@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Education, Faculty of"@en, "Educational and Counselling Psychology, and Special Education (ECPS), Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Timm, Maria"@en ; dcterms:issued "2016-08-03T02:01:12"@en, "2016"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description "Despite an increase in the use of online crisis counselling services, little research has been conducted on how the therapeutic relationship is negotiated online. The prevalence of suicide in Canada as well the established importance of the therapeutic relationship when working with suicidal individuals led to the development of this study. The current study consisted of a discourse analysis of client-counsellor interactions in online crisis chats with individuals who reported thoughts of suicide. Data sources consisted of 24 transcripts obtained from two online crisis services: one for youth, one for adults. Of these, 16 were considered rapport-containing; eight were considered non-rapport-containing. Chats were separated into three phases: Initial Contact, Suicide Assessment, and Termination. Content analysis followed by discourse analysis found that, in the first two phases of the rapport-containing chats, the client tended to express a dialectic of wanting help and simultaneously feeling hopeless and of a person with a story to tell, and the counsellor positioned themselves as an empathic witness to the client’s narrative. The client-counsellor relationship in Phase 1 was characterized by themes of informality and equality/mutual respect. Main relational themes in the Suicide Assessment phase were client/counsellor collaboration and counsellor authenticity. In the Termination phase, main relational themes included shared humour, counsellor self-disclosure, and client trust. In the non-rapport-containing chats, in the Initial Contact phase, the client positioned themselves as a consumer of services and the counsellor positioned themselves as a service provider. The client-counsellor relationship in this phase was characterized by client frustration and counsellor helplessness with respect to the client’s unmet needs for counsellor directiveness, authenticity, and self-disclosure. In the Suicide Assessment phase, three main relational themes were found: client-perceived circularity of the conversation, feeling misunderstood, and feeling unheard. In the Termination phase, chats were frequently ended abruptly by the client, and the predominant theme was one of client rejection of the counsellor. Across all chats, client-perceived mattering (or lack thereof) was an observed theme. The results have important clinical implications for those working with suicidal individuals online."@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/58652?expand=metadata"@en ; skos:note " “Are you a robot?” A discourse analysis of rapport-building in online crisis chats from a suicide prevention centre by Maria Timm M.A., The University of British Columbia, 2011 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate and Postdoctoral Studies (Counselling Psychology) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) July 2016 © Maria Timm 2016 ii Abstract Despite an increase in the use of online crisis counselling services, little research has been conducted on how the therapeutic relationship is negotiated online. The prevalence of suicide in Canada as well the established importance of the therapeutic relationship when working with suicidal individuals led to the development of this study. The current study consisted of a discourse analysis of client-counsellor interactions in online crisis chats with individuals who reported thoughts of suicide. Data sources consisted of 24 transcripts obtained from two online crisis services: one for youth, one for adults. Of these, 16 were considered rapport-containing; eight were considered non-rapport-containing. Chats were separated into three phases: Initial Contact, Suicide Assessment, and Termination. Content analysis followed by discourse analysis found that, in the first two phases of the rapport-containing chats, the client tended to express a dialectic of wanting help and simultaneously feeling hopeless and of a person with a story to tell, and the counsellor positioned themselves as an empathic witness to the client’s narrative. The client-counsellor relationship in Phase 1 was characterized by themes of informality and equality/mutual respect. Main relational themes in the Suicide Assessment phase were client/counsellor collaboration and counsellor authenticity. In the Termination phase, main relational themes included shared humour, counsellor self-disclosure, and client trust. In the non-rapport-containing chats, in the Initial Contact phase, the client positioned themselves as a consumer of services and the counsellor positioned themselves as a service provider. The client-counsellor relationship in this phase was characterized by client frustration and counsellor helplessness with respect to the client’s unmet needs for counsellor directiveness, authenticity, and self-disclosure. In the Suicide Assessment phase, three main relational themes were found: client-perceived circularity of the conversation, feeling misunderstood, and feeling unheard. In the Termination phase, chats were frequently ended abruptly by the client, and the predominant theme was one of client rejection of the counsellor. Across all chats, client-perceived mattering (or lack thereof) was an observed theme. The results have important clinical implications for those working with suicidal individuals online. iii Preface This study was designed and conducted by myself, Maria Timm, using information obtained in collaboration with the Vancouver Crisis Centre and supervised by Dr. Ishu Ishiyama. Permission was granted by Vancouver Crisis Centre Clinical Director Elizabeth Robbins. Program Coordinator, Josh Cytrynbaum, assisted with data collection. Melissa Penner, BA, assisted with data analysis. Ethics approval was granted from UBC Behavioural Research Ethics Board on July 28, 2015 (number H15-00714). Dr. Ishu Ishiyama, Dr. Colleen Haney, and Dr. Marvin Westwood served as committee members, and Dr. Leslie Roman and Dr. Jennifer Vadeboncoeur were consultants in the data analysis. iv Table of Contents Abstract ........................................................................................................................................... ii Preface............................................................................................................................................ iii Table of Contents ........................................................................................................................... iv List of Tables ................................................................................................................................. ix List of Figures ..................................................................................................................................x Acknowledgments.......................................................................................................................... xi Dedication ..................................................................................................................................... xii Chapter One: Introduction ..........................................................................................................1 Statement of Problem ...........................................................................................................1 Gaps in the Literature...........................................................................................................1 Rationale for Study ..............................................................................................................2 Chapter Two: Literature Review .................................................................................................4 Suicide: Definition of Terms ...............................................................................................4 Suicide: Etiological Factors .................................................................................................6 Psychopharmacology and Suicide: Risks to Youth .............................................................7 Suicide Risk Assessment: Process .......................................................................................9 Prediction of Suicide: Methodological Limitations ...........................................................10 Suicide Risk Assessment: Content.....................................................................................11 Do Suicide Prevention Programs Work? ...........................................................................13 Protective Factors...............................................................................................................14 Therapist Factors in Reducing Suicide Risk ......................................................................14 Help-Negation ..................................................................................................................16 Online Counselling: A Growing Field ...............................................................................16 Online and In-Person Counselling: Clinical Outcomes ....................................................18 Rapport: The Importance of the Therapeutic Connection .................................................20 The Real Relationship ............................................................................................22 The Importance of Rapport ....................................................................................22 Communication of Rapport....................................................................................23 Building Rapport Online: Challenges and Preliminary Findings ......................................25 Online Calming Hypothesis ...................................................................................27 v Rapport-Building Online: Suggestions from the Research ................................................28 Research Questions .....................................................................................................................29 Site of the Study: Vancouver Crisis Centre ................................................................................31 Chapter Three: Method ..............................................................................................................33 Theoretical Assumptions and Rationale .....................................................................................33 Epistemological Stance: Social Constructionism .......................................................................35 Trustworthiness ..........................................................................................................................36 Researcher Reflexivity ...............................................................................................................37 Participants (Data Sources) ........................................................................................................38 Rapport-Containing versus Non-Rapport-Containing Chats .....................................................43 Identification of Rapport-Containing Chats...............................................................................44 Other Measures of Rapport ................................................................................................45 Satisfaction Scale ...............................................................................................................48 Rationale for Size of Data Set ....................................................................................................49 Delineation of Phases .................................................................................................................50 Data Analysis .............................................................................................................................51 Stage One: Content Analysis .............................................................................................51 Stage Two: Discourse Analysis .........................................................................................54 Discourse Analysis: Theory and Practice ..........................................................................56 Applicability of a Foucauldian Lens .................................................................................59 The Virtual Space: Applicability of Discourse Analysis ..................................................62 Discourse Analysis Tools Used .........................................................................................63 The Identities-Buildings Tool ............................................................................................63 The Relationships-Building Tool .......................................................................................65 The Activities-Building Tool: ............................................................................................66 The Figured-Worlds Tool ..................................................................................................66 Ethical Concerns ........................................................................................................................68 Informed Consent...........................................................................................................68 Confidentiality ...............................................................................................................69 Increasing Trustworthiness: Data Verification ..........................................................................70 vi Chapter Four: Findings ...............................................................................................................73 Rapport-Containing Chats ..........................................................................................................73 Phase 1 (Initial Contact Phase): Client Identity ................................................................73 Phase 1 (Initial Contact Phase): Counsellor Identity ........................................................74 Youth versus Adult ...........................................................................................................75 Theme 1: Informality ........................................................................................................75 Theme 2: Equality and Respect .........................................................................................76 Youth versus Adult ...........................................................................................................76 Phase 2 (Suicide Assessment Phase): Client Identity .......................................................76 Phase 2 (Suicide Assessment Phase): Counsellor Identity ................................................77 Youth versus Adult ............................................................................................................77 Phase 2 (Suicide Assessment Phase): Client-Counsellor Relationship .............................78 Theme 1: Client/Counsellor Collaboration ......................................................................78 Theme 2: Counsellor Genuineness ....................................................................................78 Youth versus Adult ............................................................................................................79 Phase 3 (Termination Phase): Client Identity ....................................................................80 Phase 3 (Termination Phase): Counsellor Identity ............................................................80 Youth versus Adult ............................................................................................................80 Phase 3 (Termination Phase): Client/Counsellor Relationship ..........................................80 Theme 1: Counsellor Self-Disclosure ................................................................................81 Theme 2: Shared Humour .................................................................................................81 Theme 3: Client Trust ........................................................................................................81 Youth versus Adult ............................................................................................................82 The Activities-Building Tool .........................................................................................................82 Fourth Phase: Social Conversation ....................................................................................83 The Figured-Worlds Tool ..............................................................................................................84 Non-Rapport-Containing Chats ....................................................................................................90 Youth versus Adult .................................................................................................................90 Phase 1 (Initial Contact Phase): Client/Counsellor Relationship ...........................................91 Theme 1: Criticism of Counsellor Communication ...................................................91 vii Theme 2: Client Requests for Personal Information .......................................................91 Youth versus Adult ............................................................................................................92 Phase 2 (Suicide Assessment Phase): Client Identity ........................................................92 Phase 2 (Suicide Assessment Phase): Counsellor Identity ...............................................93 Youth versus Adult ............................................................................................................93 Phase 2 (Suicide Assessment Phase): Client/Counsellor Relationship..............................93 Theme 1: Misunderstanding ..............................................................................................94 Theme 2: Feeling Unheard ................................................................................................94 Youth versus Adult ............................................................................................................93 Phase 3: (Termination Phase): Client Identity ...................................................................94 Phase 3 (Termination Phase): Counsellor Identity ...........................................................95 Youth versus Adult ...........................................................................................................95 Phase 3 (Termination Phase): Client/Counsellor Relationship ..........................................95 Theme 1: Client Rejection of Counsellor ..........................................................................95 Youth versus Adult ............................................................................................................96 Activities-Building Tool: ...........................................................................................................96 Figured-Worlds Tool ..................................................................................................................97 Gender ......................................................................................................................................102 Chapter Five: Discussion ...........................................................................................................103 Client/Counsellor Identities ....................................................................................................103 Importance of Client Narrative ................................................................................................104 Client/Counsellor Relationship ................................................................................................105 Social Power ............................................................................................................................107 Client/Counsellor Activities.................................................................................................108 Use of Humour/Informal Language ....................................................................................110 Counsellor Self-Disclosure ................................................................................................111 Common Themes: Mattering ..............................................................................................113 Age Differences ..................................................................................................................118 Gender Differences .............................................................................................................119 Potential Clinical Applications ............................................................................................120 viii Counsellor Flexibility ...........................................................................................................122 Counsellor Transparency/Authenticity .................................................................................122 Limitations ............................................................................................................................124 Directions for Future Research .............................................................................................125 Chapter Six: Conclusion ...........................................................................................................127 Epilogue .....................................................................................................................................130 References ...................................................................................................................................131 Appendices ..................................................................................................................................145 Appendix A: Volunteer Consent Form ..................................................................................145 Appendix B: Satisfaction Scale .............................................................................................147 Appendix C: Disclaimer Placed On Online Sites ..................................................................148 Appendix D: Letter of Consent from Vancouver Crisis Centre ............................................149 Appendix E: Sample Wordle .................................................................................................150 ix List of Tables Table 1: Nomenclature of Suicidal Behaviour.................................................................................5 Table 2: Demographic Information: Rapport-Containing Chats ...................................................42 Table 3: Demographic Information: Non-Rapport-Containing Chats ..........................................43 Table 4: Rapport Measures ............................................................................................................47 Table 5: Explicit versus Implicit Messages: Illustrative Example ................................................55 Table 6: Contributions of Discourse Analysis .............................................................................56 Table 7: Rapport-Containing Chats: Themes by Phase/DA Tool ................................................100 Table 8: Non-Rapport-Containing Chats: Themes by Phase/DA Tool .......................................101 x List of Figures Figure 1: Expansion of the Therapeutic Space ..............................................................................17 Figure 2: Sample Coding of Rapport-Containing Chat .................................................................86 Figure 3: Sample Coding of Rapport-Containing Chat ................................................................87 Figure 4: Sample Coding of Rapport-Containing Chat .................................................................89 Figure 5: Sample Coding of Non-Rapport-Containing Chat ........................................................98 Figure 6: Ishiyama’s Validationgram .........................................................................................121 xi Acknowledgements I would like to first thank my research supervisor, Dr. Ishu Ishiyama, for his insights and consistent support, as well as the introduction to the validationgram and Morita Therapy. I also appreciate the input and guidance provided by committee members Dr. Colleen Haney and Dr. Marvin Westwood, as well as their supportive and encouraging presence during my proposal defense. Dr. Leslie Roman and Dr. Jennifer Vadeboncoeur provided invaluable assistance in the data analysis portion of this project. Their contributions and insights regarding discourse analysis are gratefully acknowledged. I would also like to thank Dr. Jennifer White (who I have not yet met in person) for writing so evocatively on youth suicide prevention and for unknowingly providing the inspiration for my choice of research methodology. This project would not have been possible without the support of the Vancouver Crisis Centre. I will always be grateful to the staff and volunteers who run this life-affirming organization that inspired me to study counselling after my first suicide call as a volunteer many years ago. In particular, I would like to thank Liz Robbins, Clinical Director, and Josh Cytrynbaum, Online Services Coordinator, for their assistance and patience with the data collection, and Akhila Blaise for her efforts in developing and implementing volunteer training. I would like to thank Melissa Penner, BA, for her assistance in data analysis. Melissa provided not only data verification assistance in her role as research assistant but also contributed valuable insights which shaped how the results were expressed and inspired me to think differently about the data. I am grateful to my family and friends for their support and encouragement. In particular, my parents, Dr. Johanna Timm and Dr. Eitel Timm, provided the inspiration and tangible proof that pursuing a Ph.D is not only possible but also life-enriching. Lastly, I am grateful for my Ph.D cohort: Ada, Deepak, Jessie, Kevin, and Lynn, for your solidarity and commiseration, which made this process so much more enjoyable. As we say in Austin, “I appreciate you!” xii Dedication To the inmates, with deep gratitude for inspiring me to offer “mattering” in the toughest places. To Trinity, the little girl who taught me the importance of meaning and connection. 1 1. INTRODUCTION Suicide may also be regarded as an experiment - a question which man puts to Nature, trying to force her to answer. The question is this: What change will death produce in a man's existence and in his insight into the nature of things? It is a clumsy experiment to make, for it involves the destruction of the very consciousness which puts the question and awaits the answer. - Arthur Schopenhauer, Parerga and Paralipomena (cited in Shea, 2008 p. 6). Statement of Problem A distressed person takes their own life every 40 seconds somewhere in the world, according to the Canadian Mental Health Association (Canadian Mental Health Association, 2013). Concerningly, it has been observed that those at highest risk of suicide are often least likely to reach out for help from mental health professionals. This phenomenon is often referred to in the suicidology literature as help-negation (Wilson & Deane, 2010). Help negation has been explained by invoking a number of potential cognitive and affective client factors: the stigma surrounding mental health issues in general and suicide in particular (Curtis, 2010), lack of awareness about existing resources, concerns about those resources being unhelpful or inacessible (Wilson & Deane, 2010), and concerns surrounding the confidentiality of disclosure of suicidality (Menna & Ruck, 2005). Recently, online counselling has emerged as an alternative to the traditional face-to-face or telephonic crisis intervention for individuals who are feeling suicidal, and has been increasingly utilized by these individuals (Martin & Stuart, 2011; Richards & Vigano, 2013). This increase has been attributed to numerous factors, including both the ease (Cook & Doyle, 2002; Greidanus & Everall, 2010), and anonymity (Lapidot-Lefler & Barak, 2015; Suler, 2005) of the provision of services online. Gaps in the Literature The literature review below discusses in more detail several key themes with respect to gaps in the literature on suicide prevention, intervention, and etiology. Four specific gaps in 2 knowledge were focused on in the current study. First, obtaining access to conversations about suicidality directly from the perspective of the suicidal person (as opposed to via third-party or after-the-fact questionnaires) was hoped to provide valuable information to assist helpers in determining how to best support suicidal individuals. Second, the stigma of suicide specifically and mental health in general, and the frequently observed finding that those most in need of help do not seek it, points to a need to examine other increasingly used modalities of helping, such as online counselling (Hanley, 2006). Third, the importance of the therapist-client relationship when working with suicidal individuals has been frequently cited as one of the most valuable contributor towards positiveclient outcomes (Granello, 2010; Joiner, 2005; Shea, 2008; Shneidman, 1998), and was therefore a primary focus of this investigation. Lastly, this investigation strove to include the voices of those not traditionally represented in the existing suicidology research due to both sampling difficulties and bias. This includes those in remote communities (including some self-identified as from Aboriginal communities in Northern British Columbia) who may not have had access to in-person counselling resources. It also includes the voices of those who are underrepresented in the existing suicidology research: male clients and adolescent clients, who have historically shown high rates of help-negation towards traditional in-person counselling (Wilson & Deane, 2010). Rationale for Study Despite a growing body of literature examining how rapport is built online (e.g., Hanley, 2006; Martin & Stuart, 2011; Reynolds et al., 2011; Roy & Gillett, 2008), there is a significant lack of research on how rapport is built online when individuals are suicidal. Specifically, searches conducted using the databases PsycINFO, Academic Search Complete, PsycARTICLES, PsycBOOKS, PsycEXTRA, eBookCollection (EBSCOhost), and Humanities 3 and Social Sciences Index Retrospective and employing the keywords “suicide,” and “rapport,” or “alliance” and “online,” “cyber,” or “internet” yielded only five results. This lack of research in the area of addressing suicidality online can be attributed to ethical concerns about collecting data on those who are suicidal online, lack of a reliable and valid measure of rapport for online counselling, difficulty accessing data from crisis lines, and the relative newness of this modality of delivering counselling services. Therefore, the following section will provide a general overview of current research regarding how rapport can be built online, with suggestions for how these findings may be applicable to individuals who are suicidal. Use of online counselling has increased in recent years and represents an important avenue of further exploration with respect to its applicability for those who are suicidal. The anonymity (e.g., Sade-Beck, 2004), ease of access (e.g., Cook & Doyle, 2002), and increasing use of technology by both therapists and clients (e.g., Martin & Stewart, 2011; Reynolds et al., 2013) have been cited as key reasons for this increase in provision of online counselling services. Lastly, the research reviewed below appears to point to the importance of building a therapeutic connection with those who are suicidal (e.g., Durkheim, 1951; Granello, 2010; Joiner, 2005; Shea, 2008; Shneidman, 1998; Van Orden et al., 2010). The social stigma associated with suicide, the high stakes of this conversation, and the heightened sense of isolation and need for connection all emphasize that building rapport in online counselling is of vital importance. This study therefore examined how rapport was built in crisis chats from an online suicide prevention service. 