V105N1(201611), 114–129.Find this resource:

                                                                                                                                                                                                                          Smithson, J., Sharkey, S., Hewis, E., Jones, R., Emmens, T., Ford T., & Owens, C. (2011a). Problem presentation and responses on an online forum for young people who self-harm. Discourse Studies, 13(4), 487–501.Find this resource:

                                                                                                                                                                                                                            Smithson, J., Sharkey, S., Hewis, E., Jones, R., Emmens, T., Ford T., & Owens, C. (2011b). Membership and boundary maintenance on an online self-harm forum. Qualitative Health Research, 21(11), 1567–1575.Find this resource:

                                                                                                                                                                                                                              Sneijder, P., & Te Molder, H. F. M. (2004). “Health should not have to be a problem”: Talking health and accountability in an Internet forum on veganism. Journal of Health Psychology, 9(4), 599–616.Find this resource:

                                                                                                                                                                                                                                Sneijder, P., & Te Molder, H. F. M. (2005). Moral logic and logic morality: Attributions of responsibility and blame in online discourse on veganism. Discourse & Society, 16(5), 675–696.Find this resource:

                                                                                                                                                                                                                                  Sneijder, P., & Te Molder, H. F. M. (2009). Normalising ideological food choice and eating practices. Identity work in online discussions on veganism. Appetite, 52(3), 621–630.Find this resource:

                                                                                                                                                                                                                                    Stokoe, E. H. (2012). Moving forward with membership categorization analysis: Methods for systematic analysis. Discourse Studies, 14(3), 277–303.Find this resource:

                                                                                                                                                                                                                                      Stommel, W. (2009). Entering an online support group on eating disorders. A discourse analysis (Published doctoral thesis). Amsterdam: Rodopi.Find this resource:

                                                                                                                                                                                                                                        Stommel, W. (2012). Salutations, closings and pronouns; some aspects of recipient design in online counselling. Journal of Communication and Medicine, 9(2), 145–158.Find this resource:

                                                                                                                                                                                                                                          Stommel, W. (2016). Information giving or problem discussion? Formulations in the initial phase of web-based chat counseling sessions. Journal of Pragmatics, 105, 87–100.Find this resource:

                                                                                                                                                                                                                                            Stommel, W., & Koole, T. (2010). The online support group as a community. A micro-analysis of the interaction with a new member. Discourse Studies, 12(3), 357–378.Find this resource:

                                                                                                                                                                                                                                              Stommel, W., & Lamerichs, J. (2014). Communication in online support groups: Advice and beyond. In H. Hamilton & W. S. Chou (Eds.), Handbook of language and health communication (pp. 191–211). New York: Routledge.Find this resource:

                                                                                                                                                                                                                                                Stommel, W., & Meijman, F. J. (2011). The use of conversation analysis to study social accessibility of an online support group on eating disorders. Global Health Promotion, 18(2), 18–26.Find this resource:

                                                                                                                                                                                                                                                  Stommel, W., & Te Molder, H. (2015). Counselling online and over the phone: When pre-closing questions fail as a closing device. Research on Language and Social Interaction, 48(3), 281–300.Find this resource:

                                                                                                                                                                                                                                                    Stommel, W., & Te Molder, H. (2016). When technical affordances meet interactional norms: The value of pre-screening in online chat counselling [Special issue on mediated interaction]. PsychNology Journal, 13(2–3), 235–258. Retrieved from this resource:

                                                                                                                                                                                                                                                      Stommel, W., & Van der Houwen, F. (2013). Formulations in “trouble” chat sessions. In I. Kupferberg & I. Gilat (Eds.), Computer-Mediated Troubled Talk [Special issue: Language@Internet, art. 3, n.p.]. Retrieved from this resource:

                                                                                                                                                                                                                                                        Stommel, W., & Van der Houwen, F. (2014). Complaining and the management of face in online counselling. Qualitative Health Research, 24(2), 183–193.Find this resource:

                                                                                                                                                                                                                                                          Stommel, W., & Van der Houwen, F. (2015). Counselling and new media technologies: A comparison of problem presentations in e-mail and in chat [Special issue on telemedicine]. Journal of Communication and Medicine, 12(2–3), 243–256.Find this resource:

                                                                                                                                                                                                                                                            Thornicroft, G., Rose, D., & Kassam, A. (2007). Discrimination in health care against people with mental illness. International Review of Psychiatry, 19(2), 113–122.Find this resource:

                                                                                                                                                                                                                                                              Thurnherr, F., Rudolf von Rohr, M., & Locher, M. A. (2016). The functions of narrative passages in three written online health contexts. Open Linguistics, S. 450–470.Find this resource:

                                                                                                                                                                                                                                                                Varga, M. A., & Paulus, T. (2014). Grieving online: Newcomers’ constructions of grief in an online support group. Death Studies, 38(7), 443–449.Find this resource:

                                                                                                                                                                                                                                                                  Vayreda, A., & Antaki, C. (2009). Social support and unsolicited advice in a bipolar disorder online forum. Qualitative Health Research, 19(7), 931–942.Find this resource:

                                                                                                                                                                                                                                                                    Vehviläinen, S. (2003). Preparing and delivering interpretations in psychoanalytic interaction. Text, 23(4), 573–606.Find this resource:

                                                                                                                                                                                                                                                                      Virtanen, T., Stein, D., & Herring, S. C. (2013). Pragmatics of computer-mediated communication. Berlin, Germany: De Gruyter Mouton.Find this resource:

                                                                                                                                                                                                                                                                        Voutilainen, L., Peräkylä, A., & Ruusuvuori, J. (2011). Therapeutic change in interaction: conversation analysis of a transforming sequence. Psychotherapy Research, 21(3), 348–365.Find this resource:

                                                                                                                                                                                                                                                                          Wallace, P. (1999). The psychology of the Internet. Cambridge, UK: Cambridge University Press.Find this resource:

                                                                                                                                                                                                                                                                            Winzelberg, A. J., Classen, C., Alpers, G. W., Roberts, H., Koopman, C., Adams, R. E., . . . Barr Taylor, C. (2003). Evaluation of an Internet support group for women with primary breast cancer. Cancer, 97(5), 1164–1173.Find this resource:

                                                                                                                                                                                                                                                                              Wood, L., & Kroger, (2000). Doing discourse analysis: Methods for studying action in talk and text. Thousand Oaks, CA: SAGE.Find this resource:

                                                                                                                                                                                                                                                                                Yuen, E. K., Goetter, E. M., Herbert, J. D., & Forman, E. M. (2012). Challenges and opportunities in Internet-mediated telemental health. Professional Psychology: Research and Practice, 43(1), 1–8.Find this resource:


                                                                                                                                                                                                                                                                                  (1.) Polylogal interaction or (text-based) “polylogues” refers to the potentially unlimited and unknown number of visitors to any given forum or thread in an online forum who might drop in and out of the interaction at any time (Bou-Franch, Lorenzo-Dus, & Garcès-Conejos Blitvich, 2012).