4 2. LITERATURE REVIEW This chapter will summarize the theoretical, epistemological, and professional evolution of suicide prevention as a discipline, providing an overview of proposed etiology, risk factors, and new directions in suicide prevention. Specifically, the development of online services to support suicidal individuals will be summarized and critiqued, with a focus on how the client-counsellor relationship is both established and impacted when crisis counselling takes place online. Suicide: Definition of Terms Suicide has been succinctly described as “a self-injurious act with the sole intent to die” (Silverman et al., 2007, p. 250). Despite this seemingly parsimonious definition, Silverman et al. (2007) note that the lack of a universally agreed-upon nomenclature for suicidal behaviour has adversely affected any attempts to systematically study it. Lack of a universal nomenclature impedes the undertaking of a systematic meta-analysis or attempts to compare the results of one study with those of another; this in turn impacts the ability of quantitatively oriented researchers to generalize beyond the findings of a particular study. Silverman et al. (2007) note that “[…] the term suicide refers not to a single action but more broadly to a great many varied behaviours […] one can speak of suicidal thoughts, intentions, ideation, gestures, attempts, completions, equivalents” (p. 428). The ambiguity inherent in many of these commonly used terms in suicidology thus adversely affects the ability of researchers to make meaningful inferences about the results of studies. For example, individuals who attempt suicide display different characteristics from those who engage in self-harm behaviours (Silverman et al., 2007), and placing these in the same category may adversely affect the generalizability of the findings. 5 O’Carroll et al. (1996) proposed a nomenclature of suicidality to be adopted by those in the field of suicidology, including coroners, police, mental health professionals, and psychiatric nurses. This table has been adopted by the APA in their recent practice guidelines for suicide assessment (APA, 2003) and is noted below. Table 1 Nomenclature of suicidal behaviour * obtained from O’Carroll et al., 1996, p. 251; used with permission. 6 The nomenclature proposed by O’Carroll (1996) highlights three essential features of suicidality: intent to die, evidence of self-inflicted injury, and outcome (injury, no injury, or death). This nomenclature, and the above-noted definition of suicide, will be used when referring to “suicide” in this paper. Suicide: Etiological Factors Emile Durkheim’s seminal book On Suicide viewed suicide through a sociological lens, concentrating mainly on 19th century Europe. His work shifted the etiology of suicidal ideation from residing in the individual to originating in a social context. He noted that individuals with lower social integration were more likely to end their lives (Durkheim, 1951). This proposed etiology was linked to his observation that when societies were in flux resulting in decreased levels of social regulation, suicide rates increased. This led to two interrelated hypotheses regarding the etiology of suicide: lack of social integration in the individual or a major change in the individual’s societal environment. Durkheim argued that these factors resulted in a sense of anomie – a feeling of hopelessness due to the perception of goals and aspirations as unattainable (Durkheim, 1951). Today, Durkheim’s initial etiology has been extended: it is recognized by most suicidologists that suicidal ideation arises from the interaction of numerous biopsychosocial factor and current researchers have elaborated on Durkheim’s initial hypothesis. For example, Shneidman (1998) refers to the emotional state that often leads to suicide ideation as “psychache” (p. 10) which he states is composed of intense negative feelings; for example, shame, fear or loneliness, and is thought to be due to frustrated psychological needs. Shneidman posits that the simultaneous presence of three factors is required for suicidal behaviour to occur: psychache, press (stress) and perturbation (agitation). Hopelessness (Winter 7 et al., 2013), feeling that one is a burden either to society or significant others (Van Orden et al., 2008), and a lack of social belonging (Joiner, 2005) are also cited as intra/interpersonal correlates of suicidality. The interpersonal theory of suicide (Joiner, 2005) views belonging and social connection as a fundamental human need, the thwarting of which can lead to suicidality. The most commonly agreed upon risk factors for suicidality are depression and/or other mood disorders involving mood disturbance such as bipolar disorder, schizoaffective disorder, and schizophrenia (APA, 2013), stressors of a situational or environmental nature, and social isolation (Westefeld, Range, Rogers, & Hill, 2010). Some psychological disorders are associated with particular factors that predispose individuals to suicidality; for example, the affect instability associated with Borderline Personality Disorder is associated with behavioural dysregulation, which increases risk of suicide behaviour (Linehan, 1994). Personality traits such as neuroticism and extraversion have also been implicated (Rogers & Lester, 2009). Gender differences in suicide completion have been attributed to personality (Witte, Gordon, Smith & Van Orden, 2012) and interpersonal/social factors (Van Orden et al., 2010). Overholser et al. (2011) note that a precipitating event occurs in the majority of cases of completed suicide. Psychopharmacology and Suicide: Risks to Youth A significant body of research has emerged regarding youth use of SSRI’s (selective serotonin reuptake inhibitors) following the “black box” warning issued by the FDA in 2006 that these drugs may cause suicidal thoughts in those under 18 and the subsequent ban on prescribing these to youth in the UK. Prior to this, SSRI’s had been frequently prescribed to young people between 1988 and 2004. Following the warning, prescription of anti-depressants to youth sharply declined by 31% within the first two years of the warning (Sparks & Barry, 2013). It has been noted that individuals under age 25 are at greatest risk of suicide following prescription of an 8 SSRI and that risk increases as age decreases (Sparks & Barry, 2013). Interestingly, emotional arousal rates of youth treated with anti-depressants were found to be 3-10 times higher than those in individuals treated with placebos (Sparks & Barry, 2013). It has been posited that the increase in suicide attempts may result partly from the increase in energy and motivation provided by the influx of serotonin (Isacsson & Ahlner, 2014). Therefore, the general guidelines for practitioners as per the Guidelines for Adolescent Depression in Primary Care Treatment (GLAD-BC) are to offer a course of cognitive-behavioural, interpersonal, or family therapy for a period of at least three months prior to suggesting medication, and that medication should only be offered when psychological intervention is unsuccessful, inaccessible, or declined (Gordon & Melvin, 2014). On this note, Cutcliffe et al. (2014), the editors of the Routledge International Handbook of Clinical Suicide Research, observe that it is important for clinicians to distinguish between situational and endogenous factors so that they ensure that they are medicating a serotonin imbalance and not a circumstance. As noted above, suicide is a particularly pressing concern for youth. About 50% of young people who complete suicide have a diagnosable mood disorder, such as depression (Cheung & Dewa, 2007). Additionally, Berman (1991) observes that almost all published studies on suicide notes – what Shneidman (1998) would refer to as “psychological autopsies” – have focused on adults. He proposed that young adults may differ in patterns of linguistic and emotional expression in their suicide notes. Considering the high rate of suicide among youth (CMHA, 2013), examining how youth conceptualize and speak about suicide represents an important aspect of suicide prevention to explore. When clinical prediction of suicide was examined, Garb (1998) found that several variables increased the interrater reliability of clinicians assessing risk using clinical means: the 9 source of information (an interview format was seen as more effective than a case consultation), the length of time elapsed from meeting to prediction (better predictions were made for predicting suicidal behavior within the week), and the amount of information [italics mine] available to the clinician” (Brown & Lent, 2008, p. 542). The importance of gathering pertinent information from the client cannot be overstated. However, the clinician walks a fine line between collecting potentially sensitive client information and maintaining rapport. This underscores the importance of building a therapeutic alliance in suicide risk assessment, the process and content of which will now be discussed. Suicide Risk Assessment: Process The most commonly employed model of suicide assessment is the crisis intervention model (e.g., Westefeld et al., 2010). This model has been widely employed in suicide prevention centres and by emergency mental health providers. It generally includes assessment of risk, validation and empathy of client emotions, and follow-up planning. However, research attempting to determine what makes an “effective” suicide risk assessment is limited by the lack of professional agreement on operational definitions of several key constructs. For example, there has been much disagreement among researchers on the definitions of suicidal intent, a key component of the nomenclature proposed by O’Carroll et al. (1996). Silverman et al. (2007) note that different stakeholders (statisticians, medical examiners, coroners, public health practitioners) may view intent differently and have different requirements for establishing it. They summarize the four key components of intent that tend to be required across all professions: 1) intention to take the action; 2) intention to harm himself or herself by the action; 3) intention to die as a result of the action; 4) at the time of acting, a capacity to understand the likely consequences of the act and form the desire to die (Silverman et al., 2007, p. 255). 10 An additional methodological concern with respect to assessing intent is the importance of determining which mode of assessment should be used to infer it: Rudd (2006) describes the difficulty in determining whether suicidal intent – and perceived risk – should be viewed as subjective (client-rated) or objective (clinician-rated) in nature. He further recommends that clinicians attempt to clarify and resolve potential discrepancies between observed behavior (e.g., cutting), and the client’s self-reported cognitions. This underlines the importance of the clinician’s ability to build rapport and facilitate client self-disclosure in the suicide risk assessment process. Prediction of Suicide: Methodological Limitations Suicide risk assessment, by nature, assumes an ability to predict the occurrence of suicide. Edwin Shneidman (1998), who is frequently cited as the “father” of modern suicidology, states succinctly that reducing the individual’s level of perturbation (perceived emotional distress) will result in a consequent reduction in what he terms “lethality” (p. 153). Lethality has historically been a difficult construct to operationally define (Silverman et al., 2007). Researchers generally include availability/proximity to method, chance of intervention, and familiarity with a given method as variables that influence method lethality (Silverman et al., 2007). While many current suicide assessments are structured according to the notion that lethality of chosen method is positively correlated with suicide risk, a number of studies have questioned this finding; Brown, Henriques, Sosdjan, and Beck (2004) found only a small positive association between these two variables, calling into question the validity of method lethality as a predictor of suicide risk. Of interest to practitioners was the finding that over half of the participants in their study were inaccurate in their estimations of the lethality of their chosen method. These findings highlight the importance of taking into account subjective client perceptions of risk in addition to objective, clinician-rated assessments. Additionally, the lack of 11 a common definition of lethality calls into question the reliability and validity of studies including lethality as a predictive variable. Suicide Risk Assessment: Content Shneidman (1998) states that the therapist’s function in suicide assessment could be likened to that of an “anodyne” (p. 151) – a psychological pain reliever. He emphasizes that understanding the nature of the client’s pain and the source of their frustrated needs is vitally important if the therapist is to fulfill this pain-relieving role. In terms of follow-up planning, Shneidman (1998) notes that once immediate risk has diminished, the counsellor and client can both “learn, in admittedly different ways, the mental landscape of the beleaguered person” (p. 151). This entails, according to Shneidman, an examination of their unique psychological needs (for example, for nurturance, understanding, and achievement), which he believes are vital in understanding their suicidality as the frustration of these is seen as a common stimulus for a suicidal act (Shneidman, 1998). Furthermore, Shneidman notes the importance of attending to the practicalities of the suicidal person’s life. He describes this as acting as a temporary “ombudsman,” undertaking such concrete activities as helping the client schedule daily tasks. This approach has been criticized by some as disempowering for the client by placing the clinician in a position of authority, which has been compared to the qualities endorsed by emergency room physicians: assuming implicit responsibility for the client’s life and decisions without taking their life context into account. Other criticisms of the model include a lack of empirical evidence concerning its effectiveness, difficulty establishing a therapeutic alliance, and inconsistency in the application of the model (Westefeld et al., 2010). In terms of the process of suicide assessment, Shneidman (1998) outlines the difference between a traditional conversation and a therapeutic one. He notes that the former is generally 12 focused on manifest content and usually implies equality between participants, whereas the latter contains both manifest and latent content and is a more hierarchical relationship. Furthermore, the therapeutic conversation is unique in that clients come to therapy to discuss things “ordinarily not mentioned” (Shneidman, 1998, p. 141). Therefore, it is important for the therapist to be aware of this difference in conversational tone and – particularly in the case of suicide assessment – to be alert and aware of subtle nuances in a person’s conversation that may indicate that they are contemplating ending their life. Shneidman (1998) addresses the limitation of language to describe pain “in ordinary words, through the admittedly limited funnel of language, to travel from private inchoate experience to interpersonal communication through the use of culturally defined words and phrases” (p. 10). Often, it is not the specific words but the images evoked that can provide insight into the experience of being suicidal. Metaphors can provide a window into the inner landscape of the suicidal person. This landscape, according to the accounts of those who have survived or contemplated a suicide attempt, is often silent, empty, dark, enclosed, and replete with quicksand, enemies, natural disasters, and nautical mishaps such as shipwrecks (e.g., Kaviani & Hamedi, 2013; Shneidman, 1998). Attempting to distill the experience of psychological pain into a measurable construct, Shneidman (1998) developed the Psychological Pain Survey. This survey uses the method of paired comparisons to help gauge the severity of an individual’s perceived psychological suffering. The vignette that serves as a baseline for the most intense pain is that of an individual in a Nazi concentration camp, a culturally embedded example that may not be an appropriate or known emotional reference point for all individuals. 13 Do Suicide Prevention Programs Work? The difficulty in evaluating the effectiveness of suicide prevention programs lies partly in the methodology of most research designs on this topic. The majority of investigations have focused on client post-hoc self-reports (e.g., Gould et al., 2007; Mishara et al., 2007; Neimeyer & Pfeiffer, 1994), which are subject to both hindsight and informant bias (Silverman et al., 2007). Other studies have used trained raters to assess the effectiveness of suicide prevention programs (e.g., Gould et al., 2007). These studies, while helpful, likely miss important information since they consist of self-reports made by third parties. In the 1970’s, efforts were made to demonstrate the effectiveness of suicide prevention centres in reducing suicide via the crisis intervention model, outlined above. While some studies indicated that geographic locations with suicide prevention centres had lower suicide rates (Lester, 1997), the correlation between decreased rates of suicide and presence of these centres could also be attributed to a third factor, such as the size of the city itself which would likely be positively correlated with access to mental health screening and resources. Additionally, as noted in the above chapter, much research on suicide prevention centres cites the primary users of these services as being young European American females (Westefeld et al., 2010). Since men are more likely to complete suicide (APA, 2013) and since suicide occurs frequently among the growing immigrant population in Canada and the United States and among certain Aboriginal communities (Clifford, Doran, & Tsey, 2013) the existing data may be lacking the voices of those who are most in need of help: males, immigrants, Aboriginal communities, and those high in help-negation. For example, Clifford et al. (2013) conducted a meta-analysis of suicide intervention programs targeting indigenous populations in Canada, the United States, Australia, and New Zealand. They found that out of nine suicide intervention strategies mentioned in their 14 search of 17 electronic databases and 13 websites, only three of these directly measured changes in rates of suicide or suicidal behaviour. The authors note several methodological issues in the study designs: most notably, reliance on self-report measures, strong variation in participant consent and follow-up rates, and a lack of cost evaluation procedures. Considering that Canada’s Aboriginal communities suffer a suicide rate at least twice that of the general population (Clifford et al., 2013), it seems important to conduct research into the predictive and protective factors of suicide among Aboriginal Canadian populations. Protective Factors The preceding section has focused on methodological limitations in the study of the etiology, assessment, and therapeutic response to those in suicidal crisis. Etiological explanations of suicidal behaviour have historically included situational, interpersonal, and intrapersonal risk factors. Situational factors that may protect against suicidal ideation are generally considered the inverse of those that contribute to risk (e.g., social isolation versus connection). Most researchers seem to agree that a combination of situational, interpersonal, and intrapersonal factors contribute to decreasing an individual’s suicide risk. For example, Berman (1991) outlines three categories of protective factors: a) dispositional attributes such as self-esteem, and feelings of autonomy and control, b) family cohesion and warmth, lack of family discord, and c) the availability and use of external supports and resources (Berman, 1991, p. 106). It is clear from viewing this list that protective factors against suicide seem to comprise a combination of internal (intra-personal) and external (contextual/situational) factors. Therapist Factors in Reducing Suicide Risk Which therapeutic factors have been found to influence suicide risk? Many clinicians who work in the field of suicide prevention note the importance of the therapeutic alliance. For 15 example, Shea (2008) notes that “[…] the degree to which the patient shares these [suicidal] thoughts and shares them truthfully, is a limiting factor in how accurate the clinician can be in predicting suicide” (p. 18) thus implying that the client’s level of comfort with the clinician is likely to impact his or her disclosure and therefore the assessment process. Once more, this underscores the importance of building rapport when working with clients who are suicidal. Interestingly, research has found that use of humour appears to be a moderator of the effects of stressful life events; in particular, it appears protective against developing depressive symptomatology (Nezu, Nezu, & Blissett, 1988), which in turn is a risk factor for suicidality, as noted previously. A methodological limitation of most suicidology research is the inability of researchers to capture important aspects of the suicidal experience. Namely, most research is conducted, for obvious practical reasons, with participants who have either survived a suicide attempt or, most commonly, who have reached out for help prior to attempting suicide. The voices of those who have completed suicide are, for all intents and purposes, lost – with the exception of the “psychological autopsies” conducted by Shneidman (1998) and others who attempt to reconstruct the deceased person’s life and post-humously identity risk factors. This is important to note because, as detailed above, those who do not reach out for help may have different characteristics from those who do. Thus, there is a lack of knowledge in the field of the lived experiences and “in the moment” cognitions and emotions of those who are contemplating suicide. On a related note, it has been observed that those at highest risk of suicide are often least likely to reach out. This is often referred to in the suicidology literature as “help-negation” (Wilson & Deane, 2010) and will now be expanded on as it represents an important area of suicide prevention due to its potentially lethal impact on the individual. 16 Help-Negation Related to social/interpersonal processes impacted in suicidality is the construct of help-negation. Help-negation has been defined as an inverse relationship between psychological distress and the intention to seek help for it (Wilson & Deane, 2010). Alarmingly, help-negation has been found most prominent in those at highest risk of suicide (Wilson & Deane, 2010). A number of explanations have been put forward to explain the phenomenon of help-negation as related to suicidality: a perception of being “weak” and the stigma of mental illness (Curtis, 2010), lack of awareness about existing resources, and concerns about confidentiality (Menna & Ruck, 2005) are some hypotheses. The provision of help in an online environment may mitigate some of these help-negating factors. Online Counselling: A Growing Field It has been recognized by clinicians and researchers that the traditional space inhabited by counsellor and client (i.e., the counselling office) has expanded to include the online realm. While varying definitions of “online counselling” exist, I will use the succinct yet comprehensive definition provided by Richards and Vigano (2013): “the delivery of therapeutic interventions in cyberspace where the communication between a trained professional counselor and client(s) is facilitated using computer-mediated communication (CMC) technology, provided as a stand-alone service or as an adjunct to other therapeutic interventions” (p. 994). Distinctions have also been made between synchronous (i.e., chat) and asynchronous (i.e., email) communication, as well as the phenomenon of “chat rooms” peopled by numerous individuals sharing a common goal or issue. Counselling via online video-conferencing has also been increasingly implemented, especially for those in remote locations (Richards & Vigano, 2013). There has been some conflict in the literature with respect to whether online counselling 17 should be considered a mere transposition of face-to-face counselling in a new arena or a distinct type of therapeutic intervention (Richards & Vigano, 2013). Martin and Stuart (2011) describe their new conceptualization of the ‘life space’ inhabited by clients and counsellors as “[…] a unified concept with physical, mental, relational and virtual dimensions” (p. 55) and recognize that the online world is a new arena in which “[…] people interpret and reinterpret, constructing new meanings within their relationships, drawing in their own social and cultural context and shaping the relational dimension of their life-space” (p. 55). They further question the implicit notion that the “virtual” realm is somehow less “real” than traditional life-spaces such as counselling offices. Wrzesien (2014) provided the following illustration of the expansion of the therapeutic space: Figure 1: Expansion of the therapeutic space (Wrzesien, 2014, p. 55). As can be seen from Figure 1, the computer has increased the possibilities of the therapeutic space, expanding it from an office to any space the client chooses. However, the dynamics of the client-counsellor relationship must likely also shift and change to accommodate this new space. Michel Foucault’s (1965, 1973) concept of heterotopia has been invoked in current discussions about online spaces and their impact on the construction of self (e.g., Grbich, 2013). Foucault imagined a heterotopia as functioning like a mirror: parts of the self are reflected back, 18 yet the mirror is, on its own, a real object that modifies the subject’s position by its very presence. Grbich (2013) predicts that particular characteristics of the online space may alter the behaviour of those who inhabit it: “[…] people will use the web like a stage but without the confining parameters of time and space, but a stage where there is no scripted dialogue, unfinished discourse and flexible movement […]” (p. 156). These unique characteristics of online spaces necessarily inform the way in which communication is structured within them and thus have implications for online counselling practice. Online and In-Person Counselling: Clinical Outcomes Online counselling appears to have clinical outcomes that are comparable to those achieved in in-person counselling with respect to the development of rapport (Hanley, 2006; Martin & Stuart, 2011; Reynolds et al., 2011; Roy & Gillett, 2008); in fact, it appears that a positive working alliance can be established in a single session of online counselling (King et al., 2006; Roy & Gillet, 2008) which highlights its applicability for crisis counselling of the type that was examined in this study. However, Fenichel et al. (2002) question the applicability of using an online medium for crisis counselling, citing concerns about counsellor capacity to conduct appropriate assessment. These contradictory views highlight the contentious nature of this area and the need for further research. Most crisis counselling has historically taken place via telephone (Mishara et al., 2007). As telephone counselling shares several important features with online counselling (relative anonymity, the client’s ability to terminate sessions, lack of non-verbal cues, and single-session format), it is relevant to examine what features of these services were considered helpful by clients in the formation of rapport. Based on a meta-analysis by Mishara et al. (2007), telephone crisis counselling clients preferred the anonymity of the service, counsellor attention to 19 intonation and language, and “Rogerian” non-directive responses were considered helpful in reducing symptoms of depression. Despite the literature cited above that suggests that comparable results may be obtained using in-person and online communication, certain characteristics do differentiate the online realm from the physical. As noted above, the relative anonymity of both counsellor and client is one such difference. Additionally, in the case of the online chat service that is the focus of the study, the fact that individuals can choose when they initiate and end the session provides them with a level of autonomy and control that differs from traditional in-person counselling. The reliance solely on written communication may promote cognitive engagement in a way that differs from in-person counselling: Martin and Stuart (2011) note that “writing is a recursive process through which thoughts and ideas are formulated, expressed, organized, and elaborated on” (p. 59). On a related note, the potential accessibility of session transcripts to both client and counsellor can facilitate reflection and insight. Use of therapeutic documents (such as journaling, letters of encouragement, or letters written to lost loved ones) in therapy appears to facilitate client involvement in therapy, promote client autonomy (since they alone are the author of the document), and allow for introspection. Empirical investigation supports the value of therapeutic writing in promoting change; expressive writing has been shown to reduce rumination and depressive symptoms (Gortner, Rude & Pennebaker, 2006). Randomized controlled trials indicate that expressive writing may have physical benefits as well: it has been shown to reduce symptom reporting and lower Impact of Event (EIS) scores in patients with renal cell carcinoma (Milbury et al., 2014). Therefore, online counselling appears to offer unique benefits (anonymity, accessibility, potential for increased disclosure and insight) that make it a worthy area of further research. Martin and Stuart 20 (2011) call for an investigation into how the following key issues in in-person counselling are negotiated in an online realm:  Physical, real-time presence  Immediacy of intervention  Synchronicity and rhythmicity  Reflective moments  Therapeutic use of activities  The nature of the engagement and relationship The final issue identified as necessitating further examination (the nature of the engagement and relationship) could arguably be subsumed in all of the above issues named by Martin and Stuart (2011) and was the focus of the current study. Building rapport, as outlined above, is a vitally important part of the counselling process, and its importance is heightened when the client is suicidal. The increase in use of online crisis services necessitates an examination of how rapport is built online. Hanley (2006) notes the need for further research in this area: “[...] a major question revolves around the ability of counsellors to establish relationships of a sufficient quality to create psychological change online” (p.36). Rapport: Importance of the Therapeutic Connection The important emotional connection between client and counsellor is often referred to as “rapport.” Researchers have frequently acknowledged the reciprocal nature of rapport: rapport is defined by Hackney and Cormier (2009) as “the psychological climate that emerges from interpersonal contact between you and the client” (p. 43). In order to further explore how rapport may be built online, it is important to more thoroughly examine this construct and how it has been studied in order to determine which aspects may be most impacted in online counselling. In his discussion of the “real relationship,” Gelso (2009) elaborates that both magnitude (i.e, weak to strong) and valence (i.e., negative to positive) of the client-counsellor relationship 21 need to be taken into account. Thus, rapport could be described as a strong, positive, reciprocal emotional connection between counsellor and client. However, there is some confusion in the literature between working alliance and rapport. It appears that while there is substantial overlap between the two constructs, there is also a subtle distinction between the two. In order to understand which elements of rapport may be relevant online, it seems prudent to more deeply examine this difference. In their review of the therapeutic alliance in counselling, Horvath and Lester (1993) note that the words therapeutic alliance, working alliance, and helping alliance have often been used interchangeably. Similarly, the literature on the therapeutic relationship implies that rapport and working alliance are often viewed as interchangeable concepts; researchers frequently use measures of working alliance such as the Working Alliance Inventory (WAI) to assess rapport (Sharpley & Ridgway, 1992). Others (Auerbach et al., 2008) have found that rapport is most closely correlated with the bond dimension of working alliance as conceptualized by Bordin (1979). A number of studies (e.g. Efstation, Patton, & Kardash, 1990) have found high correlations between rapport and the bond subscale of the Working Alliance Inventory (WAI) developed by Horvath and Greenberg (1989). Similarly, in a study that aimed to assess the degree of rapport between client and counsellor, the WAI was used as a measure of rapport and researchers observed that the bond dimension of the WAI is “very similar to rapport” […] (Sharpley & Ridgway, 1992, p. 1). In fact, one version of the Working Alliance Inventory (the AWAI-A, Schlosser & Gelso, 2005) has a subscale titled Rapport. Hence, for the purposes of this study, working alliance is considered a broader construct which subsumes rapport. Because this element of the therapeutic relationship seems most difficult to establish in an online realm (as discussed above), rapport was the primary focus of 22 this investigation, with the understanding that other elements of working alliance (e.g., agreement on goals and tasks) are also important, particularly as they relate to suicide risk assessment, and have a strong impact on the establishment of rapport. The real relationship. Gelso (2005) proposes a construct related to but conceptually separate from both rapport and working alliance: the “real relationship.” This has been defined as “the personal relationship that exists between therapist and client from the first moment of contact” (p.640). Gelso distinguished the notion of the real relationship from the concept of working alliance on the grounds that the real relationship is considered to exist independently of the working alliance that is formed over time and is “more basic than the alliance and existing apart from the work of therapy” (Gelso, 2005, p. 640). The real relationship is said to be composed of two main features: genuineness and realism. Genuineness is described as “the ability to be who one truly is, as opposed to being phony or inauthentic” (Gelso, 2009, p. 254), and realism refers to the extent to which client and therapist experience each other “in ways that fit him or her rather than projections based on fears and wishes related to significant others from the past” (Gelso, 2009, p. 254). The real relationship is thus a construct related to rapport but not synonymous with it, and may have particular relevance for online counselling, as described below. The importance of rapport. The clinical and research literature strongly support the assertion that rapport-building is a vital component of the counselling encounter. The presence of rapport has been linked to positive therapeutic outcomes (Hackney & Cormier, 2009; Timulak, 2007). Meta-analyses indicate that the client-counsellor relationship is the single most significant predictor of successful therapy. Specifically, studies have found that it accounts for between 22% (Knaevelsrud & Maercker, 2007) and 30% (Fletcher-Tomenius & Vossler, 2009) of 23 variance in therapeutic outcome, regardless of the practitioner’s theoretical orientation and regardless of whether it was rated by the client, counsellor, or a third party (Hilsenroth, Cromer, & Akerman, 2012). In his review of intra and extratherapeutic indicators of client change, Groth-Marnat (1997) concludes that: “the overall quality of the therapeutic relationship accounts for at least as much of the outcome variance as specific techniques” (p. 596). It should also be noted that there is some dissent in the literature with respect to the role of transference in the therapeutic relationship (Horvath & Luborsky, 1993). Depending on the clinician’s perspective, the client-counsellor relationship may be a pure reflection of the client’s unresolved relational conflicts, a here-and-now relational encounter, or some mixture of the two (Horvath & Luborsky, 1993). For the purposes of this study, I position myself somewhere in between, recognizing that a client’s past relational style and history may interact with the counsellor’s history and style, and that the strength of the impact of either one will vary with each individual. Communication of rapport. Rapport must be communicated by both parties in order to be felt. This may appear self-evident; however, interpersonal interaction takes on additional significance when the communication takes place online, as the absence of non-verbal cues may cause other features of the interaction to take on increased significance (e.g., Feng, Li, & Li 2016). Research on client-counsellor dynamics online appears to support this general observation: Anderson, Ogles, and Patterson (2009) found that the ability of clinicians to understand and relay interpersonal content had a significant impact on online session outcome. Specifically, client outcomes were more positive when therapists exhibited greater interpersonal fluency (i.e., emotional expression, warmth, and persuasiveness). Some of these qualities may not be as easily demonstrated without the benefit of voice tone or facial expression, pointing to a 24 need to examine how text-mediated communication can be amended to attempt to facilitate these important facilitators of rapport. On a related note, how clients perceive the therapist has important implications for how the alliance is formed. For example, Saunders (1999) notes that therapist confidence, focus, involvement, emotional engagement, and positive feelings are key to effective therapy and positive clinical outcomes. Therefore, how language is used online can have specific and unique effects on the therapeutic connection. With respect to rapport as defined above, two dimensions seem particularly salient when conducting counselling online: the degree to which the counsellor is perceived as a real person – what Gelso (2007) would describe as “realism” – and the degree to which the counsellor shares his or her experiences, perceptions, and personhood with the client – what Gelso (2007) would call “genuineness.” These dimensions seem conceptually related to the shift in social power observed in online communication noted by Hanley (2006). In traditional in-person counselling, many aspects of the counsellor’s “realness” are directly accessible to the client: the counsellor is visible, has a known identity (e.g., the client generally knows the first and usually last name of the counsellor and vice versa). Furthermore, both client and counsellor are able to respond to each other’s verbal and non-verbal communicative behaviours. Therefore, in-person counselling, regardless of the therapist’s orientation, could be characterized as being higher on the dimension of “realness,” merely by virtue of existing in the “real world.” Online counselling, on the other hand, especially online crisis counselling, tends to be more anonymous – client and counsellor are often represented by “chat aliases” (as is the case with the online service which was the focus of the current study). Additionally, the interactive cues of voice tone and body language are absent. 25 Since both realness and genuineness seem important components of rapport, and since the “realness” dimension (in other words, the absence of anonymity), seems to be severely compromised in this medium, the second dimension, genuineness, may take on additional importance. Previous research by this author on a similar data set (Timm, 2011) found that a common theme emerged in terms of threats to rapport in online chats: that of the client questioning the crisis counsellor’s humanity – for example, asking “are you a human? Are you a robot?” This finding suggests that the perception of the counsellor as both real and genuine takes on additional importance in the online realm, and that if counsellor realness cannot be verified, counsellor genuineness may emerge as a significant predictor of the establishment of the therapeutic relationship. Building Rapport Online: Challenges and Preliminary Findings Wrzesien et al. (2014) note that the current lack of research in rapport-building online within the counselling domain necessitates drawing from other domains of knowledge. They offer concepts from Computer-Supported Collaborative Learning (CSCL) to help understand how rapport may be built online. For example, they describe the process of communication online (referred to as ‘grounding’ in their research) as composed of two interpersonal elements. The first element entails the content of communication and involves behaviours such as giving feedback or interpersonal cues. The second element references the process of communication, which involves behaviours such as initiating and ending conversations and taking turns (Wrzesien et al., 2014). With respect to conducting therapy online, Wrzesien et al. (2014) suggest that “[...] the quality of grounding would [...] depend on whether the therapist and/or client makes their contribution understandable, allows the partner to ask questions, gives feedback regarding the issue, or shows understanding” (p. 131). When working with suicidal 26 clients, showing understanding and paying attention to interpersonal cues (such as pauses in the dialogue and capitalized letters or punctuation) would likely be important aspects of CSCL to facilitate. Wrzesien (2014) note that […] the technology characteristics such as the quality of graphics, display, or interaction metaphors can influence the sense of being in the anxious situation (presence) or the sense of feeling this situation as real (reality judgement). These in turn have an effect on the reactions of clients (increase in anxiety or decrease in anxiety after a certain time of exposure and, in turn, on the final effectiveness of the therapeutic process (p. 136). Therefore, the context in which the therapeutic interaction occurs (online space as opposed to therapist’s office) necessitates other ways of building rapport with the means at one’s disposal (e.g., emoticons, capitalization, use of punctuation). Wrzesien et al. (2014) suggest that “[…] the interaction with technology by both the therapist and client allows them to create a common ground, which […] contributes to the construction and reinforcement of the therapeutic alliance” (p. 133). The research indicates, however, that subtle distinctions may exist with respect to how rapport may be perceived online versus in-person. Williams et al. (2009) found that online counsellors spent a greater proportion of time on rapport-building than they did on task-oriented activities. This highlights the importance of building rapport online and may suggest that some online counsellors may struggle with doing so. Also, Reynolds et al. (2013) examined the Session Evaluation Questionnaire (SEQ; Stiles, Gordon, & Lani, 2002) scores of clients and therapists who participated in either in-person or online counselling. It was found that SEQ scores were higher for those therapists who provided services online; online clients’ SEQ scores 27 were also either equal to or greater than the scores of clients who had received services in-person. Additionally, the Agnew Relationship Measure (ARM; Agnew-Davies, Stiles, Hardy, Barkam, & Shapiro, 1998) scores of those who had received online counselling were comparable with those who had received face-to-face counselling. However, interestingly, the Arousal component of the SEQ was lower for clients who had received online counselling, and client scores on the Openness dimension of the ARM were also lower in those who received online services (Reynolds et al., 2013). This last finding contradicts other research which indicates that individuals tend to disclose more sensitive information online (e.g., Sade-Beck, 2004), an indication of the many unresolved debates in this emerging field. The lowered arousal scores are explained by invoking the online calming hypothesis, which is described below. Online calming hypothesis. The finding that client scores were lower on the Arousal dimension of the SEQ has been explained by invoking the online calming hypothesis (Reynolds et al., 2013). This hypothesis attributes the lower arousal of clients to the lack of anxiety-producing cues present in the online environment. Reynolds et al. (2013) conclude by suggesting that receiving counselling online may be particularly suited for those struggling with anxiety disorders (in particular, forms of social anxiety) who may otherwise not seek help. Hanley (2006) echoes this idea by suggesting that individuals may seek help online for issues that they may have discomfort disclosing in-person. Interestingly, on this note, it has been proposed that the lack of visual and verbal cues, cited by some critics as a key drawback to online counselling, may actually promote disinhibition – and thus more open self-disclosure – as it removes the salience of the other person’s reaction to their narrative (Richards & Vigano, 2013). This may potentially increase their level of comfort with the interaction. These projected 28 benefits of online counselling may be relevant to suicidal individuals who may experience social anxiety and fear of disclosing suicidality in-person due to concerns about being perceived negatively. Rapport-building online: Suggestions from the research. An obvious interactive cue that is missing in online counselling is the presence of eye contact, which has been linked to rapport, perceptions of empathy, and other positive therapeutic outcomes (Dowell & Berman, 2013). Therefore, other ways of building a therapeutic connection must be found. Preliminary research indicates that use of emoticons to enhance communication and attendance to potential technological glitches such as “freezing” of computers were found to be important features to attend to when establishing rapport online in a study of 15 users of an online counselling service for youth (Hanley, 2006). On a related note, Hanley (2006) found that level of skill using “netiquette” (adhering to norms of online communication) may impact rapport in online counselling. Mitchell and Murphy (1998) have suggested that use of abbreviations and “bracketing” emotional reflections are possible ways in which to communicate understanding. As noted by Wrzesien (2014), the online space creates a potential “common ground” for client and counsellor to interact in, potentially creating opportunities for re-imagining the traditional client/counsellor relationship. The therapeutic interaction between client and counsellor may thus approach a more egalitarian model online. Hanley (2006) notes that due to the anonymous nature of the interaction, the traditional discourse of the counsellor possessing greater social power may shift. He suggests that “clients may therefore feel more confident to openly challenge decisions made by the counsellor” (p. 41). This shift in social power and more informal communicative style, in addition to the lack of verbal and visual cues, may cause discomfort in some counsellors providing services online (Richards & Vigano, 2013), which 29 speaks to the importance of investigating the efficacy of training practices regarding how rapport may be built online. To summarize, the research indicates that when conducting counselling online, familiarity with online linguistic conventions, attention to potentially shifting social power, and recognition of the possibility of increased counsellor self-disclosure are all areas that merit particular focus. Research Questions Rapport, as noted above, has been found to facilitate positive therapeutic outcomes regardless of the therapeutic orientation of the practitioner or the presenting issue of the client. Larsson et al. (2012) note that “[...] the therapeutic relationship is one of the most pertinent discourses in counselling psychology literature and talk [...] and often [...] used as a raison d’etre for counselling psychology and something that makes it unique, while at the same time positioning itself as different from other fields of applied psychology, such as clinical psychology” (p. 37). Studies of therapeutic outcome frequently note that the common factors underlying a person-centered approach (e.g., empathy, client/counsellor relationship variables) are more highly correlated with client change than are specific treatment interventions (Lambert & Barley, 2001), and meta-analyses of client-perceived helpful events frequently invoke aspects of the therapeutic relationship such as experiencing safety, feeling understood, and experiencing the counsellor as a fellow human being (Timulak, 2007). The mechanisms by which this change comes about include facilitation of client insight, catharsis, and the witnessing of one’s suffering by an emotionally attuned other (e.g., Rogers, 1957). Reciprocal communication, in itself, has been discussed as potentially facilitative of client change (e.g., Westwood & Ishiyama, 1990); thus the ways in which communication may facilitate rapport online seems an important avenue to explore. 30 As noted above, my Master’s thesis (Timm, 2011) focussed on how crisis counsellors (volunteers at a local crisis centre) built rapport online with suicidal youth. The finding that counsellor genuineness emerged as a key factor in maintaining rapport is consistent with the literature reviewed above which indicates that building a connection and communicating one’s humanness is especially important in the online world. Furthermore, the stigma of suicidality and the high-stakes nature of the therapeutic conversation necessitate an examination of how rapport is built and maintained online. This need is highlighted by the lack of research in this specific area as outlined above. As discussed, the lack of research in the area of addressing suicidality online can be attributed to ethical concerns about collecting data on those who are suicidal online, lack of a reliable and valid measure of rapport for online counselling, difficulty accessing data from crisis lines, and the relative newness of this modality of delivering counselling services. I believe I was fortunate to be well-positioned (via my involvement with an online suicide prevention service) to conduct this research, as I was able to address many of the above barriers to conducting research on online crisis counselling. The investigation described below was conducted with the aim to extend and deepen the preliminary findings of my Master’s thesis in three main ways: a. Since adults are increasingly communicating online, adding data from the newly launched crisis centre chat for adults b. Rather than focusing solely on counsellor behaviours, expanding the data analysis to include ways in which the clients spoke about and experienced rapport c. Chats defined as “non-rapport-containing” (not containing rapport, as per the operational definition outlined below) were included for comparative purposes Two overarching research questions that informed the data analysis are stated below: Q1. How is the client-counsellor relationship negotiated by the crisis counsellor in an online chat where the client (youth or adult) is suicidal? 31 Q 2. How is the client-counsellor relationship negotiated by the client (youth or adult) in an online chat where the client is suicidal? Following the procedure used in my Master’s thesis (Timm, 2011), each transcript was separated into three sections, Initial Contact, Suicide Assessment, and Termination. Based on these, the following specific research questions guided the analysis in each section: Q3. How is the client/counsellor relationship negotiated in the initial contact stage of the chat? Q4. How is the client/counsellor relationship negotiated during and following disclosure of suicidality? Q5. How is the client-counsellor relationship negotiated when the chat is ending? Q6. If threats to rapport occur, how does the counsellor relate to the client to maintain or re-establish rapport, and what is the impact on the client? Threats to rapport were defined as any behaviour of client or counsellor that threatened existing rapport or hindered the development of rapport, as identified above. For example, the client’s trust in the therapeutic process could be threatened by the counsellor’s refusal to answer direct questions, or the bond between client and counsellor could be threatened by the client’s expression of dissatisfaction with the service provided. Site of the Study: Vancouver Crisis Centre The Vancouver Crisis Centre was the site at which the research was conducted, as noted above. Volunteers at the Vancouver Crisis Centre (hereafter referred to as “crisis counsellors”), are paraprofessionals who receive 102 hours of basic training in crisis intervention and suicide prevention on both the online and telephonic services. They all possess the equivalent of a high school diploma; many possess at least a bachelor’s degree, and many are pursuing entry into graduate school in the helping professions. They are not professional counsellors and most do not possess a master’s degree in counselling psychology. Volunteers are provided with on-site supervision and 24/7 on-call support. Staff (who often possess at least a Master’s degree in a 32 related profession) are available for debriefing and provide support and guidance as needed, and assist in the case of an emergency intervention. The online services offered by the Vancouver Crisis Centre, www.youthinbc.com and www.crisiscentrechat.ca, offer crisis counselling in real-time using a text-based interface using Sitemax software. This platform allows the use of emotiocons and also permits the sending of links between users. The youthinbc.com site is distinguished from the site for adults by the inclusion of more youth-specific resources (e.g. regarding bullying), and by its youth-oriented graphics. 33 3. METHOD Theoretical Assumptions and Rationale It is recognized that a large body of research on suicide prevention, including much of the research reviewed in Chapter 1, has been conducted within a post-positivist framework, including correlational research regarding risk and protective factors, use of post-hoc surveys, and Likert scales assessing various constructs. While the nature of the questions asked in the field of suicidology often necessitates a quantitative approach (e.g., identifying risk and protective factors, designing assessment tools), and while this body of research has contributed valuable information to the study of suicide, a quantitative approach also suffers from drawbacks. Namely, because quantitative research can be less detailed and contextual, important information may be missed (Kouri & White, 2014). Furthermore, the often-endorsed assumption in quantitative research that the sample used is representative of the population may be overstated in these studies, considering the low base rate of suicidal behaviour and the difficulty obtaining data from this population, pointing to the potential utility of smaller, more detailed exploratory studies. On this note, for ethical reasons, suicide research is often conducted in simulated artificial environments, and the pre-set answers on questionnaires may not capture the complexity of the experience of someone contemplating suicide. It has also been argued that the complex and unpredictable nature of suicide itself lends itself well to indepth qualitative exploration. Borrowing from Rittell and Weber (1973), policy analysts who distinguished between “tame” and “wicked” problems, Jennifer White (2012) distinguishes between “tame” and “wild” problems, describing “tame” problems as those that are static (unchanging), have a singular meaning and clearly identifiable causes, and are thus linked to linear solutions. These problems, she suggests, can be effectively studied using a post-positivist framework. However, White (2012) proposes that suicide (in particular, youth suicide, 34 her area of focus) is a “wild” problem that has been studied as though it were a tame one, and that this has limited the range of solutions identified. She references the complexity, multiplicity of causes, and importance of context in situating youth suicide as a “wild” problem. This suggestion is consistent with the developmental systems theory of suicide that “implies there is a story behind each suicidal action and not a simple cause” (Valach, Young, & Michel, 2011, p. 137). I would like to propose that the tame/wild distinction could be viewed as a continuum rather than a dichotomy, and that the building of rapport in online suicide prevention counselling could also be characterized as on the “wild” problem end of the spectrum – in addition to possessing the characteristics of youth suicide outlined above, it is also a new modality that can benefit from an exploratory analysis and a detailed, contextual exploration. The nature of the data obtained from examination of issues such as online suicide intervention is complex, multi-faceted, and often context-specific, especially considering the biopsychosocial correlates of suicidality outlined above. It has been frequently noted that the complexity of counselling practice cannot fully be captured by employing quantitative methods. For example, Woolsey (1986) notes that “[…] traditional methods of research that focus on quantification and experimentation are inadequate to meet the challenges of counselling practice” (p. 242). The presence of “outliers”, often discounted in quantitative research, may provide rich sources of information about the specific lived contexts of suicidal individuals. For example, a strong negative reaction to a given online counselling interaction may facilitate additional research into what factors may prevent the establishment of a therapeutic alliance online. Furthermore, the newness of the online counselling modality and the lack of research 35 examining this specific area, (as detailed in the literature review above) precludes the adoption of explicit a priori assumptions which may limit the finding of potentially useful information. Epistemological Stance: Social Constructionism Qualitative research, informed by social constructionism, acknowledges and attempts to delineate any potential researcher assumptions, worldview, or bias (Grbich, 2013). On this note, it is acknowledged that, due to my previous research and role, I held certain implicit assumptions that needed to be “bracketed” prior to my data collection (e.g., Flick, 2006; Grbich, 2013) which are detailed below. It is recognized that the data were viewed through my subjective lens and that the picture that emerged of each therapeutic conversation would inevitably differ from that recorded by another researcher. Therefore, the methodology described below was informed by a social constructionist perspective. Social constructionism has been described by Andrews (2013) as “[…] essentially an anti-realist, relativist stance [that] places great emphasis on everyday interactions between people and how they use language to construct their reality” (p. 2). This framework assumes that researchers’ worldviews or background may shape their expectations, which was certainly the case for myself in this study. Being exposed to both the literature on the topic of suicide and having worked with youth in crisis for a number of years, I inevitably brought certain expectations and a priori beliefs to my research. These beliefs, and my methods of reflexively acknowledging them throughout the research process, are detailed below in the section entitled “Researcher Reflexivity.” Working within the epistemological framework outlined above, I operated from an understanding that there may be multiple realities in terms of the definitions of the constructs studied (suicide, rapport, online counselling), and that both myself and the sources of data (crisis line works and crisis line clients) had, through our interaction with the world, created our own 36 beliefs about the constructs “suicide” and “rapport.” Furthermore, it was assumed that these constructions had arisen mostly due to the society in which we were born and the associated norms and beliefs that surrounded us. Therefore, the epistemology that informed this perspective was that of social constructionism (Flick, 2006) and the research and its findings are situated within a social-constructivist epistemological paradigm. On this note, it is recognized that the findings of this study serve as “[…] a narrow illumination of the chosen topic – a constructed reflection which is time-and context-bound […] a truth limited by the constructions and interpretations of both researcher and researched” (Grbich, 2013, p. 112). Therefore, the findings were not generalized to a broader context beyond the one under investigation. Trustworthiness In qualitative research, “trustworthiness” describes what would be referred to as the reliability of the findings in post-positivist terms (Tobin & Begley, 2004). However, the social-constructionist perspective that informs my research implies that there is no absolute reality that is not socially constructed. Therefore, attempts were made to ensure trustworthiness from this social-constructionist perspective. Trustworthiness when viewed from this perspective is generally broken down into four categories: credibility, transferability, dependability, and confirmability (Tobin & Begley, 2004). To address credibility (the degree to which the presented information is a “credible” representation of the raw data), I ensured that the transcripts were viewed independently through the lenses of multiple trained individuals (one trained research assistant as well as my research supervisor and two consultants who were experts in the area of discourse analysis). Credibility of observations was determined by using the method suggested by DeCuir-Gunby et al. (2011): dividing the number of agreements on codes by the total number 37 of agreements and disagreements. However, since the aim of qualitative research is not to seek results that are generalizable to a larger population but to offer a rich and nuanced picture of the phenomenon under investigation, disagreements between raters are included and discussed. Creswell (2009) observes that one of the ways to add rigour to qualitative research is to “provide detailed descriptions of the setting […]” since this can help the ensuing results “become more realistic and richer” (Creswell, 2009, p.192). When presenting the results of this study, it is recognized that provision of “thick description” to facilitate a rich picture for the reader must be balanced with respect for participants’ confidentiality. To address transferability (the ease with which other researchers may be able to replicate or extend this study), I include examples of my coding system and attempt to describe the procedures used to analyze the data in as much detail as possible, using “thick description” balanced with the need for confidentiality, as noted above. With respect to both credibility (as described above) and confirmability (acknowledgement that the research is never objective, and striving to reflect the research context as accurately as possible), I brought questions regarding the study’s design and the processes of data collection to my research supervisor and two experts in qualitative data analysis to ensure that the quality of the processes of data collection was sound and that the presented findings represented the actual data. Furthermore, I strove to ensure that my own biases were clearly delineated and engaged in reflexivity with respect to my engagement with the data (Grbich, 2013). Ways in which this was done are outlined below. Researcher Reflexivity Grbich (2013) notes that when working from a social constructionist perspective, it is important to consider the multiple researcher “selves” that operate in relation to the data. She notes the potential presence of “[…] the central historically constructed self, the self that is currently undergoing change and […] the reflexive observer of this process” (p.113). Throughout 38 the process of data analysis, attention was paid to which “self” was operating in response to the data and how this potentially impacted its interpretation. Researcher reflexivity was cultivated in the following ways, following suggestions outlined by Grbich (2013):  Examining and disclosing my biases and/or a priori assumptions as well as acknowledging the sources of information that may have influenced analysis of the data  Seeking outside sources (research assistant; faculty members; supervisor) to check coding  Acknowledging ambiguity and resisting the temptation for closure/black and white observations  Engaging in self-reflective practices throughout data collection (e.g., journaling; dialogue with colleagues) On this note, I recognize that I brought to this research the assumption that the constructs of “suicidality” and “rapport” – as defined by myself and by those in crisis who are the focus of this study - arose out of our respective interactions with the world. In addition, the ways in which the data were collected imply certain assumptions on my part; namely: a. All communication is relational b. The presence of rapport as defined in this study was a necessary condition for a chat to be deemed “successful.” c. My operational definition of rapport and the measures used encapsulate the experience of rapport built online with suicidal clients d. In most cases, suicide is an event that needs to be prevented Furthermore, it was recognized that the data were viewed through my subjective lens and that the picture that emerged of each case would inevitably differ from that recorded by another researcher. Following this theoretical framework and abiding by the assumptions outlined above, data were collected in the following way. Participants (Data Sources) Little research has been conducted using transcripts of actual crisis chats, which, due to their archival nature, have the benefit of being uninfluenced by the presence of the researcher at 39 the time of their creation. They represent, in real time, the experience of suicidality in purely linguistic and paralinguistic terms. My data sources consisted of 24 transcripts of chats obtained from the Vancouver Crisis Centre’s www.youthinbc.com website (for youth) and from www.crisiscentrechat.ca (for adults). The two services are comparable in function; however, the youthinbc service is geared towards youth (defined as those aged older than 16 but younger than 19; however, see exclusion criteria below). Sixteen chats from each service were classified as “rapport-containing” chats (client/counsellor rapport is established and client reported satisfaction with the encounter); eight chats from each category were classified as “non-rapport-containing” (rapport was not present and client reported dissatisfaction with the encounter). Of the rapport-containing chats, nine were obtained from the chat service for adults; seven were obtained from the chat service for youth. Of the non-rapport containing chats, four were obtained from the chat service for youth; four from the chat service for adults. It was understood that working from a social constructionist framework, the research design might evolve and be amended as new information emerged. This did occur in response to the characteristics of the data obtained. For example, it was initially thought that the two services (adult and youth) would be compared with respect to how rapport was built with youth versus adults. However, upon examination of the data, it was discovered that the themes observed were found across all ages of clients. Furthermore, an individual’s choice of service was sometimes inconsistent with their age (adults would sometimes use the youth service and vice versa), and sometimes the distinction between adults and youth seemed arbitrary (i.e., a 19-year-old seeking services on the youthinbc service likely considered themselves a “youth” but would be classified as an adult as per the law). Hence, it was deemed more important to focus on how rapport was built and less important to focus on the age of the client. However, to increase the 40 trustworthiness of this study by providing a “rich description” of a given content domain, and to delineate between youth and children (who would likely have different developmental needs which might impact rapport), “youth” was defined as a person aged over age 16 but under age 19 as per the provincial government of British Columbia (Government of British Columbia, 2016). The inclusion criteria were as follows: only electronically recorded internet-based chats with “suicide” as a problem code (chats are coded in this way if there is a suicide assessment undertaken by the crisis centre volunteer) were included in the analysis. Since the aims of qualitative research are exploratory and descriptive rather than generalizable (Grbich, 2013; Woolsey, 1986), attempts were made to include a diverse array of transcripts with respect to chatter age, gender, and other demographic variables, with the aim of providing “[…] complete coverage of the content domain” (Woolsey, 1986, p. 245). Furthermore, an additional consideration related to diversity was the requirement that no two chats from the same client (identified via IP address) or the same counsellor (identified via their alias) were included. Tables 2 and 3, below, outline the demographic information obtained. It is notable that, as Tables 2 and 3 illustrate, of the 24 chats obtained, only 15 disclosed their gender, whereas all disclosed their age. It is further important to mention that the characteristics of the data (archival transcripts of online crisis chats) preclude the verification of any demographic data. Thus, the information provided below is solely based on self-report. With respect to the category of “primary issues,” these are labeled by the Vancouver Crisis Centre volunteer at the conclusion of each chat. In this case, each chat had already been labeled “suicide ideation” as per the inclusion criteria; however, in order to create a more nuanced picture of the chat transcripts, it was decided to include the secondary issue other than suicide ideation that was experienced by the client. These were concurrently labeled by myself and research assistant, and 41 no disagreements were found. In one chat (chat #19), there was no presenting concern other than suicide ideation found; this may be due in part to the shorter duration of this chat. As detailed below, ages of chatters ranged from 17 to 48, with a mean age of 23. Presenting concerns were diverse, with depression being the most commonly cited, followed by relationship concerns of some kind (bullying; peer conflict; family conflict; relationship conflict). Chatters identified as being from diverse geographic locations across BC. The most frequently cited locations were Vancouver (four chats) and Vancouver Island (three chats) with four chatters simply stating their location as “BC” and one as “planet earth.” Some remote, northern communities were represented (Haida Gwaii and Prince Rupert). Six chatters did not disclose their location. Chats were received from nine chatters who identified as female, seven as male, and nine who did not disclose their gender. Length of chats ranged from 17 minutes to 102 minutes with a mean time of 57.378 minutes in the rapport-containing category, and a mean of 38.625 minutes in the non-rapport containing category. The means are reported for the separate categories because of the notably shorter length of the non-rapport-containing chats, which was attributed to chatter dissatisfaction and subsequent premature termination of the chat. It must also be noted that length of chats was not necessarily synonymous with amount of client-counsellor interaction, as some chats containing significant gaps in conversation where the client was otherwise engaged and therefore responses were slower. Counsellor gender is not reported, as volunteer aliases at the Crisis Centre are gender-neutral in order to prevent client requests for a specific gender, and due to the way in which data were collected (to protect the anonymity of the volunteers in light of my role at the Centre), I was not able to have access to counsellor gender. 42 Table 2 Demographic Information: Rapport-Containing Chats Chat # Client age* Client gender* Client location* Chat length Secondary issue(s) Chat 1 20 Female Vancouver 62 minutes eating disorder Chat 2 18 Female Port McNeil 102 minutes bullying Chat 3 21 Male Prince George 59 minutes depression Chat 4 30 Undisclosed Undisclosed 71 minutes panic attacks Chat 5 48 Female Haida Gwaii 52 minutes depression Chat 6 20 Male Planet Earth 63 minutes schizophrenia Chat 7 18 Undisclosed Surrey 48 minutes family conflict Chat 8 23 Female BC 54 minutes social anxiety Chat 9 18 Female Vancouver 62 minutes depression Chat 10 22 Undisclosed Vancouver 47 minuttes depression Chat 11 26 Male North Island 45 minutes peer conflict Chat 12 19 Undisclosed BC 61 minutes social isolation Chat 13 18 Female Vernon 41 minutes depression Chat 14 18 Female Vancouver 47 minutes depression Chat 15 35 Undisclosed BC 51 minutes trauma Chat 16 23 Male Undisclosed 53 minutes depression * chatter-identified 43 Table 3 Demographic information: Non-rapport-containing chats Chat # Client age* Client gender* Client location* Chat length Secondary issue(s) Chat 17 18 Female Smithers 33 minutes depression Chat 18 29 Undisclosed Victoria 88 minutes workplace stress Chat 19 30 Female Canada 40 minutes suicide ideation Chat 20 17 Male BC 17 minutes anxiety Chat 21 29 Male Undisclosed 24 minutes depression/housing Chat 22 19 Undisclosed Undisclosed 32 minutes relationship conflict Chat 23 22 Undisclosed Undisclosed 27 minutes depression Chat 24 22 Male Undisclosed 48 minutes self-harm * chatter-identified Rapport-Containing versus Non-Rapport-Containing Chats The procedures for distinguishing rapport-containing chats from non-rapport-containing ones are now outlined. Practical considerations related to the online nature of the interaction prevent researchers from being able to determine chat “success” in preventing suicide. As stark as this may seem, the primary determinant of chat success in this study was thus not suicide prevention itself but the etablishment of a strong rapport between client and counsellor. Considering the stigma of suicide and the often intense loneliness and lack of connection cited in many suicide prevention models as key predictors of suicide risk (e.g., Durkheim, 1951; Joiner, 2005; Van Orden et al., 2010), being able to establish rapport with one’s counsellor online assumes additional importance. Thus, the attainment of rapport in online suicide counselling was 44 considered a worthy goal and therefore served as a necessary and sufficient condition for a chat to be deemed “successful” in this study. Identification of Rapport-Containing Chats In order to distinguish between rapport-containing and non-rapport-containing chats in this study a number of measures of rapport were used. Prior to this, the construct of rapport was operationalized. Operationalization in qualitative research could be described as the process of attempting to capture a phenomenon that may not be directly measurable, although its existence is signified by the presence of other, often related, phenomena (Grbich, 2013). It involves delineating the dimensions and limits of a construct, specifying, often in behavioural terms, what the construct is and what it is not. However, it is important to acknowledge that, working from a social constructionist perspective as described above, the construct of rapport is viewed as socially constructed and amenable to change through social discourse. My Master’s thesis (Timm, 2011) involved the creation of an operational definition of rapport for online counselling that was developed through consultation with colleagues and research assistants and is supported by current research in the domains of online counselling, suicide intervention, and youth counselling. Specifically, considering the nomological network of rapport, it is reasonable to assume that when assessing the convergent validity of a given measure of this construct, moderate correlations would exist between measures of rapport and measures of trust, real relationship, and working alliance, and this seems to be the case based on the research cited above. Therefore, the operational definition used included these constructs. As the data set had similar characteristics as the one used in my Master’s thesis (Timm, 2011) with the exception of the age of the participants studied (chats from adults were included in the present study), and since there was no operational definition of rapport applicable for use 45 with online counselling with suicidal individuals, the operational definition of rapport used in my Master’s thesis was used in the current study, with a number of additional measures of rapport added to ensure trustworthiness of the results. Based on the existing research and nomological networks of rapport, outlined at length in my Master’s thesis (Timm, 2011), the following definition of rapport was used: “The existence of a genuine emotional connection between client and counsellor in which a climate of mutual trust, collaboration, and understanding is created, as operationalized by verbal statements or actions made by the client that imply the existence of such a connection.” As noted above, my Master’s thesis (Timm, 2011) explored how rapport was built online in the unique context of a crisis chat; this was undertaken in response to the finding that existing measures of rapport were insufficient to capture all salient aspects of rapport in an online environment due to both the online and short-term nature of crisis chats which rendered existing measures insufficient for use on their own. This research (Timm, 2011) found that four clear, discrete categories of client responses indicating rapport were consistently highlighted by the primary researchers and the research assistants. Therefore, the following criteria were sought when determining whether a chat was “rapport-containing” (the presence of more than one category was required for inclusion). These were: (1) Expresses clear agreement with counsellor’s statements (2) Collaborates with risk assessment and safety planning (3) Openly discloses suicidality/sensitive information (4) Expresses liking of/connection with counsellor Other measures of rapport. In order to increase the trustworthiness of the findings, multiple measures of rapport were used, with the understanding that, as noted above, some 46 measures, having been developed for in-person counselling, may not be fully appropriate for evaluating online crisis counselling, based on both the nature (in-person versus online) and duration (multiple sessions versus a single session). The measures used were the Working Alliance Inventory – Observer Form (Archuck, Wang, Weibel, Fende, Anderson, & Horvath, 2000), and the Kids Helpline Online Counselling Transcript Coding Instrument (Williams, 2009). A second rationale for the use of multiple measures was to assess the potential concurrent validity of the measure developed in my Master’s program. A research assistant was employed who rated chats independently. Chats were considered “rapport-containing” if they had a minimum score of 30 or higher on the KHLTCI (obtained by receiving a mean score of 5 on a seven-item 10-point Likert-type scale) and a score higher than 4 on the WAI (as per the developers of the WAI, a score of 7 indicates an “ideal alliance” and a score of 4 indicates an “average alliance”). They were considered “non-rapport-containing” if they had a score of lower than 30 on the WAI and a score lower than 4 on the KHLCI. Initially, it was proposed to include only chats with one or less of the four categories of rapport-containing statements found in my Master’s thesis (Timm, 2011). However, once coding of chats began, it was observed that category 3 (openly discloses suicidality/sensitive information) appeared in all but one chat. Considering that the chats came from a suicide prevention service, openness about suicidality was not considered necessarily indicative of rapport (although in one chat, Chat #22, the chatter was overtly hesitant in disclosing information). Therefore, the decision was made to include chats containing Category 3 responses and no more than one other additional category of response in the delineation of non-rapport-containing chats. The number of ratings in agreement was calculated, as was the total number of ratings. This fraction was then expressed as a percentage. A minimum 75% agreement 47 was considered sufficient as per established guidelines of inter-rater agreement with two raters using numerical data with four or fewer categories (Wongpakaran et al., 2013). In cases where agreement was not reached on a given category, the disagreement was discussed. An inter-rater agreement of 89% was initially found; discussions were held until consensus was reached on all measures of rapport. The scores on each instrument for the rapport-containing and non-rapport-containing chats are noted below: the chats containing rapport are noted in blue; the chats not containing rapport are noted in yellow. Table 4 Rapport Measures Chat number WAI Score KHLTCI score Timm 4 conditions* 1 4.36 42 1,2,3,4, 2 4.52 43 2,3,4 3 5.80 55 1,2,3,4, 4 5.11 31 1,2,3,4 5 5.16 43 1,2,3,4 6 5.88 54 1,3,4 7 5.63 53 1,3,4 8 4.61 31 1,2,3,4 9 4.41 33 1,2,3 10 4.41 30 1.2.3 11 4.52 35 1,2,3 12 5.14 32 1,2,3 13 5.04 41 1,2,3,4 14 4.16 32 1,2,3 15 5.12 37 1,2,3,4 16 4.17 31 1,2,3 17 3.72 21 1,3 18 3.75 23 2,3, 19 2.05 2 3 20 2.38 6 3 21 2.13 0 3 22 2.44 1 None 23 3.36 23 1,3 24 2.60 2 3 *Condition 1: Expresses clear agreement with counsellor’s statements Condition 2: Collaborates with risk assessment and safety planning Condition 3: Openly discloses suicidality/sensitive information Condition 4: Expresses liking of/connection with counsellor 48 Satisfaction scale. In addition to rapport, which is a researcher-generated theoretical construct, whether or not the client expressed satisfaction with the encounter was an additional criterion for a rapport-containing chat. This added the client’s explicit expression of positive emotion regarding the therapeutic encounter and thus provided additional trustworthiness to the chat selection process. Hackney and Cormier (2009) noted that one of the most reliable predictors of a strong client-counsellor relationship is client-expressed satisfaction with the relationship. Thus, chatter satisfaction was used as a criterion for determining the presence or absence of rapport. “Chatter satisfaction” was quantified in the following manner: All calls or chats are documented by crisis line workers as per Crisis Centre protocol. At the conclusion of each call or chat, crisis line workers are required to assess the client’s satisfaction with the encounter on a 5-point scale. This is a behaviourally-based scale with 5 indicating the highest level of satisfaction and 1 indicating the lowest level of satisfaction (see Appendix B). This scale was developed by the Vancouver Crisis Centre and fulfills the professional standards set by the American Association of Suicidology, the accrediting body of the Centre (Vancouver Crisis Centre, 2010). To avoid ambiguity and increase objectivity, short behavioural definitions have been provided by the Vancouver Crisis Centre describing each level of the scale (see Appendix B). In order to increase the trustworthiness of the findings (Grbich, 2013; Morrow, 2005), a trained research assistant (described below) assessed satisfaction ratings independently, based on the chat transcript, as it was thought that some counsellors might over or underreport satisfaction based on individual counsellor characteristics (for example, newer crisis counsellors may have been more concerned with being perceived as effective and therefore may have assigned themselves higher satisfaction ratings). However, the possibility of this rating inflation was 49 considered lower than when rating crisis phone-calls, since volunteers using the online services are aware that staff have access to full transcripts of their sessions; hence, they were considered less likely to rate their chats in a way that is inconsistent with actual content of the chat. When there was disagreement between satisfaction ratings, this disagreement was discussed and the statements re-rated by all raters. The procedure for achieving inter-rater agreement was the same that undertaken when determining the presence or absence of rapport, described above. The inter-rater agreement prior to discussion was 91%. If agreement was not reached, then the chat in question was discarded (this occurred in one instance). Chats rated 4 or 5 were considered rapport-containing; chats rated 1 or 2 were considered non-rapport-containing. Rationale for Size of Data Set Considering the epistemological stance of social constructionism and the adoption of a qualitative research method, it was recognized that the aims this qualitative research study were idiographic and emic – seeking depth rather than breadth (Morrow, 2005). Grbich (2013) notes that “[…] if the assumption [is] that truth is multifaceted, that reality is multiply constructed and that large-scale research which homogenizes differences is no longer appropriate […] we fall back on smaller-scale depth research with multiple data sources and individual narratives” (p. 112). Therefore, rather than seeking an extensive data set, a smaller-scale data set of 24 transcripts was examined in depth, as noted above. In addition to rapport and satisfaction as defined above, inclusion criteria were as follows: (1) Only chat transcripts with Contemplating Suicide as a problem code were selected (chats are identified in this way if there is a suicide assessment undertaken by the crisis centre volunteer). (2) With respect to chat duration, a timeframe of 40-90 minutes was initially considered. Since the average chat received by the Crisis Centre is 67 minutes long (Vancouver Crisis Centre, 2010), this was considered a timeframe that represents an average chat and was thought to be long enough to capture the start and potential resolution of a suicidal crisis. 50 However, due to the finding that “unsuccessful” chats were often ended prematurely (due to expressed dissatisfaction by the client), this criterion was further amended and the duration of non-rapport-containing chats was shortened to 20-60 minutes, since the longer timeframe did not result in a sufficient number of chats. This criterion was amended again in response to the nature of the data – as observed above, chat length was not necessarily an indication of amount of interaction present, permitting the inclusion of one chat lasting 17 minutes, which was considered by myself, the research assistant, and the data analysis consultants to contain enough material (client and counsellor responses) to be included, and a concurrent loosening of the criteria to reflect amount of data versus chat length. (3) On December 24, 2009, the Vancouver Crisis Centre placed a disclaimer on the youthinbc.com website indicating to youth that their (non-identifying) information may be used for research purposes (see Appendix C). Only chats received after this date were included in the analysis. A similar disclaimer was placed on the adult crisis chat service on May 13th, 2014. (4) Only chats where volunteers had consented to participate via signing of a consent form (see Appendix A) were included. (5) After data collection had begun, it was determined that chats received prior to 2014 were not to be examined due to difficulties obtaining volunteer consent from volunteers who had since left the Centre. Delineation of Phases Following the procedures operationalized in my Master’s thesis (Timm, 2011), each transcript was separated into three phases: Initial Contact, Suicide Assessment, and Termination. The delineation of these phases was hoped to increase the trustworthiness of the results and to illustrate to the temporal progression of client and counsellor tasks in the process of suicide risk assessment. Based on the existing research on suicide assessment (e.g., Granello 2010; Shea, 2008), it was expected (and found) that each “phase” would involve distinct relational goals and needs for both counsellor and client, and thus different phases were expected to involve different ways of building rapport. As was proposed and carried out in Master’s thesis research, the Initial Contact phase was determined to start from the crisis centre volunteer greeting the client and to end with the disclosure of suicidality. The Suicide Assessment phase was determined to begin with the volunteer’s asking the question “are you suicidal?” or its equivalent and to end with the 51 client or volunteer indicating that the chat would end soon, for example by referring to activities to be done after the chat; the Termination phase was determined to begin at this point. Data Analysis Stage one: content analysis. Transcript analysis was conducted using an inductive rather than deductive approach; starting from more detailed observations, then examining patterns, and lastly putting forward tentative hypotheses about the observations. Chat transcripts were examined first on a micro-level (examining the content of client and counsellor statements), then on a macro-level (examination of themes). The examination of themes and patterns in the data set was done in a systematic manner informed by research that has been conducted on similar data sets. Stake (1995) encourages the development of “foreshadowing questions” which he emphasizes may be changed as the research progresses. For the purposes of this study, the following guiding questions were asked of each dyadic interaction (“unit of meaning”, as described below): Q1. How does the client relate to the counsellor in this segment? Q2. How does the counsellor relate to the client in this segment? Q3. When threats to rapport occur, how are they managed by the client? The counsellor? It was understood that the above guiding questions assumed the existence of a priori themes. Ryan and Bernard (2003) note that “even with a fixed set of open-ended questions, one cannot anticipate all the themes that arise before analyzing the data” (p. 88). Therefore, I sought a combination of a priori and inductively derived themes in this research, and amended or eliminated themes as dictated by the emergence of the data. The goal of my data analysis process was to move reflexively in an iterative style between enumerative data, emerging contextual 52 themes, and the existing literature on rapport-building, online counselling, and suicidality. Based on the characteristics of the data, these guiding questions were amended to be more general in order to facilitate a more open-ended initial exploration of the data. Therefore, the following question was asked of each interactive segment: Q1R. What is happening between client and counsellor in this segment (meaning unit)? It was recognized that a reflexive approach does not begin with deferring to the literature; therefore, this was done only once raw data had been analyzed for the existence of researcher-generated themes as described below. Ryan and Bernard (2003) point out that “in theme discovery, more is better” (p.103), referring to the often exploratory nature of qualitative research and the utility of gaining a rich pool of themes to draw from in the initial analysis. Therefore, to increase the likelihood of finding relevant themes, the answers to the preceding question were examined using multiple methods for identifying themes; these are outlined below. The six categoies of code components referenced by DeCuir-Gunby (2011) were included in the coding process: code name/label, brief definition, full definition, inclusion and exclusion criteria, and an example of a given code. Following open coding (the creation of new codes using raw data), axial coding was undertaken (identifiying any connections between codes). A number of related scrutiny techniques (Ryan & Bernard, 2003) were used in this process. The meaning units were initially sorted using the “block and file” approach (Grbich, 2013), grouping related responses under broad labels. Client and counsellor responses were kept intact to avoid missing important information. Coding was done by “units of meaning” (blocks of related content) rather than line-by-line, as this was considered more meaningful with respect 53 to the nature of the data (DeCuir-Gunby, 2011). Units of meaning, or meaning units, were defined in a manner similar to how they are described by Wong et al. (2013). A meaning unit was considered as such if it was characterized by a clear, discrete idea and a transition in meaning from the previous meaning unit. For this study meaning units were considered relevant if they related to the therapeutic relationship as per the discourse analysis “tools” described below. Similar units were combined and themes were then extracted. Many of the labels initially chosen for the meaning units were changed or consolidated as new information was added and after consultation with the research assistant and consultants in the area of discourse analysis. Following the block-and-file approach, conceptual mapping was used in order to visually identify connections between particular themes or constructs (Grbich, 2013). Repetition of conceptually similar content, and, conversely, lack of repetition or “negative cases,” were noted. Rather than pure word frequency, I focused on category frequency (examining how often words that were conceptually related appeared in a given cluster of text). For example, one theme that emerged in the Initial Contact phase of all non-rapport containing chats was labelled “client requests for personal information.” This theme was characterized by frequent use of personal pronouns (“you, I”) and references to counsellor feeling states or motives (for example, “do you get bored hearing this stuff over and over?” or “sorry to depress you with my stuff”). In order to provide a preliminary visual representation of the frequency of repeated words on the part of the crisis centre volunteer and to examine general trends, Wordles (see Appendix E) of both counsellor responses and client responses were created to assist in generating themes, as suggested by Grbich (2013). 54 Transcripts were examined until saturation was reached. Saturation was defined as per Glaser and Strauss (1967) where new data did not contribute any further information to the data being studied. Recognizing that data saturation is a somewhat arbitrary concept as discussed by Grbich (2013), and considering the large amount of complex data at my disposal, data were collected until the next 10 chats in each category (rapport-containing and non-rapport-containing) did not contribute new categories because they were subsumable under the existing thematic categories. Of interest was the observation that saturation was reached much more rapidly with the non-rapport containing chats; this could be partially attributable to their shorter length. Stage two: discourse analysis. Based on preliminary examination of potential data sets (sample transcripts provided by the Vancouver Crisis Centre), as well as on the data collected as part of my Master’s thesis (Timm, 2011), it was expected that the data obtained would be abundant, complex, deep in meaning, and possessing ambivalence, discrepancies, and dissonances. These characteristics were expected in part because of the ambivalence often expressed by those who are suicidal (Rudd, 2006; Shea, 2008), the stigma of suicide that may prevent direct expression, and the online nature of the interaction which can create greater potential for multiple interpretations of a given sentence (e.g., sarcasm mistaken for excitement). This prediction was supported by the characteristics found in the current data set and informed the decision to use discourse analysis as a method of examining it. A second reason for the use of discourse analysis was the limitations of content analysis. It is recognized that there are a number of drawbacks to using content analysis as the sole method of data analysis for a data set that possesses these above-named characteristics. First, content analysis cannot analyze what is not visible. Working from the assumption that all communication is relational (e.g., Westwood & 55 Ishiyama, 1990), a given statement made by either client or counsellor has both an explicit (said) and implicit (unsaid) message. Content analysis is only able to capture the former, while the latter may provide access to rich, unexplored themes. Table 5, below, is paraphrased from the data obtained in this study and illustrates the potential utility of paying attention to the implicit messages in a given interaction. Table 5 Explicit versus Implicit Messages: Illustrative Example Second, as Table 5 above illustrates, contextual factors and common discursive practices may also be missed when attending solely to content. For example, failure to listen to client cues, as well as implied difference in social class or power between client and counsellor, are important aspects of the dyadic interaction that cannot be attended to using solely content analysis. Therefore, while content analysis provided valuable information regarding what was explicitly said by counsellor and client (thus providing information that may help in the creation of new training protocols), its failure to examine what was not present (e.g., the implicit messages conveyed by client and counsellor) as well as its inability to take into account Sample Counsellor Response Sample CounsellorResponse Sample Client Response “What are you watching on Netflix at the moment?” “I totally understand that. I screen my calls too sometimes.” “How do you help suicidal people if you can’t easily tell them where to do for help?” Explicit: Asking about activities after the call Explicit: Self-disclosure of counsellor activities Explicit: Asking about/questioning counsellor’s ability to help Implicit: “I care about you.” Implicit: “I have been where you are. You are not alone.” Implicit: “I feel frustrated with this interaction” or “I am suspicious of you/you need to prove yourself to me.” 56 contextual factors (e.g., issues of power and social class), made it an insufficient method to use on its own in the examination of my data. The characteristics of the data outlined above and the gaps left by content analysis, as well as the nature of my research questions that aim to explore a new domain, necessitated a data analysis method that was sensitive to these factors. Specifically, a method was needed that took into account language use, paid attention to issues of social power, expected and accounted for ambiguities, and was open to challenging the dominant discourse on a given issue. Discourse analysis was selected as a method that fit these needs. Table 6, below, enumerates how discourse analysis can add to the information obtained through content analysis. Table 6 Contributions of Discourse Analysis Discourse analysis: Theory and practice. Discourse analysis (DA) was developed in the late 1980’s by a group of social psychologists in the United Kingdom. It arose partly in opposition to the dominant post-positivist experimental research paradigm of the time (Potter & Wetherall, 1987). Discourses have been defined as “[…] specific ways of thinking and speaking about the world generated by social power” (Phelan, Wright & Gibson, 2014, p.2). DA is based on several key assumptions: first, that language is context-dependent and action-oriented; for Content Analysis Discourse Analysis Explicit Implicit Frequency of word use Context of word use Themes and patterns Themes and patterns in context What is said What is unsaid/missing No attention to social power Attention to social power 57 example, it informs how people perform various social actions. Second, that discourses should therefore be critically examined as an indicator of social practices that are taken for granted – and reinforced using language – by members of a given society. Spong (2004) succinctly summarizes this process: “[…] the way we talk about things does not merely describe the world, but makes the world what it is (p. 68).” Common practices when conducting a DA on a given set of data include paying attention to positionality (how people place themselves and others in relation to a given social role or narrative), and looking for exceptions to common social discourses and their effect on the interaction. For example, Spong (2010) provides a potential discursive analysis question applied to the counselling domain: “how are our clients constrained and liberated by their engagement in counselling discourses?” (p. 68). Discourse analysis is often focused on the co-construction of meaning between two parties: narrator and listener. It is assumed that two parallel processes are often at play in a given discourse: a “heuristic” process in which “[...] A narrator, puzzled by her own experience, narrates it in hope of arriving at a clearer understanding of its meaning, and a “negotiative” process in which “a process of exploration, negotiation, and reformulation of narrative meaning is instigated by the listener” (Nye, 1998, p. 272). Transposing these two processes into the counselling domain, Nye (1998) notes that “[…] the client is the perpetual narrator, the one who tells her story. The worker is the listener, engaged either in the heuristic process of co-construction of meaning or in initiating a negotiative process by challenging the client’s taken for granted meaning. Tensions about power and control in the narrative process may be exacerbated in the treatment relationship by this structurally fixed role inequality” (p. 273). The narrator (client) thus implicitly positions herself in a 58 vulnerable space by virtue of her narration, implying expertise on the part of the listener (counsellor) to make sense of her experience. In the context of discussing suicidality online, the vulnerability of the client is further highlighted due to the sensitivity of the subject matter. Also, the neutral, more egalitarian space of the online context implies the need for a potential re-examination of dominant modes of interaction. Nye (1998) notes that “being conscious of the [imbalance of power] in the therapeutic process precedes using it most effectively for therapeutic ends” (p. 273). It is hoped that the focus on therapeutic process in this study (specifically, on rapport-building) will assist clinicians in becoming more conscious of (and potentially amending) their work with suicidal clients online. Therefore, discourse analysis was viewed as consistent with both the underlying epistemology and the aims of this study’s research questions: it addresses the changing forms of discourse as an interactive activity in a specific (online) context. Discourse analysis has been used to examine a wide variety of mental health concerns. For example, it has been used to explore how problem gamblers seek help online (Mudry & Strong, 2012), how ideas of schizophrenia are co-constructed by clients and counsellors (Larsson, Loewenthal, & Brooks, 2012), and how suicide is spoken about by clients and counsellors (Reeves et al., 2004). Most studies employing discourse analysis to examine suicidality have focused on interview and focus group data obtained from those who were previously suicidal (Roen et al., 2008). Some researchers have attempted to re-create the suicidal situation by employing suicidal “client actors” (Reeves et al., 2004). This research aimed to expand the transferability (Morrow, 2005) of the findings by examining actual transcripts of real conversations about suicide occurring between clients and crisis counsellors. A criticism leveled at much research in counselling psychology concerns the notion of rapport and the observation 59 that most individuals conducting such research have training in active listening skills and may interact with participants in a manner that is more “counsellor” than “researcher” (Morrow, 2005). Because the data sources in this study were archival in nature, this was hoped to increase the trustworthiness of the findings by elimintating the potential of the researcher interacting with the data sources (clients and counsellors). Applicability of a Foucauldian lens. A critical Foucauldian perspective on discourse analysis (Grbich, 2013) involves tracking a given discourse over time and identifying key players as well as the social, political, and economic climate that influenced how it developed. It seems particularly relevant for the data under examination. Namely, the newness of this mode of counselling and the vulnerability of the population under study were considered a good fit for the Foucauldian emphasis on historical discourses and distribution of social power, respectively. As noted by Hanley (2006), and Martin and Stuart (2011), the neutrality and shared nature of the online space implies a shift in social power between the dominant discourse that tends to be played out in the arena of counselling and mental health of “counsellor as expert” (Cowan & McLeod, 2004). Looking at this discourse from a critical, Foucauldian perspective, it was deemed important to first examine the genesis of this dominant discourse. Many discursive practices in counselling are based on the medical model which implies pathology on the part of the client and expertise on the part of the clinician. This discourse has its origins in historical events and cultural context: The Surgeon General's staff administered intelligence and personality tests during World War 1 to the almost two million military recruits. This required the creation of diagnostic categories and, simultaneously, the development of the implicit assumption that people can be categorised as mentally “sick” or “well.” 60 This dominant discourse has been further perpetuated by the institutionalization of the mentally ill and by the portrayal of those with mental illness as dangerous or unpredictable in popular culture such as films like One Flew Over the Cuckoo’s Nest. On a more subtle note, the language used when describing individuals who have been diagnosed with a mental disorder tends to be totalizing language that equates illness with identity, a discourse that Foucault (1973) critiqued in his book The birth of the clinic: An archaeology of medical perception. For example, a psychologist may refer to a client as “depressed” or “schizophrenic” rather than as a “person with schizophrenia.” Many of those struggling with mental disorders have adopted this language, extending and perpetuating the dominant discourse of “clinician as expert” with its implication of social power. Critical disability studies have taken on this critique of the medical model, noting that it “[...] lends scientific credibility to the assumption that the source of the ‘problem’ related disability is located within the individual body, thereby absolving society of any complicity in the lived outcomes for people with disabilities” (Lalvani & Douglas, 2013, p. 2). An example of discourse analysis focused on critiquing the medical model is the research of Phelan et al. (2014) who examined representations of disability and normality in rehabilitation technology promotional materials. They found that “normalized” discourses constructed, through image and text, the idea of the disabled person as one needing to be fixed, implying a socially constructed notion of “normal” in which the social power rests with those without the identified disability. The newness of online counselling as a modality necessitates a re-examining of traditional modes of providing therapeutic services. Discourse analysis, with its critical, social constructionist lens, seems well-suited to the proposed deconstruction of dominant modes of interaction (and the search for exceptions to these dominant discourses) within the domain of 61 online counselling. In particular, Foucauldian themes of social power and language seem especially appropriate potential areas of focus. Cowan and McLeod (2004) note that “the effect of discourse analysis in counselling and psychotherapy is […] largely that of questioning prevailing therapeutic assumptions and practices” (p. 102). As noted above, the topic of suicide itself has been examined using discourse analysis; for example, Roen et al. (2008) used a poststructuralist, Foucauldian framework to examine how previously suicidal young people positioned themselves in relation to suicide, with attention paid to issues of social power that may impact a given discourse. Working from Foucault’s (1973) assumption that discourse is not only embedded in language but also gives rise to action, they found that youth tended “other” suicidal individuals, demonstrated a tendency towards rationalizing suicidal behavior, and understood suicidal individuals primarily through their relationships with others. Similarly, Reeves et al. (2004) conducted a critical discourse analysis examining how the discursive object of “suicide” was negotiated by client actors simulating suicidality and their counsellors. Critical discourse analysis was cited as an appropriate method of investigation for this subject as its goal was to “[…] describe and critique the discursive worlds people inhabit and explore their implications for subjective experience” (p.63). In this study, it was discovered, interestingly, that the discursive subject (suicide) was referred to mostly in indirect terms (via metaphor, imagery, etc). The researchers note that “both [client and counsellor] seem to collude in not specifically naming suicide as an active possibility” (Reeves et al., 2004, p.66) and attribute this finding to the stigma surrounding suicide. As these studies indicate, discourse analysis has been shown to be an effective tool for understanding the phenomenon of suicide in a critical and nuanced manner. 62 The nature of the data used in this project (transcripts of online chats with individuals who express thoughts of suicide) also renders it appropriate for discourse analysis: a potential concern regarding the data collected from this virtual space is the question of the veracity of the “stories” obtained from the chats. It could be argued that there is no way of verifying whether the information (demographic, contextual, and content-related), is objectively “true.” However, discourse analysis acknowledges the ambiguity and potential contradiction that may exist in a single sentence (Grbich, 2013). The “reality” of a given client’s demographic information could therefore be viewed as a social construction rather than objective reality. The virtual space: applicability of discourse analysis. The context in which this examination took place was also considered to lend itself well to discourse analysis: Martin and Stuart (2011) note that references to the “virtual world” (referring to online communication) sometimes imply that “[virtual] is something that has an effect but is not real […] if reality is accepted as a subjective concept and trying to understand the life-space of the other person is important, then it is necessary to explore the aspects of life-space that affect that person […]” (p. 56). Thus, my adoption of the epistemological framework of social constructionism is consistent with the maxim “the map is not the territory” first posited by linguist Alfred Korzybski (1990): namely, that what is represented in the chat transcripts will always be an imperfect reflection of the client’s internal experience, and that I as a researcher inevitably transposed my own views and beliefs onto the data. Discourse analysis as a method is therefore consistent with the social constructionist epistemology that informed this investigation. In this secondary phase of data analysis, the implicit messages of both counsellor and client were examined as they pertained to the client/counsellor relationship, and emergent themes were examined with attention paid to 63 issues of social power, language, and context. Of interest for the purposes of the study was how rapport was created and built in all three phases of a given chat. Discourse analysis: Tools used. In order to answer the research questions above using discourse analysis, four main “tools” for conducting discourse analysis were implemented, as described by Gee (2011). Gee provides twenty-seven tools for use in discourse analysis, noting that all twenty-seven are overlapping and may all be simultaneously applied to a given portion of data. However, he acknowledges that “[…] for some data, some tools will yield more illuminating information that for other data” (p. x). Taking this guidance into account and considering the characteristics of the data, the following four tools were selected: a. The identities building tool b. The relationships building tool c. The activities-building tool d. The figured worlds tool The first three tools were selected because the questions they ask were considered consistent with the questions asked in this study; namely, how the client and counsellor express their identities and build their relationship in an online context. The figured worlds tool (described in detail below) was chosen because it is designed to facilitate a “big-picture” examination of how a given discourse is constructed and/or perpetuated (in this case, the discourse of a counselling encounter online in which the client expresses suicide ideation). These tools and how they were applied to the segments identified to answer the research questions above are now described. The identities-building tool. This tool builds on the assumption that individuals construct different identities depending on their social context, and that their language use reflects this. For example, a psychologist may use different vocabulary, grammar, and syntax when speaking with 64 colleagues, with clients, and with the general public. Gee (2011) notes that this latter context, the general public domain, often requires what he refers to as the enactment of the “life world identity” (p. 107) in which an individual, while engaging in the linguistic and social norms of their culture, assumes the role of the “everyday” person, separate from any professional or socially affiliated identity. The “life-world” identity described by Gee (2011) is inspired by the work of philosopher Jurgen Habermas (1984) who first used this definition. This tool takes into account the observation that as individuals construct and express their own identity through language, they also co-create the identities of others with whom they relate through their interaction with them. For example, a client consistently using profanity in conversation with his counsellor may be positioning the counsellor as a friend or as someone on the same social level. Conversely, the counsellor’s response to this (e.g., complicity or rejection) may solidify or change this assigned identity. In the online environment, the concept of identity is markedly more fluid – both participants in the discourse have more freedom to construct their identities independent of the constaints of physical presence, as noted by Murdy and Strong (2012) who examined how the identities of problem gamblers are constructed in an online support group. Thus, the identities -building tool was considered especially relevant given the online context of the interaction. Gee (2011) suggests that when using the identities-building tool, researchers should “[…] ask what socially recognizable identity or identities the speaker is trying to enact or get others to recognize. Also ask how the speaker’s language treats other people’s identities, what sorts of identities the speaker recognizes for others in relationship to his or her own” (p. 110). Therefore, the following questions were asked of each segment of client/counsellor text (unit of meaning, as described above): 65 Q 1. What type of identity is the client enacting, and what impact does this have on the counsellor? Q 2. What type of identity is the counsellor enacting, and what impact does this have on the client? Q 3. Are there any contradictions or tensions in how client and/or counsellor enact their respective identities? The relationships-building tool. The relationships-building tool is built on the assumption that language use builds and enacts relationships with people, groups, or institutions. Gee (2011) notes that while the relationships-building and the identity-building tools are closely linked (our identities impact our relationships and vice versa), they are not one and the same. For the purposes of this research project, also, it was considered prudent to separate them in order to more clearly examine how rapport was built in the absence of voice tone. For example, the youth using profanity in the example above might have been constructing an identity of toughness and rebellion for himself, and co-constructing a similar identity on the part of the counsellor. The relationship he was forging between himself and the counsellor, on the other hand, could be described as one of complicity or solidarity. Therefore, the relationships-building tool was implemented by asking the following questions: Q 1. How do the enacted identities of client and counsellor impact the relationship between them? Q 2. What kind of relationship is the client building through use of language? Q 3. What kind of relationship is the counsellor building through use of language? Q 4. Are there any tensions or contradictions in the ways in which client and counsellor are building or construing their relationship? 66 The activities-building tool. As per Gee (2011), this tool is typically used to illuminate how communication facilitates or enacts particular activities. It is based on Foucault’s (1973) observation that social discourse is inextricably linked to action. In the case of this data set, the delineation of phases (Initial Contact, Suicide Assessment, and Termination) clearly illustrates the assumption of particular activities taking place in each phase. Also, the conversation that takes place between a suicidal individual and a clinician has been viewed by action theorists as constructing a joint action, the details of which merit further exploration (Valach, Young & Michel, 2011). How these activities are spoken about and negotiated was therefore assumed to impact the client-counsellor relationship (the primary focus of this study); hence the activities-building tool was applied to the data set. The questions asked of each transcript using this tool were as follows: Q1. What activity is this communication facilitating or enacting? Q2. How is this activity related to the client-counsellor relationship? Q3. Are there common themes across phases in terms of how activities are spoken about or facilitated? The figured worlds tool. The figured worlds tool draws on the observation that social practices (in this case, modes of interaction in an online suicide prevention service) constitute a set of assumptions about the both the world at large and the current situational context. Gee (2011) refers to this set of assumptions collectively as “typical stories, […] folk theories, mental models, frames, cultural models, or discourse models” (p. 170). Part of undertaking a discourse analysis includes discovering the rules of the discourse (Grbich, 2013). For example, when examining how counsellors typically interact with clients in the general arena of mental health, several potential rules of discourse (and associated behaviours or lack thereof) may be identified. 67 In a traditional counselling setting, the client typically enters the counsellor’s space (office), and the client assumes the counsellor possesses skills or knowledge that the client is lacking (a difference in social power). Because online counselling takes place within a shared space and the physical markers of difference or sameness are less present, exploring the set of assumptions governing a therapeutic conversation in this space seemed pertinent. For the purposes of this study, therefore, it was assumed that each conversation between counsellor and client reflected a particular set of assumptions about a) what a counselling session is b) how the relationship between client and counsellor was built and maintained in the context of suicide prevention online. These assumptions led to the development of the following questions asked of the data in the employment of the figured worlds tool: Q 1. What must the speaker (client or counsellor) assume about the counselling relationship to have spoken in this way? Q 2. What would the counselling relationship look like if one accepts the figured world expressed by the speaker? (adapted from Gee, 2011, p. 173). A related tool that was used to inform the questions above was the “context is reflexive” tool, which encourages the researcher to consider how a given context (in this case, the online nature of the interaction), impacts the speakers (client and counsellor). It was understood, as the above description illustrates, and as acknowledged by Gee (2011), there may be significant overlap between the results obtained from use of the four DA tools utilized. For example, as observed by Parker (1992), discourse analysis assumes that social identities are implicitly linked to specific social activities. However, it was also hoped that the different focus of each one would contribute a more nuanced view of what occurs in an online conversation where a client is suicidal. Because of the broader picture captured by the figured worlds tool as described above, 68 chats were not broken into phases when this tool was used, but viewed holistically with attention to themese across phases. Similarly, the “filling in” tool and the “making strange” (Gee, 2011) tool were both used across phases in all chats. These tools ask the researcher to consider that which is inferred or implicit in the case of the “filling in” tool, and to consider what aspects of the client-counsellor conversations might be perceived as “unclear, confusing, worth questioning” (Gee, 2011, p. 12) in the case of the “making strange” tool. Ethical Concerns Informed consent. It was understood that the archival nature of the data precluded the ability to obtain informed consent in the traditional sense. Freeman and Mathison (2001) cite Murphy and Dingwall’s position on adhering too rigidly to ethical codes: “[…] ethical codes that are not method-sensitive may constrain research unnecessarily and inappropriately […]” (p. 340). I was aware of the difficulty in balancing ethical issues around informed consent with the goals of this study and addressed this through consultation with other researchers and examination of similar studies, as outlined below. Battle (2010) notes that when accessing youth data online, “the transient nature of the Internet in combination with anonymity means that some users may be impossible to track down” (p. 32). She observes that in these cases, research procedures may need to be amended to reflect this fact. This was the case in the present study: due to the anonymous nature of the services, traditional informed consent was not possible. The Crisis Centre declined to provide a “check box” indicating active assent that clients’ non-identifying information may be used for research purposes. The rationale behind this decision was the belief that active assent would compromise the building of trust and rapport, as well as hinder self-disclosure during a period when it is most critical. However, when I conducted my Master’s thesis research, a compromise was reached regarding balancing rapport with informed consent: the Crisis Centre agreed to 69 place a disclaimer on the youthinbc.com website indicating to clients that their (non-identifying) information may be used for research purposes. This occurred on December 2009 (see Appendix C). The Crisis Centre created a similar disclaimer on their new www.crisiscentre.ca site for adults, which was posted on the site on May 13, 2014. Only chats from each respective service received after these dates were included in the analysis. It should also be noted that as per the Tri-Council Policy Statement (TCPS-2, http://www.pre.ethics.gc.ca/eng/policy-politique/initiatives/tcps2-eptc2/Default/) specifies that informed consent should be obtained when conducting research with secondary data when “information provided for secondary use in research can be linked to individuals, and when the possibility exists that individuals can be identified in published reports, or through data linkage.” As such, my study was considered in compliance with TCPS standards even without directly obtaining consent from the clients or counsellors. However, since it was technically possible to obtain consent from counsellors (as they were not completely anonymous, unlike the clients), it was considered prudent to do so. Confidentiality. Chat transcripts with identifying information removed were provided to me by the Online Services Coordinator at the Vancouver Crisis Centre. Transcripts were selected using the inclusion criteria described above. With respect to client confidentiality, Battle (2010) notes that “[…] online and real-life identities should be equally protected in research” (p. 32). To prevent confidentiality from being breached, identifying features were obscured by eliminating identifying details and by blending different cases as needed. It is recognized that that this blending of cases needed to be balanced with the value of providing the reader with “thick description” of unique features of individual chat transcripts to facilitate transferability of the findings (Morrow, 2005). These features included illustrations of a particular pattern or noteworthy interactions worthy of further study. Decisions on whether to include particular 70 examples were directed to my research supervisor and two faculty members with experience in discourse analysis. Direct quotes were paraphrased and combined from different transcripts in order to avoid potential identification of a client or volunteer. On a related note, Grbich (2013) raises the issue that, when conducting research in an online environment, the anonymity of research participants results in an “[…] “inability to determine the extent of [the sample’s] diversity” (p. 158). Therefore, was recognized that the demographic information collected (age, gender, location) could not be verified as correct in the same way that it could when conducting research with participants in-person. Increasing Trustworthiness: Data Verification It was recognized that since the data would be viewed through the subjective lens of me as a researcher, the picture that emerged of each case would inevitably differ from that recorded by another researcher. A research assistant was therefore used to ensure that the written reports conformed to the actual transcripts and to provide a concurrent rating of rapport or lack thereof. Completed analyses were sent to my supervisor for feedback. Furthermore, two faculty members with experience in discourse analysis were consulted numerous times throughout the data analysis process to ensure that proper procedures were being followed and that the observations both reflected the data and were consistent with the principles of discourse analysis. The research assistant was recruited based on the following criteria: (a) possessing at least undergraduate training in counselling skills (b) no current affiliation with the Vancouver Crisis Centre (c) possessing a moderate level of fluency in online communication (assessed at interview). The first two criteria were required as the aims of the study are most applicable to those working in the helping professions. Further, having a research assistant familiar with counselling skills was thought to enable them to effectively note (and critique) the establishment 71 of the therapeutic relationship online and how it may follow or stray from the dominant discourse of the counsellor-client relationship. Specific experience conducting crisis counselling online was not considered a prerequisite. This is first because of the difficulty in finding research assistants meeting this specific criterion, and second because exposure to another model of online helping was thought to potentially bias the research assistant in their evaluation of the current data set. However, the research assistant was required to successfully complete the Suicide Intervention Response Inventory (SIRI; Neimeyer & Bonnelle, 1997) prior to beginning data analysis to ensure her familiarity with suicide intervention and risk assessment procedures. The SIRI is a self-report inventory comprised of 25 items which assesses an individual’s level of competence in suicide intervention in response to a number of hypothetical scenarios involving suicidal clients. It has shown high reliability and validity in both in-person and telephonic counselling environments (although, like many measures, it has not been evaluated for online counselling). It has also been demonstrated to be free of social desirability bias (Neimeyer & Bonnelle, 1997). A moderate level of online fluency (e.g., understanding of the basic mechanics of an online exchange; fluency with online-specific linguistic conventions such as use of emoticons and abbreviations) was also sought as this was considered a necessary skill when undertaking the analysis described. Training in discourse analysis was also conducted via assigned readings and provision of sample chats for practice purposes prior to the actual data analysis. Two research assistants were initially selected meeting these critera; however, the first was unable to commit to the duration of the project due to relocation and therefore her initial analyses were discarded. The research assistant selected for the duration of the project had experience conducting both online crisis counselling and telephonic online counselling, as well as a bachelor’s degree in Psychology. Her 72 experience conducting crisis counselling in both modalities was considered an asset with respect to her ability to analyze the chats. 73 4. FINDINGS In order to facilitate a multidimensional view of the data as suggested by Grbich (2013) and Denzin and Lincoln (1994), the findings of this study are presented in numerous formats. First, a written description of the themes observed in each phase with respect to the client-counsellor relationship is provided, as are the results of the application of the various DA tools described above. Second, the themes found and salient examples thereof are presented in tabular form. Third, detailed vignettes including examples of how the coding process was conducted are provided in order to facilitate transferability of the findings and hence increase the trustworthiness of the study (Morrow, 2005). Credibility was enhanced by using the research assistant as an additional analyst who coded transcripts concurrently. Agreement with respect to identification of themes was determined using the method suggested by DeCuir-Gunby et al. (2011): dividing the number of agreements on codes by the total number of agreements and disagreements. However, since the aim of qualitative research is not to seek results that are generalizable to a larger population but to offer a rich and nuanced picture of the phenomenon under investigation, disagreements between raters are included and discussed. The inter-rater agreement for the discourse analysis portion of the research was 82%, which was considered sufficient for the nature of the data as outlined earlier (Wongkaparan et al., 2013). Rapport-Containing Chats Phase 1 (initial contact phase): client identity. Two main themes were found with respect to how the client constructs their identity in the Initial Contact Phase: that of a person both desiring help and rejecting it, and that of a person with a story to tell. In this phase, the client presents with an interesting dialectic: a person both needing/desiring help and feeling 74 beyond it: “I do want things to change but tbh [to be honest] I cant see it happening anytime soon”).This conflicted identity has implications for the implied needs of the client: the client seems to want the counsellor to simultaneously provide acceptance and witnessing as well as concrete solutions for a given presenting concern (examples of these concerns in three separate chats include bullying, sexual trauma, and depression). A further facet of the client’s identity consistent across cases is that of a person with a “story to tell” – the assumption of a narrative that reaches beyond the current suicidal crisis, and the need for someone to hear it. This client need is manifested by linguistic markers of time; examples include “How far back do you want me to go?” “Should I start when I was ten and molested?” and “its been happening since gr.7 and im in gr.9 now. i wonder if everything really does get better or people just say that.” Phase 1 (initial contact phase): counsellor identity. The counsellor in the rapport-containing chats appears to construct their identity as that of a witness in the Initial Contact phase. This identity seems to complement that of the client’s implied need for someone to witness and accept their expressed hopelessness (“I feel more calm I think even though I’m still very anxious if that makes sense […] I just want you to listen to me”). This counsellor identity of witness is evidenced by the counsellor’s reflections of the client’s experience without changing or questioning them, even when the statements appear bleak or hopeless (“mhm like no matter what happens at this point, you feel trapped in a corner of overwhelming anxiety about everything that you’re supposed to be doing but don’t feel like you could actually accomplish in the time at hand”). In four of the rapport-containing chats, the “witness” counsellor identity is responded to by an overt expression of relief on the part of the client and the implication that the counsellor has provided a corrective emotional experience (“thank you for listening and letting me let it out without like telling me i should just get off my ass”). 75 The counsellor’s witness identity remains consistent in Phase 1 even at those times where the client’s identity shifts to that of a person wanting concrete help. However, in the rapport-containing chats, any potential tension between these conflicting identities is resolved either through the counsellor employing genuineness and transparency in his or her role (“Wow, never had anyone ask me that before, haha”) or the client showing acceptance of the limitations of the counsellor’s role (“It’s ok, I get your position lol”). As illustrated by the preceding quotes, humour and informal language tend to be used by the counsellor in these moments of potential tension. Youth versus adult. The identity of the client as described above did not appear to be impacted by their status as adult or youth. Main differences between adults and youth in this phase thus concerned the strength (intensity) of the dialetic described above - in chats with youth, the strength of the belief that they were “beyond help” was more salient. Their desire for concrete solutions also seemed to be more strongly and explicitly stated than was the case with adult clients. Phase 1 (initial contact phase): relationship-building. Two main themes were found in this phase, which are outlined below. Theme 1: informality. Both client and counsellor appear to be building a relationship that is somewhat informal as evidenced by use of abbreviations, profanity, and casual language, as well as, in eight of the cases sampled, a more relaxed grammatical style abandoning typical punctuation and capitalization. The pattern appears to be for the counsellor to modify his or her response style in relation to the client: for example, in Chat #8, the client states that he is “just so fucking tired” – the first time profanity is introduced. The counsellor responds with “Yeah, I get it…you’re exhausted and frankly pissed off with the way things have aligned themselves.” 76 Theme 2: equality and respect. The relationship between counsellor and client in this phase appears egalitarian, as evidenced by the counsellor’s use of tentative language (as opposed to making concrete statements, diagnostic implications, or suggestions – types of counsellor responses that might be expected in a professional counselling context). For example, in Chat #4, the counsellor states “correct me if I’m wrong, but I’m sensing a real conflict here…you want to be safe but these thoughts of suicide are tempting you…did I get that right?” The counsellor thus appears to position him or herself as a curious, respectful witness to the client’s emotional experience in this phase. Youth versus adult. In this phase as in the preceding one, the difference between how the client and counsellor relationship is negotiated does not appear to be impacted by the client’s age. Subtle differences with respect to the two identified themes (informality and equality) were observed with respect to the first category (informality) where chats with youth contained more profanity, abbreviations, and less attention to grammatical constructs (for example, lack of punctuation and capitalization). Youth responses also tended to be more succinct in this phase, sometimes consisting of only a word or two (e.g., “yup. about right”). Phase 2 (suicide assessment phase): client identity. The client’s identity in this phase seems to be a deepened and more complex manifestation of the “conflicted individual” presented in Phase 1: the dialectic of wanting and rejecting help comes into sharper focus with the explicit introduction of suicide into the conversation. In this phase, the client seems to present as someone who not only wants to be heard, but wants to be engaged with on a more practical level (with respect to safety planning and resources, specifically). The dialectic appears when the client expresses ambivalence about both staying safe and the potential helpfulness of offered 77 resources/coping strategies (“Yeah. I get that, but I’m honestly not too sure distracting with music is gonna cut it for me if you know what I mean. Cutting is better”). Phase 2 (suicide assessment phase): counsellor identity. The identity of the counsellor in this phase seems to shift in response to the more action-oriented needs of the client: while retaining the identity of empathic witness, the counsellor additionally introduces an element of suggestion; however, this is still tentatively framed: “Do you think you could agree to put the knife away for the night? Maybe give it to your sister to hold onto if you feel safe doing that?” Therefore, in this second phase, the identity of witness appears to have expanded to include the identity of “collaborative emotional first-aid responder” – including the elements of assessment (via use of closed questions versus the open-ended questions used in Phase 1), as well as overt behavioural suggestions, as in this counsellor response from Chat #5: “If you like, we could even call you and check in…so that way all you’d have to do is answer […] it’s not always easy to call when you feel so amped up and on edge.” The word “collaborative” was added to the description of the counsellor identity following discussions among research assistants and the primary investigator. It was thought to capture the positioning of the counsellor more accurately, as evidenced by frequent counsellor language indicating a shared goal: for example, “we” “us,” and “together.” Youth versus adult. In this phase, the difference between adult and youth seems more pronounced. Specifically, those clients identified as youth appear to display less knowledge of counselling resources and/or the counsellor role, which impacts the identity they assume (the identity of someone seeking help is more pronounced). The option of self-harm as a coping mechanism and the positioning of the self as someone who is “broken” or beyond help also seems more apparent in the conversations with youth. 78 Phase 2 (suicide assessment phase): client/counsellor relationship. In Chat #5, as noted above in the example of the collaborative counsellor identity, the counsellor states: “If you like, we could even call you and check in…so that way all you’d have to do is answer […] it’s not always easy to call when you feel so amped up and on edge.” The counsellor statement above encapsulates the themes present in the relationship between client and counsellor in this second phase, which is related to the counsellor’s identity described above. The tentativeness highlights the implied respect of counsellor for client; acceptance of the difficulty calling exemplifies the empathic understanding, and use of the expression “amped-up” illustrates the informal nature of the exchange. Therefore, the relationship between counsellor and client in this phase could be described as similar to the one that has been built in the initial contact phase (informal and egalitarian), with the addition of more facilitation of collaborative action on the part of the counsellor, as well as the addition of counsellor genuineness. The two main themes present in the Suicide Assessment Phase were therefore themes of collaboration and counsellor genuinness, and are described below. Theme 1: client/counsellor collaboration. It seems that in this phase, the counsellor is able to modify his or her identity in response to the evolving expressed needs of the client. For example, in Chat #4, the counsellor states: “honestly, I’m still kinda worried about you since seems like these [suicidal] thoughts are pretty strong. Can we plan together how to keep you safe?” Use of the word “we” implies a shared goal; use of the word “plan” implies future shared actions, and use of the word “can” implies tentativeness and respect for the input of the client. Theme 2: counsellor genuineness. The relationship between client and counsellor, similar to how it is negotiated in the Initial Contact phase, seems characterized by the themes of informality and shared humour, however, in this phase, counsellor genuineness emerged as 79 another key theme. This theme initially represented an area of disagreement between myself, the primary investigator, and the research assistant. It was initially coded by myself as “counsellor vulnerability” due to the comfort the counsellor seemed to have in disclosing information that rendered them less than perfect. For example, in Chat #7, in response to a client’s request to “give me a reason to live” the counsellor responds: “to be honest, im not sure what to say to that. Give me a sec while I collect my thoughts…is that ok?” However, discussions concluded with the agreement that “counsellor genuineness” was a broader code that subsumed the expression of counsellor vulnerability. The theme of counsellor vulnerability encapsulated by counsellor genuineness is illustrated by an interaction in Chat #4 where, in response to a lengthy segment of text by the client, the counsellor responds with: “sorry, i just gotta catch up reading and my eyes are stinging.” The client responds: “I can imagine, i bet you feel really tired having to stare at the screen for so long. I know it has that affect on me” to which the counsellor responds: “phew, thanks for waiting …yeah, my eyes are a super high prescription so like my glasses give me headaches sometimes lol.” The client, in turn, seems to reverse the typical counselling process by showing empathy for the counsellor, even using the “sounds like” tentative lead previously used by the counsellor in the earlier part of the transcript: “I am sorry to hear that. Sounds like they need better technology or you need newer glasses.” Youth versus adult. In this phase, the difference between adult and youth chats as it pertains to the relationship seems related to the level of directiveness of the counsellor (more directiveness was observed in chats with youth). Also, similar to the pattern in the Initial Contact phase, there was more informal and online-specific language (“lol;” “OMG”) used by both 80 parties in the chats with youth, which makes the theme of informality more pronounced in these chats in both phases. Phase 3 (termination phase): client identity. In the final phase, the client seems to position themselves as a friend or peer of the counsellor. This identity, in ten of the chats, appears to be a reaction to the counsellor’s identity as a non-judgmental witness where the client expresses acceptance of their current situation “I know, it will suck tonight, thanks for getting that lol.” This acceptance is tempered with, in four chats, some trepidation regarding their ability to regulate their emotions and/or stay safe: “honestly, I dunno if I can put the pills in another room but I guess I could call u guys if I get tempted [to take them].” Phase 3 (termination phase): counsellor identity. In this phase, the counsellor again takes on the role of witness and the role of “collaborative emotional first-aid responder” found in the previous two phases. However, in this phase, the identity of “promoter of safety” seems to emerge more prominently in this final phase: this identity seems salient based on the multiple times the word “safe” is used in this phase, as well as the counsellor’s more structured focus on post-chat activity planning. Interestingly, in this final phase the identity of the counsellor seems to expand to one of “friend” or “peer;” this is evidenced by the counsellor’s self-disclosure, initiation of conversations about shared activities, making jokes, and informal language. Youth versus adult. The client’s identity as resigned and somewhat ambivalent seems consistent across chats regardless of the age of the chatter. In the cases of youth chats, there is more frequent expression of resistance to other forms of communication or support (i.e., phone or in-person counselling). Phase 3 (termination phase): client/counsellor relationship. In this final phase, the relationship between client and counsellor appears, once again, to take on an egalitarian quality, 81 evidenced by not only informality but also counsellor self-disclosure, and, particularly in Chat #10 and Chat # 12, counsellor self-disclosure and reference to shared elements of pop culture (Netflix; literature; pets). Three key themes thus emerge in this final phase: counsellor self-disclosure, shared humour, and client trust. Theme 1: counsellor self-disclosure. Of interest is the observation that the forays into personal territory described above are often initiated or strongly encouraged by the counsellor. For example, when asked about self-care in Chat #4, the youth states: “Watch Netflix.” The counsellor responds: “Oh yeah, what? Theres so much good stuff on there.” On a similar note, in Chat #9, in response to the client’s statement that he will read a novel to relax, the counsellor responds: “ah…started that one but couldn't finish it. L-o-v-e-d Mansfield Park though.” This lengthens the conversation by a good ten minutes, which seems to violate the dominant discourse of crisis counselling as a “time-limited” or “problem-focused” therapeutic encounter. Theme 2: shared humour. The theme of shared humour emerged strongly in the termination phase. For example, in Chat #9, the conversation in the final phase is punctuated by both dark humour and non-crisis related content. Similarly, in Chat #7, in response to the client’s expressed wish to have “aliens just take over this damn planet cause we’re killing ourselves,” the counsellor responds: “before we go further…would you want the aliens to be small and green, or exceptionally good looking?” This observed theme of humour seems to violate the dominant social discourse of “counsellor as professional” and “suicide as serious.” Theme 3: client trust. A salient theme in this final phase was of trust on the part of the client. This theme was not as apparent in the first two phases, which contained frequent instances of client questions about counsellor competence, training, and trustworthiness: “How do u know u won’t call the cops on me if I tell u how I feel?” In this final phase, the client appears to trust 82 the counsellor enough to disclose their fears and vulnerabilities: “To be honest, I’m not sure I can stay safe I wanna try.” Youth versus adult. The themes of counsellor self-disclosure, trust, and humour appear consistently in all rapport-containing chats in this phase, regardless of the age of the client. As in the previous phases, the egalitarian nature of the relationship is more strongly emphasized in the youth chat transcripts as evidenced by the more frequent use of informal language. The Activities Building Tool In all three phases of the rapport-containing chats, the range of activities promoted or facilitated appears not to be as tightly constrained to a particular phase as they are in the non-rapport-containing chats (which are described below). In the rapport-containing chats, the counsellor appeared to fluidly move from discussion of suicide risk assessment to self-care and then back to risk assessment in response to the client’s expressed needs. In three of the rapport-containing chats, a finding emerged that was somewhat alarming with respect to the process of risk assessment: in some chats, focus on the therapeutic relationship (evidenced by counsellor statements of caring, self-disclosure, and acceptance of the client’s state) seems to supercede a focus on suicide risk assessment. For example, in one chat, the client expresses feeling “[…] like every day may be the last…not sure how long i can keep this up even thinking how i would like end it all” [italics mine]. While the counsellor expresses non-judgemental acceptance and empathy for this experience (“I get that…you’re feeling totally exhausted from trying to survive each day”), the italicized portion of the client’s statement also implies the existence of a suicide plan; this is not picked up by the counsellor. However, in this same chat, towards the end, the client’s level of risk seems to have been mitigated on its own, with the client stating “thanks for hearing me…i feel a lot calmer now and will just stay in and 83 watch tv…you are awesome and thanks for doing this work.” Additionally, it is notable that in the rapport-containing chats, the counsellor seems to have less of a focus on time, as evidenced by fewer linguistic markers indicating the passage of time, such as “now, soon, later.” Fourth phase: social conversation. Interestingly, in the seven chats ratest highest in rapport, an additional “phase” is found towards the end of the chat, shortly before the Termination phase (although in one case it occurs in the Suicide Assessment phase). This phase was labelled “social conversation” and could be described as a mutual exchange of information on non-serious shared subjects, for example, books, movies, and well-known celebrities or political figures. This phase, interestingly, is often initiated by the counsellor, although in some instances the client introduces a topic and the counsellor expands on it. Of particular note is the temporal placement of this shared activity, which could be potentially be inferred as “rapport-building” in intention from a therapeutic perspective. Typically, the dominant discourse regarding the counselling relationship suggests that rapport-building in the form of social conversation be “done” as a counsellor activity at the beginning of a therapeutic conversation, in order to establish a connection and build trust. In fact, this is also a common social norm in everyday conversations, often referred to as “small talk.” The temporal placement of this phase and the active involvement of the counsellor in this social activity despite having “completed” the chat’s tasks (suicide assessment and safety planning) rendered it interesting and worthy of the creation of an additional “phase.” Additionally, with respect to the activities negotiated in this phase, the counsellor and client appear to collaboratively engage in action planning, as opposed to one party positioning themselves as the individual promoting or suggesting a given activity. For example, in Chat #12, 84 the counsellor states: “Seems like these [suicide] plans are pretty firm in your mind, but I really want to help keep you safe.” Insertion of the word “help” into the counsellor’s response here seems to help position the counsellor as a co-facilitator of the client’s safety rather than a person possessing sole authority over the client’s actions. The Figured Worlds Tool This tool was employed across all phases (rather than examining individual phases as done above) with the aim of providing a comprehensive picture of the themes related to the counselling relationship. This is consistent with the aims of the Figured Worlds tool, which is designed to examine larger themes related to a given discourse (Gee, 2011). The themes noted below were consistent across phases. Several assumptions regarding the counselling relationship as it was navigated in the selection of “successful” chats are outlined below.  Counsellor positioned as vulnerable; acknowledging limitations  Counsellor deferring to client’s experiences/opinions  Informal, “friendly” relationship, counsellor self-disclosure; shared popular culture references (Netflix; books).  Joining in humour; explicit and humorous discussion of the limitations of the counselling relationship. Because a Foucauldian analysis requires seeking the limits of a given discourse and looking for exceptions to it, this was a focus in the analysis of this section. Below is a sample transcript that appears to violate the “counsellor as expert” dominant discourse. This sample segment of conversation was chosen to illustrate the identified themes for two reasons: first, because it violates the dominant discourse about “helpers” and “helped” in the discourse of the counselling relationship; second, because it appears that these violations, rather than having a negative 85 impact, appear to positively impact the working alliance. Therefore, consistent with a Foucauldian analysis, the emphasis when conducting this type of analysis was on “tracking disunity” and looking for “challenges; traces of ideas that changed direction” (Grbich, 2013, p. 262). Consistent with a critical discourse analysis perspective, the coding scheme below takes into account “rules, norms, hierarchies that maintain power.” The highlighted portions represent themes of emancipation from the dominant discourse. The preliminary identified themes were as follows: Green: Acknowledgement of limits of knowledge (versus positioning as expert) Purple: Informal language (versus formal scientific language) Blue: Counsellor self-disclosure (versus deflection of personal questions) 86 Figure 2: Sample coding of rapport-containing chat youth1 1:11:07 AM how do you feel not knowing the person you’re helping might not be alive in a couple weeks? hypothetically youth1 1:11:42 AM ..that escalated Volunteer1 1:12:18 AM you know how you were saying how we can't know everything? Volunteer1 1:12:42 AM I don't know, I honestly wouldn't be able to explain to you in words youth1 1:12:59 AM shoot that got deep hey i got this.. i feel like your okay with oblivion, but expressing it is too hard did i do job your right? Volunteer1 1:15:24 AM wow, haha, never had that happen before Volunteer1 1:15:35 AM ...pretty accurate 87 In addition to the categories noted, which focus on the content of the client-counsellor conversation, the structure of the conversation is noteworthy since it clearly violates the dominant social discourse: the client asks the counsellor about his experience and provides an empathic reflection of the counsellor’s experience. This is a direct reversal of the typical course of events in a counselling session. Another theme found across cases illustrated in the transcript that is of note from a Foucauldian perspective is the use of humour; this is noted in the continuation of the transcript below. The back-and-forth banter between the youth and crisis centre volunteer seems unexpected in light of the seriousness of the subject matter: suicide prevention. Because it seems to violate a social norm, it seems worthy of further investigation: Yellow: aligns with youth; makes fun of own position Light blue: Aligns with youth; joins in humour Figure 3: Sample coding of rapport-containing chat  youth1 1:15:59 AM  mhm and how does that make you feel  Volunteer1 1:16:21 AM  now that's just a \"counseling\" stereotype isn't it?  youth1 1:17:10 AM  but the way you guys say mhhmm is just solid truth  Volunteer1 1:18:06 AM  you really did pick that up  youth1 1:18:13 AM  mmmhmmmm  Volunteer1 1:18:53 AM  that sounds convincing to me 88 In the preceding segment, the volunteer acknowledges “counselling stereotype,” speaking to how power and social norms can have an impact on how counsellors are perceived. “Isn’t it” appears to open up the discourse for the youth to respond. “That sounds convincing to me” implies collusion with the youth against the dominant social discourse of “counsellor as expert.” Further along in the transcript, it becomes clear that a strong emotional connection has been established: the youth specifically states that they will log in again, and (with humour) requests this specific volunteer, as seen below. An example of how the coding scheme for client responses was applied is as follows: Green: reference to the present relationship/conversation Yellow: Role reversal Blue: Specific, personal reference to the volunteer. 89 Figure 4: Sample of coding of rapport-containing chat youth1 1:19:14 AM this convo should go on ellen youth1 1:20:05 AM its like making my life.. like ill live just so i can laugh with some suicide robot dude Volunteer1 1:20:31 AM i'm glad you found something here tonight youth1 1:20:52 AM i counsel you you counsel l me Volunteer1 1:21:16 AM mm hmmmm Volunteer1 1:23:08 AM on another note, I do have to close up the system soon, because the chat's supposed to close at 1am, and I know the last time maybe didn't work so well for you, but you would log on and chat again if you want a robot to share it with? youth1 1:24:20 AM can the robot be a good robot? Volunteer1 1:24:33 AM I sure hope so .....systems upgrade all the timee youth1 1:24:42 AM can it be you 90 In the segment above, we see the youth express the impact of the interaction on them. It is interesting how the youth references laughter as a key part of the interaction and makes a joke about the volunteer being a robot (which initially was a point of contention). The counsellor also appears to pick up on and extend the “robot” joke, responding “I sure hope so…systems upgrade all the timeee…” The youth’s final words “can it be you” (requesting to speak to that specific person at a future date in time) seem to speak to the strength of the therapeutic connection that has been established. Non-Rapport-Containing Chats Phase 1 (initial contact phase): client identity. The client positions themselves in this phase in the same dialectical manner as in the rapport-containing chats: as someone who is simultaneously looking for help and beyond it. As in the rapport-containing chats, the identity of the “person with a story to tell” is present in this phase of the non-rapport-containing chats, evidenced by temporal markers as well; for example: “I have been thru a lot this year...shit that’s happened to me would fill like a hundred notebooks.” The theme of “help” seems more salient in the first phase of the non-rapport-containing chats and seems related to a second theme: that of client-perceived circularity in the conversation. “This place is difficult as you are mostly telling me you understand what I said and how it must be hard, but that doesn’t really help.” The client also positions themselves as more of a “consumer” of a service, from suggesting feedback: “There must be a better phrase than “what I hear you say is […] can you send this feedback to someone please,” to asking for concrete assistance: “I am looking for help here, not someone to repeat what I say.” Phase 1 (initial contact phase): counsellor identity. The counsellor appears to enact an identity that is more detached and professional than in the rapport-containing chats in this first 91 phase. This is evidenced by frequent use of the word “we” versus “I” when clarifying the nature of the services (found in seven of the eight non-rapport-containing chats), as well as lack of informal language or abbreviations. The identity of “professional provider of services” is further reinforced by language often found in the corporate realm: “service, feedback, unfortunately.” Youth versus adult. In the chats with clients identified as youth, the feedback to the counsellor, as well as the desire for help, is stated more directly and more aggressively. This aggression is manifested as sarcasm “you have great hearing by the way” or outright verbal attacks “this places makes ppl kill themselves much faster if u just repeat that I wanna die.” However, the themes observed are present in all chats regardless of client age. Phase 1 (initial contact phase): client-counsellor relationship. In this phase, the relationship between client and counsellor seems to be one of frustration on the part of the client and helplessness on the part of the counsellor. The two main themes in this phase were client requests for personal information and criticism of counsellor communication style. Theme 1: criticism of counsellor communication. In this phase, the client expresses dissatisfaction with two key aspects of the counsellor’s communication style: their lack of directiveness and their repetition of client-expressed issues (referred to as “scripted,” “circular,” “robotic” and “regurgitation” by four different clients). This theme appears related to the differing definitions of “help” that impact the identities each participant assumes, as well as the communication style of the counsellor (in Chat #23, the client states: “I just wanted someone to talk to and help me out of this rut”). Theme 2: client requests for personal information. In this first phase, in five of the eight non-rapport containing chats, the relationship is impacted by the client’s desire to know more about the counsellor. This desire is manifested in varied ways, from asking slightly 92 confrontational personal questions (“y are u doing this…getting sum credit on your resume?”) to asking process-oriented ones (“are u bored with me yet…lol”). Questions about counsellor gender are also asked in two of the chats; both requests are framed with the intention of determining whether the counsellor will understand their situation (“no offense but if you’re a dude i’m not sure you will get where I am coming from”). The counsellor tends to respond to these direct questions evasively, stating, for example, “I would like to focus on you right now. What’s happening for you?” This, in turn, often seems to lead to client-expressed frustration: “Are you even a human being???” Youth versus adult. The frustration with respect to the client’s unmet needs for directiveness, authenticity, and self-disclosure is present regardless of the age of the chatter. However, requests for personal information on the counsellor occurs more frequently in chats with youth. Phase 2 (suicide assessment phase): client identity. In the Suicide Assessment phase, the client seems to position themselves as someone who is both desiring privacy/anonymity and desiring help – to use a metaphor, the image of a turtle periodically retreating into its shell could visually capture the client’s identity in this phase of the unsuccessful chats. This client identity is evidenced by a pattern of admission of suicidality followed by unwillingness to answer the counsellor’s follow-up questions regarding, for example, plan, means, and intent. This unwillingness is often expressed by one-word answers or ambiguous responses; for example, “idk” [I don’t know] is one youth’s response to the counsellor’s question “can you put the knife away for the night?” Client unwillingness to elaborate is rationalized in a number of different ways, with three themes being most apparent (present in all non-rapport-containing chats): a. Concern about potential of intervention (“you’re not gonna call the cops, are you?”) 93 b. Concerns over counsellor’s ability to help (“I dunno why should I tell u this…I need help not someone to repeat what I say”) c. Concerns about counsellor’s capacity to understand (“I don’t feel comfortable telling u that I don’t know u at all”) Phase 2 (suicide assessment phase): counsellor identity. In this phase, the counsellor appears to position themselves as a collector of information – an extension of their role as “professional” which was enacted in Phase 1. This phase is characterized by a larger number of questions on the part of the counsellor (an average of six across all eight non-rapport containing chats) which are interspersed with reflective and/or clarifying statements: (“you mention you imagine a lot of pain in the process…I'm wondering if you have a way of killing yourself you have planned?”). The questions are posed using more formal language than is used in the rapport-containing chats (lack of emoticons and abbreviations; greater use of multi-syllable words frequently used in corporate contexts, such as “intervention,” “professional,” and “consequently”). Youth versus adult. The themes observed with respect to the client’s identity in the suicide assessment phase of the non-rapport-containing chats are similar regardless of the age of the chatter; however, upon closer inspection and discussion, it was noted that the anger and withdrawal were more explicitly expressed in chats with youth. Phase 2 (suicide assessment phase): client-counsellor relationship. In the second phase of the non-rapport-containing chats, the relationship between client and counsellor is marked by tension on the part of both parties. Two main themes were found in this phase as they relate to the client/counsellor relationship: misunderstanding, and a sense of feeling unheard. 94 Theme 1: misunderstanding. With respect to the first theme, the tension between client and counsellor seems to partially stem from the client’s perception that the counsellor has “missed” key parts of their story. This is, in some chats, manifested as a simple misunderstanding (“no that is NOT what I meant, I am not stressed because of school I am stressed because I am LONELY!”). Theme 2: Feeling unheard. In other chats – notably, Chat #19 – the client expresses the same dissatisfaction with the counsellor’s repetition of their content that was expressed in Phase 1 (circularity); however, the reasons for this dissatisfaction are elaborated on and point to a need to feel, in one client’s words, “not just heard but understood.” For example, in this same chat (Chat #7), the client further states: I feel like I am heard. I just don’t feel I am being listened to and thought about […] in real life we use our ears to hear and our hearts to listen […] what I feel you are doing now is only using your ears to hear and then regurgitating what you heard. Further tension emerges when the counsellor reflects this back, and the client comments on this process as an observer might, noting the irony of the exchange: “Even as you’re saying you understand that I feel it’s unhelpful, you are continuing to say ‘what I hear you say is.’” Thus, the relationship between client and counsellor seems to be one of both misunderstanding and “stuck-ness” within a particular identity on the part of both counsellor and client. Youth versus adult. There does not seem to be a significant difference with respect to how the the client/counsellor relationship is negotiated in the chats in this phase. Phase 3 (termination phase): client identity. In this phase, the identity of the client appears to circle back to the “consumer” identity first enacted in Phase 1 of the non-rapport-containing chats (“Well I guess u did ur job but let me tell u I could do it better...plz send this 95 feedback k bye”). Additional evidence of the adoption of this role in Phase 3 appears illustrated by the client’s assumption of increased social power via their initiation of the end of the chat (“I am gonna log out now”). Phase 3 (termination phase): counsellor identity. In this final phase, the counsellor seems to merge the identities of “service provider” (providing descriptions of the limitations of their role and positioning self as representative of a larger organization) and “concerned parent” (expressing concern about ability to stay safe; making suggestions; providing guidance and/or reassurance). What is interesting about these is the social power implicit in the adoption of these two roles, although they differ in terms of activities and tasks (outlined below). Youth versus adult. Other than use of more informal language (emoticons and abbreviations), by youth clients, the ways in which client identities were enacted did not seem to differ based on client age. Phase 3 (termination phase): Client/counsellor relationship. As noted above, the client’s identity in this phase appears to be one of a person with more social power, and this identity impacts how the client/counsellor relationship is negotiated. The predominant theme with respect to the client-counsellor relationship in the Termination phase is thus one of client rejection of the counsellor, which shifts the balance of social power, with the client initiating the end of the chat in all eight non-rapport-containing chats. Theme 1: client rejection of counsellor. This shift in social power is evidenced by the client’s initiation of the end of the chat and the counsellor’s attempts to keep in contact (“I’m still worried about you; can you stay a few more moments?”). Further evidence of this shift in social power is observed by the client’s adoption of the “consumer” identity “this was not what I was looking for, take care now [logs off].” The online nature of the interaction means that the 96 client is then able to abruptly sign off. In this phase, the relationship’s casual, time-limited nature is highlighted by the absence of typical social conventions indicating that a conversation or encounter is finished (“Goodbye,” “Take care now,” etc.). Youth versus adult. Client age did not appear to impact the way in which chats were terminated. Activities-Building Tool With respect to the activities the client appears to want to engage in, the theme of “help” becomes salient in the non-rapport-containing chats: the client appears to want to initiate the activity of “being helped.” The counsellor’s main activity appears to be “assessment” as evidenced by the high number ofsuicide-related questions asked in all three phases. Therefore, the activities desired by client and counsellor appear to be different, thus leading to conflict. Furthermore, there appears to be a sharper distinction between all three phases in the unsuccessful chats and a shorter duration of phases as a result. In many unsuccessful chats, this pattern of shorter phases is manifested by a rapid shift between phases and a pointed focus on suicide risk assessment in response to a client’s indication of emotional distress and/or verbal “red flags” such as expressions of hopelessness. When compared to the rapport-containing chats, where there appears to be more exploration/validation of non-suicide related content (friends; coping skills; current conflicts), in the unsuccessful chats, the Initial Contact phase very quickly transitions into the Suicide Risk Assessment phase, which quickly turns into the Termination phase, often inititated by the client after a comment about the perceived unhelpfulness of the interaction. 97 Figured Worlds Tool Disourse analysis, as detailed above, is often concerned with illuminating that which is implicit and unspoken. The figured worlds tool examines what a given discourse tells an observer about the world it occurs in (Gee, 2011). Using this tool and paying attention to the implicit message, every client statement of dissatisfaction could be reframed as a wish or need that has been unfulfilled by the current interaction. If the themes contained in the statements clients made at moments of disconnection in the non-rapport-containing chats were reframed as wishes, they could be stated as wishes for the following: a. Forward movement (versus stagnation/circularity) b. More informal dialogue c. Counsellor authenticity d. Concrete solutions e. Human connection f. A feeling that they [client] are important While there are nuances in the way in which the relationship is built across phases, these seem to be more a matter of degree than type in the non-rapport-containing chats, with the overall theme appearing to be one of client-counsellor disconnection. Below is an example excerpt from a non-rapport-containing chat to illustrate the coding process. There seem to be a number of themes present in the transcript below. Three seem particularly salient as they pertain to rapport and are noted in Figure 5 below: (1) Criticism of manner of manner of communication/misunderstanding (highlighted in green) 98 (2) Expression of desire for “help” and implying that the services is not meeting this need (highlighted in blue (3) Counsellor reference to client activities/goals are highlighted in yellow. Figure 5: Sample coding of non-rapport-containing chat. Client 19 8:45:46 PM but i promised someone i would try to talk to somebody and i've tried here before but like now, it's very much got a scripted sense to it Volunteer20 8:46:01 PM it sounds like you feel lonely and the fact that you’re in a hard financial situation it makes it even harder to try to find someone to talk to Client19 8:49:08 PM i am unable to connect with people via phone, it's too difficult. it's hard to talk to people i don't know. this place is difficult as you are mostly telling me that you understand what i just said and how it must be hard. but that doesn't really help. Volunteer20 8:55:39 PM so you mentioned that talking on here hasnt really helped you in the past and im curious what brought you on to chat tonight? Client19 8:56:15 PM i cant find another online resource that is close to home but i had a specific question Client19 8:57:18 PM which is where can someone with a low income hope to find assistance in talking to someone about why they should pursue an existence? Volunteer20 8:57:58 PM it sounds like you feel because of your low income your wondering what reasons are their to live anymore Client19 8:58:07 PM no i'm wondering how i can afford to try talking to someone don't really understand how this really helps people if you go in circles 99 Figure 5 above illustrates several themes that emerged across cases in the non-rapport-containing chats. Namely, there appears to be a sense of disconnection between client and counsellor; the client’s needs (for “help” in this case), does not appear to be heard or met by the counsellor. Furthermore, the criticicism on the part of the client with respect to the nature of the communication (scripted; robotic; circular) that is seen in Figure 5 appeared across all phases of the non-rapport-containing chats. Therefore, the question asked by the Figured Worlds tool (“What can be assumed about the counselling relationship?”) could be answered describing a relationship characterized by disconnection, frustration, and mismatched needs and expectations. Tables 7 and 8 (below) detail the specific differences between chats containing rapport and those that do not. They illustrate a number of dichotomies: informality versus formality, fluidity versus rigidity; time-sensitive versus open-ended, and egalitarian versus differing social power. These dichotomies are present in each phase, and each DA “tool” was able to illuminate a particular facet of their expression. For example, when examining the counsellor’s positioned identity, activities undertaken, and how the relationship is constructed and maintained, what is apparent when comparing Table 7 and Table 8 (on the following page) is the lack of change in client and counsellor identity across phases in the non-rapport-containing chats. This seeming rigidity appears echoed in the type of relationship that is created across phases. 100 Table 7 Rapport-Containing Chats: Themes by Phase/DA Tool Counsellor Identity (Identities-Building Tool) Phase 1: Witness Phase 2: Witness; collaborative emotional first-aid responder Phase 3: Witness; collaborative emotional first-aid responder; friend/peer Co: “Seems like it's just been a really rough road right now… and you've been dealing with this for what seems like ages” Co: “lets figure out how we can plan for the night to not suck as much” Client Identity (Identities-Building Tool) Phase 1: person needing help + beyond it; person with story to tell Phase 2: Conflicted individual: wanting to stay safe + doubting ability to do so Phase 3: Friend/peer (honesty; resignation; acceptance) Ct: “I get mad when people say suicide is selfish…I think it’s unfair for people to suffer I will continue to suffering the future and I am not allowed to say “stop..” Co: “Well I’m doing my best not to add to that burden right now.” Ct: “Haha that’s ok I understand your position.” What activity is this communication facilitating or enacting/How is this activity related to the client-counsellor relationship? (Activities-Building Tool) Fluid counsellor movement between phases Longer duration of phases; less focus on risk assessment “Social conversation” phase at end of Termination phase in seven chats Co: “If you like, we could even call you and check in? So that way all you'd have to do is answer, rather than it being really hard picking up the phone. it's not always easy to call when you feel so amped up and on edge” What kind of relationship are client and counsellor building through language? (Relationship-building tool) Phase 1: Informality; equality and respect Phase 2: Collaboration; counsellor genuineness Phase 3: Shared humour; client trust; counsellor self-disclosure Co:“it seems like you feel just fed up and, frankly, pissed off with the way things have aligned themselves” Co: “Wow, never had anyone ask me that before. Let me think about that a sec” Ct: “I get it…my mind goes blank sometimes too :s:” Tensions or contradictions? Phase 1: Conflict between client needs and counsellor role resolved through humour and counsellor transparency Phase 2: Counsellor concern for client safety; client ambivalence Phase 3: Counsellor acceptance of client’s choice of coping/ current emotional state Ct: “I don’t really want help I just want to go do what I need to do.” Co: “I get that. You’re not looking for a solution; you feel tired of it all.” 101 Table 8 Non-rapport-containing Chats: Themes by Phase/DA Tool Counsellor Identity (Identities-Building Tool) Across Phases: Witness/Impersonal, “professional” helper/ “customer service” persona Co: “seems like you’re wanting something more personal like advice or encouragement. Unfortunately, that’s beyond what we can do, simply because the service you’d get between volunteers would be different.” Client Identity (Identities-Building Tool) Across Phases: Consumer; person wanting answers; person wanting to be heard; person wanting a reciprocal interaction Ct: “this place is difficult as you are mostly telling me that you understand what i just said and how it must be hard. But that doesn't really help.” What activity is this communication facilitating or enacting/How is this activity related to the client-counsellor relationship? (Activities-Building Tool) Shorter duration of phases Sharper delineation of phases Counsellor main activity: assessment. Client main activity: help-seeking; support-seeking Co: “We’ve been talking for a while now and it looks like you have a plan to stay safe. I’m going to have to log out shortly.” What kind of relationship are client and counsellor building through language? (Relationship-building tool) Phase 1: Criticism of counsellor communication style; requests for personal information Phase 2: Misunderstanding; client feeling unheard Phase 3: Client rejection of counsellor; client termination of chat Ct: “I really want to kill myself. i guess im too afraid too. but i want help because i hate the way life is. i hate not knowing and i hate going through bumpy roads like how bad things happen then good stuff does then bad again. its so exhausting.” Co: “You've said a couple times that you want to kill yourself, is that something you're planning to do tonight?” Tensions or contradictions Client’s positioned identity and counsellor’s appear incongruent: client appears to seek: guidance, structure, genuineness Client perceived conversation as circular; robotic Ct: “I feel like I am heard. I just don’t feel I am being listened to and thought about.” Ct: “What kind of robot response is that.” 102 Gender As noted in the section describing the characteristics of the sample, gender of the client was not focused on in this study due to lack of data regarding client gender in slightly less than half of the chats sampled (only 17 out of the 24 clients disclosed their gender), as well as the gender-neutral pseudonyms given to the crisis centre volunteers. However, some preliminary observations are noted in order to describe the data set and findings as fully as possible with the aim to fulfilling the criterion of credibility in qualitative research (Morrow, 2005). It is noted that, in the non-rapport-containing chats, more males compared to females were represented, whereas in the rapport-containing chats the distribution of genders was more equal. However, the smaller size of the sample of non-rapport-containing chats as well as the presence of a number of chats where gender was undisclosed preclude the ability to report results comparing genders in terms of how it impacted the client-counsellor relationship. The most salient finding with respect to gender in this study therefore appears to be the way in which gender was approached by both parties and the impact on the client-counsellor relationship. Counsellor gender arose as a potential threat to rapport in three chats: one rapport-containing chat and two non-rapport-containing chats. This took place in the Initial Contact phase of the chat and in all cases took the form of the client asking the counsellor their gender (counsellors at the Vancouver Crisis Centre are provided gender-neutral aliases but are permitted to disclose their gender if asked). In all five chats where gender was referenced, asking questions about counsellor gender were positioned as concerns about client/counsellor similarity (“if you’re a dude I just dunno if u get it”). 103 5. DISCUSSION Client/Counsellor Identities The dialectic inherent in the client’s positioned identity (that of a person both wanting help and being beyond it) was seen consistently in the Initial Contact phase of both the rapport-containing and non-rapport containing chats. It would therefore seem that the limitations of the crisis counsellor’s role have the potential to impact rapport. Therefore, how these limitations are explained and addressed becomes important, particularly in the first phase of the chat, where the norms of this particular discourse are established. The counsellor’s role as empathic witness in the Initial Contact phase of the rapport-containing chats may have implications for the rapport that is established in this vital first point of contact. Individuals (especially those accessing for the first time, as was the case for half of the clients in this study), likely experience feelings of vulnerability, anxiety, and shame associated with the stigma surrounding suicide (Fullagar, 2003). Therefore, the counsellor’s positioning as a service provider in the non-rapport-containing chats may be interpreted as distancing, particularly given the online nature of the interaction and the absence of non-verbal cues such as eye contact and body language, as well as, in some cases, demographic information such as gender. On this note, upon examining Table 3 which details the demographic information from non-rapport-containing chats, it is evident that these chats contained less personal information (five of the eight clients left out their name, gender, location, or all three). This could potentially mean that these individuals were more concerned about their anonymity, (implying that they were entering the chat with a pre-existing lack of trust in the counsellor or the counselling process), resistance to disclosing sensitive information, or a heightened degree of suicide risk. It 104 could also be interpreted as a desire to assert control of the interaction by refusing to share this information, which possibly implies a perceived lack of control. This observation highlights the importance of the Initial Contact phase in building trust, clarifying roles, and refraining from overly clinical language. Importance of Client Narrative How the client’s narrative is received by the counsellor in the Initial Contact phase appears to have implications for how the Suicide Assessment phase is negotiated. It is observed by Fullagar (2003) and Elliott et al. (2005) that having the depth of their suicidal feelings acknowledged is vitally important to those considering suicide; Fullagar (2003) notes that the exploration of suicide necessitates that the person working with the suicidal individual must “enter those dark spaces of existential dread that we mostly try to avoid” (p. 295). The counsellor’s positioning as non-judgemental witness and willingness to hear the client’s story in the Initial Contact phase may assist in dispelling this initial hesitation and undoing some of the shame associated with the stigma of suicide. The finding of temporal markers (references to time) in the client’s statements in the Initial Contact phase is of interest because it occurred in both rapport-containing and non-rapport-containing chats. It is consistent with the observations of current theorists in the field of suicidology (e.g., Valach et al., 2011) who cite research indicating that while “professionals” in the field tend to think of suicidal acts in terms of unilinear causes (depression; psychopathology), suicidal individuals tend to explain their actions in terms of motives or reasons for action. These explanations are often embedded in a meta-narrative that the counsellor would do well to pay attention to. For example, a suicide “project” may be embedded in a relationship “project,” where the suicidal act represents communication with a person important to the client (Valach et 105 al., 2011). This was exemplified in the current study’s findings, where clients often situated their suicidal thoughts within a historical framework (e.g., history of bullying; long-term struggle with an eating disorder). The length of the Initial Contact phase in the rapport-containing chats in comparision to the non-rapport containing chats also seems to speak to the importance of this “witness” identity. Counsellor comfort in speaking about suicide and ability to accept dichotomous information or ambivalent feelings therefore seems to be an important factor in the establishment of rapport. Client/Counsellor Relationship It was previously noted that a number of implicit assumptions about the client/counsellor relationship appear to be repeatedly violated in those chats defined as rapport-containing. Conversely, they appear to be upheld in those defined as non-rapport-containing. These assumptions are noted below, with the violations observed in the rapport-containing chats placed in parentheses:  Counsellor possesses knowledge to impart (versus asking client’s opinion; positioning self as co-investigator)  Counsellor is expert and superhuman (versus acknowledging limitations or vulnerabilities)  Use of formal, scientific, or professional language (versus informal, colloquial language)  Lack of or limited self-disclosure (versus openness about own experiences and/or sharing one’s emotions)  Interpretation of client humour as “distraction” or “defense mechanism” (versus joining client in humour)  Setting: in traditional in-person counselling, client comes into counsellor’s space (counsellor does not visit client in home) The existence of these repeated violations of the dominant discourse of “counsellor as expert” raises the following question: what is the impact of this role reversal on the client? It is possible that this interaction paradoxically places the client in control of the interaction, providing a sense of agency that may be developmentally (in the case of youth clients) and psychologically needed. Furthermore, certain rules of the dominant discourse described above 106 may either not translate into online counselling (the setting, for example); others may be overly distancing when dealing with the subject of suicide. Furthermore, anecdotal evidence seems to indicate that one of the key barriers to building rapport is concern that the counsellor is a “robot,” implying a need or desire for verification of the counsellor’s humanity. In fact, in one of the excerpts from the data noted in the previous chapter, the youth client specifically (and humourously) questions whether the counsellor is a robot or a human. This apparent need for counsellor genuineness as evidenced by client requests for it in the unsuccessful chats, and their positive responses to it in the successful chats appears to reflect the importance of the “real relationship” as defined by Gelso (2009). As noted in the literature review above, the real relationship is thought to be composed of two parts: genuineness and realism. Realism refers to the extent to which client and therapist experience each other in as distinct human beings independent of the context in which the interaction takes place, and genuineness refers to the ability of both parties to be authentic in their interactions towards each other (Gelso, 2009). Gelso (2009) notes that the real relationship “[…] often transpires silently and is not the real focus of the [counselling] work” (p. 253). However, he acknowledges that it permeates and impacts everything that occurs between client and counsellor. Thus, it could be described as the relational core of the interpersonal process between therapist and client, the shaking of which has the potential to destabilize the relationship significantly. This can have life-threatening consequences when the client is suicidal. The findings of this study underscore the importance of the real relationship when conducting crisis counselling online. It may be of particular importance in the online environment due to the absence of cues present in in-person counselling. In particular, the importance of the genuineness component of the real relationship seems highlighted in the chats with youth clients. Research in the field of youth counselling has 107 found that youth in particular value and are sensitive to a counsellor’s perceived authenticity, and, on the flip side, a strong aversion to “fakeness” or perceived condescension (Martin, 2003). Social Power It was noted in the literature review section that the online relationship approaches a less professional model online, likely based on the lack of traditional indicators of social power in this context (Hanley, 2006; Wrzesien, 2014). The finding of a more informal relationship where the client was viewed as more of a collaborator or consultant seemed to be upheld by the results of this study, as per the observation of increased informality, collaboration, and mutual respect found in the rapport-containing chats. Of interest, conversely, was the more “professional” identity assumed by the counsellor in the non-rapport-containing chats. The word “professional” may invoke a sense of expertise, neutrality, and formality. The informal, vulnerable, and self-disclosing identity enacted by the counsellor in the rapport-containing chats seems to directly contradict this image. It is possible that, in an online context when talking with those who are suicidal, the identity of the counsellor needs to shift in response to the context in which the interaction occurs. Of note as well is that observation that “unsuccessful” chats were often ended abruptly by the client by simply logging off. It appeared that they were using the online space to communicate their anger by simply shutting off the conversation. In an in-person counselling environment, such emotion would likely have been communicated very differently - likely more subtly or not at all. The online disinhibition hypothesis (Lapidot-Lefler & Barak, 2015; Suler, 2005) dictates that people are more likely to be truthful in an online environment. The comments made by the clients in the unsuccessful chats are therefore likely much more honest and blunt than they would 108 have been in-person. In fact, it is possible that much of the negative feedback received in an online environment would never have been communicated in-person, even if it were felt, due to the norms surrounding interpersonal communication in a professional environment in which one person (the counsellor) possesess greater social power. This speaks to the importance of the results of this study, as they provide a glimpse into therapeutic interactions that are not often accessible. Client/Counsellor Activities The finding that in some rapport-containing chats, the counsellor’s focus on the therapeutic relationship seemed to supercede a focus on suicide risk assessment is important to examine and represents an area of exploration and further study. It is possible that counsellor discomfort with suicide was a factor here; however, this seems less likely given the counsellor’s response to the client’s admission of suicidality in these chats, which appears empathic and validating. To reiterate from the results section, in one chat, the client expressed feeling “like every day may be the last…not sure how long i can keep this up even thinking how i would like end it all.” While the counsellor expressed non-judgemental acceptance and empathy for this experience (“I get that…you’re feeling totally exhausted from trying to survive each day”), the client’s statement also implies the existence of a suicide plan; this was not picked up by the counsellor. However, in this same chat, towards the end, the client’s level of risk seems to have been mitigated on its own, with the client stating “thanks for hearing me…i feel a lot calmer now and will just stay in and watch tv…you are awesome and thanks for doing this work.” Therefore, it could be argued that focus on the client’s suicide plan may have negatively impacted the 109 developing rapport, and that provision of empathy and validation in this case was sufficient to de-escalate the client and hence reduce their risk. On a related note, the finding of a fourth “phase” – one characterized by informal social conversation on the part of client and counsellor, is worthy of further examination. It could be that this phase provides a “buffer” or “rest” from the intensity of the conversation about suicide. Implicitly, it could facilitate rapport by communicating to the client that the counsellor is interested in more than just the client’s immediate safety, but in the context of their lives (what shows they watch; the name of their cat). It is also possible that this phase emerged due to the unique nature of crisis counselling online: it is likely that those who access these services do so at a higher level of emotional escalation than would be present if they were to access another type of (non-crisis) counselling service. The research assistant noted that the social conversation phase may indicate client de-escalation and could therefore be an indicator of a beneficial therapeutic interaction. The previously stated observation that counsellor comfort with the topic of suicide seems an important factor in establishing rapport seems relevant here. In particular, the finding that in some chats, suicide assessment seemed to be disregarded in favour of relationship-building counsellor interventions (process-oriented comments; inquiries about self-care; humour; counsellor self-disclosure), highlights the importance of balancing these two in a way that promotes client safety while still maintaining the relationship. White (2015) acknowledges the above struggles as a paradox inherent in the field of suicidology. She quotes the words of a previously suicidal individual which capture this dichotomy: “When suicide is seen as something to be prevented, honest listening — which, to me, means listening without needing to act and without needing to find an immediate answer — is deemed irresponsible or even dangerous” 110 (White, 2015, p. 1). In the rapport-containing chats in this study, counsellors appeared to strive for - and often reach - a fine balance between listening and engaging in collaborative action. In the rapport-containing chats, the use of collaborative language on the part of the counsellor (“we; “us;” together”) when speaking about safety-planning could reflect the counsellor’s implicit message that suicide prevention is a joint goal that client and counsellor are engaged in, drawing from the perspective of action theory that sees actions as part of long and short-term psychological goals (Valach et al., 2011). The observation that a “social conversation” took place following the Termination phase in seven of the rapport-containing chats also seems to violate the dominant discourse of crisis counselling as a “time-limited” or “problem-focused” therapeutic encounter. Perhaps the counsellor’s initiation of this conversation is a way of conveying “mattering” in an online realm in a more indirect manner. It is possible that by showing interest in TV shows, asking questions, and using humour, this sends the message that the counsellor cares about the client without explicitely stating it (which could come across as scripted and not genuine). Use of Humour/Informal Language The online context seems to necessitate a “humanizing” of the counsellor, which was an expected finding based on the existing current research on online counselling which was reviewed in the Introduction section. However, the informality and humour present in rapport-containing chats, despite – or perhaps because of – the seriousness of the subject matter, was surprising. The question remains whether these characteristics would be found in rapport-containing chats that did not include suicide as a presenting concern, or whether they exist as a kind of reaction (on the part of client, counsellor, or both) against the seriousness and darkness of the subject of suicide. Research indicating that use of humour is a protective factor against 111 depression via cognitive distancing and recruiting of social support (Nezu, Nezu, & Bisset) may be relevant here. From the counsellor’s perspective, the research of Cain (2012) indicates that use of humour by social workers in palliative care protected against burnout. It is possible that counsellor use of humour functioned in a similar manner, allowing the counsellor to navigate the “landscape of the beleaguered person” (Shneidman, 1998, p. 151) without becoming consumed by it. The timing of the use of humour seems important to note as well – specifically, use of humour seemed mainly confined to the Termination phase in the rapport-containing chats. It is possible that use of humour too early in the chat may be potentially distancing or disconnecting, as it may imply that the client’s story or concerns are not worthy of serious exploration. Use of humour in the Termination phase may also be an indicator of client de-escalation in a similar manner as the Social Conversation phase seemed to be. The difference between adults and youth with respect to use of informal, online-specific language, while not large, nevertheless points to a possible need for the counsellor to be more alert to patterns of language and to amend their communication style to match that of the client. However, it must also be noted that younger clients tended to take more offense to “scripted” or “fake” dialogue; thus, it seems that balancing authenticity with mirroring of language patterns must be attempted, although it is recognized that this is a delicate “dance” and that further study is needed that includes the voices of the people being impacted, as per White (2012). Counsellor Self-Disclosure The finding that counsellor self-disclosure emerged as a theme in the rapport-containing chats warrants a closer examination of this construct and how it emerged in the data, since it is counter to the training received by crisis line volunteers and hence a phenomenon that is, in line 112 with discourse analysis, worthy of further examination by virtue of its’ “strangeness” (Gee, 2011). A recent meta-analysis found that counsellor self-disclosure is associated with positive client outcomes, including perceiving the counsellor favourably and, of particular interest to this study, increased client self-disclosure (Henretty, Currier, Berman, & Levitt (2014). The finding of increased counsellor self-disclosure facilitating client self-disclosure (the authors theorize that this may occur via modeling as per social learning theory) seemed to be replicated in this study and is of particular importance for individuals who are suicidal. Namely, the research on suicide prevention and assessment outlined in the literature review highlights the importance of obtaining relevant client information in order to facilitate risk assessment and treatment planning (e.g., Brown et al. 2004; Garb, 1998; Shea, 2008; Shneidman, 1998). However, as noted by Henretty et al. (2014), not all counsellor self-disclosures are created equal. Specifically, they outline the three most commonly investigated types of counsellor self-disclosure (Henretty et al., 2014): a. Intra-therapy (disclosures regarding the client or therapy process) versus extratherapy (disclosures regarding the counsellor’s life outside of therapy) b. Positive or negative content valence c. Whether the disclosure reveals similarity or difference with the client The most frequently used counsellor self-disclosures in this study were positively valenced extra-therapy disclosures that implied client-counsellor similarity (such as referring to books read or taking a particular course in school). These were found in seven of the 16 rapport-containing chats and in none of the non-rapport-containing chats. However, five rapport-containing chats contained negatively valenced extratherapy counsellor disclosures (e.g., self-113 reported anxiety regarding picking up the phone when overwhelmed). Of note was the finding that all of these of these negatively valenced counsellor self-disclosures implied similarity with the client. Four of the rapport-containing chats contained negatively valenced intra-therapy counsellor disclosures (“to be honest, I’m feeling a bit stuck right now, as you’re saying you’re not sure you’re safe and I’m worried about you”). It is possible that the online context renders counsellor self-disclosure more important, given the absence of non-verbal cues. Since the client only has the counsellors (unisex) pseudonym to refer to when constructing a mental image of them, the salience of small, personal details of the counsellor’s extratherapy life may be increased and have implications for the development of rapport. The behaviour of the counsellor in the rapport-containing chats is consistent with the online disinhibition hypothesis (Lapidot-Lefler & Barak, 2015; Suler, 2005), which states that individuals are more likely to disclose in an environment where their identity is concealed, but this hypothesis does not explain the finding that counsellor self-disclosure did not occur in the non-rapport-containing chats. It is possible that the client’s positioned identity in the non-rapport-containing chats inhibited counsellor self-disclosure. Common Themes: Mattering A common theme that was found in all chats in this sample was the presence or absence of client-perceived importance of self. Specifically, in the non-rapport containing chats, the client frequently expressed feeling unheard, misunderstood, and disconnected from the counsellor; in the rapport-containing chats, the inverse seemed to be true. The sociological construct of “mattering” could be used as a theoretical base for explaining these observed patterns in all three phases of the rapport and non-rapport containing chats in this study and has relevance for application in the fields of both suicide prevention and online counselling. 114 Mattering has been defined by Rosenberg and McCullough (1981) as a felt sense (composed of both affective and cognitive components) that one is important to others and that others are interested in one’s activities, accomplishments, and emotional state. Rosenberg (1980) believed that all people experience mattering in varying degrees. He also distinguished between general mattering (mattering to society in a larger sense) and interpersonal mattering (mattering to specific individuals or within a certain social context). Mattering appears to facilitate a feeling of belonging and of being appreciated (Rosenberg & McCullough, 1981) and hence seems to tap into both individual and community-based needs. Rosenberg and McCullough (1981) also note that mattering can be distinguished from social support by the presence of interest in a person’s welfare that goes beyond mere provision of internal or external resources. Social support, on the other hand, has been defined as specific forms of emotional or informational support and can theoretically be offered without the person feeling a sense that they matter (Elliott, Colangelo & Gelles, 2005). Schlossberg (1989) extended this construct by separating it into four separate dimensions, summarized by Corbiere and Amundson (2007, p. 142) in terms of their applicability for the counselling relationship: (1) Attention: the feeling that someone notices or is interested in you (2) Importance: the sense that people are interested in what you are thinking, feeling, and doing. This aspect is communicated when people inquire about your well-being and take time to listen carefully to what you have to say (3) Dependence: The feeling that your contribution is valued and needed (4) Ego-Extension: The feeling, whether right or wrong, that others are interested in how you are doing (accomplishments as well as disappointments) even when there is no longer a professional relationship [italics mine] The construct of mattering appears to be a relevant meta-theme based on client-counsellor interaction in both rapport-containing and non-rapport-containing chats. Specifically, the 115 counsellor appears to facilitate a sense of “mattering” to the client in the rapport-containing chats, whereas the specific client feedback in the non-rapport-containing chats appears to reflect a felt sense of a lack of mattering on the part of the client. Thus, the presence or absence of client-perceived mattering seems to have had an impact on the degree and valence (positive or negative) of the therapeutic relationship. This is evident in all three phases of the chat. For example, in the Initial Contact Phase of both rapport-containing and non-rapport-containing chats, the client’s adoption of the identity of a person with a “story to tell” (as evidenced by multiple temporal markers in their narrative) recruits the counsellor into the identity of “witness.” In the rapport-containing chats, the counsellor appears to facilitate mattering by showing a genuine interest in the client’s story (by using a combination of empathic reflections, open-ended questions, and occasional self-disclosure) and a willingness to give the client space to elaborate and share details of their story. Conversely, in the non-rapport-containing chats, the swift transition from Initial Contact to Suicide Assessment phase implies a more focused attention on the client’s risk factors and less focus on the context in which they arose (the client’s story). This counsellor behaviour, while likely engaged in with the intention of facilitating client safety, may adversely affect all four dimensions of mattering described above. The dialectic observed in the client’s identity through the Initial Contact and Suicide Assessment phases is one of both wanting help and being beyond it. It is possible that, in the rapport-containing chats, the experience of mattering shifts the client’s identity from being beyond help to believing help may be possible. An alternative, although perhaps overly parsimonious, explanation is that the experience of mattering to the counsellor is, in itself, the help that the client needs. This explanation is consistent, however, with Gelso’s work on the real relationship as a strong extratherapeutic factor (2009), and with much of the common factors 116 research that indicates that in addition to a strong client-counsellor relationship, feeling that one matters to one’s counsellor (perceived unconditional positive regard) explains as much of the variance in therapeutic outcomes as does the application of specific techniques (Lambert & Barley, 2001; Timulak, 2007). In the Suicide Assessment phase, “general mattering” (mattering to society in a broader sense) seems more salient as the client frequently invokes statements that imply a sense of burdensomness or disconnectedness from society. This expressed lack of purpose or meaning is consistent with current theories of etiology of suicidal behavior, from Durkheim’s initial theory of lack of social integration or belonging in society, to Shneidman’s notion of psychache (defined as frustrated psychological needs that are often interpersonal in nature, such as for belonging or social contact). On this note, White (2012) questions current dominant discourse of suicide originating from pathology in the individual. In fact, as noted in the Introduction section, suicide was originally explained by invoking sociological factors (Durkheim, 1951). It is not surprising, therefore, that the findings of this study can be partially explained by invoking explanations that involve relational factors. In the Suicide Assessment phase of the rapport-containing chats, the counsellor seems to facilitate a sense of general mattering by facilitating interpersonal mattering by focusing on the here-and-now therapeutic interaction (“I wish I knew what to say to you here… I feel really worried about you with all those pills on hand”). Thus, working within his or her sphere of influence, the counsellor appears to provide a corrective emotional experience that may countract the experience of non-mattering that the client experiences in the world and that is often cited as a trigger or rationale for their thoughts of suicide (“no one cares if i live or die”). The clients’ expressed appreciation of the counsellor’s communication of mattering in the Suicide 117 Assessment phase (“thanks for making me feel like i’m not a waste of space”) is consistent with Joiner’s (2005) interpersonal theory of suicide, which posits that suicidal behaviour arises of the interaction between thwarted belongingness and perceived burdensomeness, both of which imply a lack of what Rosenberg and McCullough (1981) would term interpersonal mattering. It is also consistent with current research on how thoughts of suicide are experienced and processed: Fullagar (2003) conducted in-depth interviews with 41 young people who had been exposed to suicide, either through the suicide of an acquaintance or through their own attempt. One youth’s response seems to encapsulate the the need for mattering and offers clues for how it may be met in the counselling relationship: Young people are taking their own lives because they are not happy. Because they don’t see any other way because there’s no one offering them help and people don’t ask for help because like why would anyone want to listen to what I’ve got to say […]. No one else knows how I’m feeling […] it’s the stigma of it [suicide attempts], being known as the person who swallowed the Panadol […] I think often things don’t get seen […] (Fullagar, 2003, p. 301). In the above citation, the participant, similar to the clients in the present study, seems to express the need for help and her perception that designated “helpers” may not possess the qualities that would invoke a sense of mattering. Specifically, she seems to equate being helped with being both listened to and being seen (which could be tentatively reframed as wanting the intensity of her suffering acknowledged) – key components of mattering. The findings of this study thus illuminate the potential therapeutic value of situating and supporting the suicidal individual within their life context in a way that facilitates both general and interpersonal mattering. White (2012) suggests that solutions for the problem of suicide should be expanded 118 from solely intra-personal factors to include an individual’s socio-cultural life context. This suggestion is consistent with the recommendations of Valach et al. (2011) who encourage clinicians to view a client’s suicidal thoughts or acts not as merely symptoms of a psychological disorder but instead to view these individuals as constructing suicide as a goal-directed action within their current life context. From an action theory perspective (Valach et al., 2011), understanding the goal of suicide for the individual (e.g., release, escape, revenge), can therefore help the clinician engage the client in exploration of how these needs and their associated emotions may be met or managed in a safe manner. The client-perceived circularity of communication found in this study (“don’t just repeat what I’m saying”) in the non-rapport containing chats likely impacts all four dimensions of mattering by communicating to the client that the counsellor is unable or unwilling engage with them in a way that involves forward movement, or the meeting of their expressed goals. However, paradoxically, the client’s identity as seeking a witness or someone to hear their story points to a need to focus on listening rather than action. The theme of counsellor flexibility and fluidity across phases seems relevant here. This flexibility, along with authenticity and informality, may be factors that influence whether the counsellor’s identity as a witness is perceived as circular and frustrating or validating and reassuring. Age Differences As noted in the Results section, there did not appear to be large age differences with respect to how the client-counsellor relationship was negotiated online. The differences observed across phases appeared to be differences of degree rather than type of interaction. Specifically, it appeared that younger chatters seemed to prefer more directiveness, informality, and counsellor self-disclosure. This finding is consistent with the research of Geldard and Patton (2007) who 119 found that young people express empathy to one another by being both evaluative (giving an opinion) and directive (giving advice). Since this is not typically how counsellors are trained to interact (particularly volunteers at a crisis line who are trained in non-directive provision of empathy), and since young people may have had less exposure to the helping professions, this may account for the more pronounced preference for this type of interaction shown by youth clients. Gender Differences As noted in the Results section, in the non-rapport-containing chats, more males compared to females were represented, whereas in the rapport-containing chats the distribution of genders was more equal. The finding that the theme of “help” (and not receiving it) as well as the positioning of the counsellor as “service provider” in the non-rapport-containing chats (in which males were more highly represented) could be partially explained by the seminal research on gender-specific communication that indicates that males are more likely to prefer instrumental support, whereas females are more likely to prefer emotional support (e.g., Tannen, 1990; Wood, 1997). However, recently this theory has been questioned, with some studies indicating no clear gender differences in preferences for type of emotional support, with both genders preferring supportive emotional support versus instrumental support (MacGeorge, Feng & Butler, 2003). The clients’ questions regarding the counsellor’s gender in the Initial Contact phase could be explained partly as a reaction against the anonymity of the online environment: research indicates that sources providing support online are considered more “trustworthy” if they provide an avatar (an image representing them), which can partially mitigate the anonymity of the setting (Taylor, 2011). This finding speaks to the potential utility of providing counsellor avatars on the online service, whether or not these are of a specific gender. The finding that the rationale 120 frequently given for wanting to know the counsellor’s gender was based on searching for client-counsellor similarity is consistent with research on the “similarity” dimension of counsellor self-disclosure which has been associated with positive client outcomes (Henretty et al., 2014). Potential Clinical Applications The results obtained point to a number of potential clinical applications; however, it s important to note that further research is needed and that the results cannot be generalized outside of the context studied. Furthermore, it is recognized that the working alliance involves the interaction of two parties, and that the client inevitably brings to the interaction a number of intra and interpersonal factors that cannot be controlled or determined pre-emptively and may have an impact on the therapeutic alliance. As discussed in the Introduction chapter, many theorists recognize that an important component of the client-counsellor relationship is the client’s previous relational history (Horvath & Luborsky, 1993). Therefore, the clinical applications in terms of potential counsellor behaviours discussed below are proposed tentatively. With that caveat in mind, a number of potential clinical applications are now discussed. The theme of counsellor interest in client activities, values, or beliefs that lie outside the realm of therapy (facilitating a sense of client mattering) has important clinical implications. It echoes the critique of Kouri and White (2014), who note that most suicide prevention work assumes that suicide occurs as a result of individual psychopathology, thus often disregarding any contextual or situational factors. If client-perceived mattering is indeed an important therapeutic factor facilitated by the counsellor in rapport-containing chats, then a relevant focus of suicide prevention online could be the exploration of sources of mattering in a client’s life as well as facilitating mattering in the therapeutic process. One example of a tool that may facilitate 121 exploration of extra-therapeutic sources of mattering is Ishiyama’s (1995) validationgram. The validationgram is an interactive activity that taps into a construct that is conceptually related to mattering but broader in scope: self-validation (Ishiyama, 1995). It involves using a circular image to assist clients in identifying various sources of validation in their lives by graphically representing them in the four quadrants of Ishiyama’s validationgram (see Figure 6 below). Figure 6: Ishiyama’s validationgram. Used with permission. 122 The validationgram could be a concrete and holistic clinical tool (and could be sent online as an image) to assist suicidal individuals in graphically representing a validation network for themselves which could be of use in moments of crisis. It could also assist in dispelling some of the impersonality of online communication by “fleshing out” the identity of the client without compromising their anonymity. Use of this technique could also facilitate intra-therapeutic mattering by showing genuine interest in the client’s life. Counsellor Flexibility Of note when considering clinical implications is the observation that the counsellors in the non-rapport-containing chats appeared to display a certain lack of cognitive and affective flexibility. This was manifested by often rigid adherence to risk assessment at the expense of the therapeutic relationship and/or continuing to engage in a behaviour despite the client’s request not to (e.g., using verbatim empathic reflections which were then perceived as robotic or scripted). This observation points to two potential solutions for crisis intervention stakeholders: either intervene at the level of training (encourage behavioural flexibility via strategies such as role-playing) or recruitment (screen for cognitive rigidity/black and white thinking). Perhaps paradoxically, the research assistant hypothesized that individuals high in the personality trait of conscientiousness might actually be “worse” at engaging in relationship-building with individuals who are suicidal online, due to a concern about “doing things wrong” and that those higher in the personality trait of openness to experience would be more skilled at relationship-building. These observations are intriguing and represent a potential area of future research. Counsellor Transparency/Authenticity Since many threats to rapport in this study consisted of client requests for information (concrete solutions or personal information about the counsellor), providing opportunities for 123 role-playing this issue in training may be helpful, as would providing examples of ways this was handled effectively. On this note, with respect to the client’s expressed need for counsellor authenticity and directiveness, it is recognized that encouraging crisis line volunteers to be authentic and direct comes with inherent risks. For example, some volunteers may either inadvertently cause harm by doing so (by being too direct or evaluative), or may create a climate of dependence on the service in the case of providing direct advice. Therefore, perhaps the safest course of action would be to encourage crisis counsellors to strive for authenticity and clarity when delineating the limits of their role. To use an example of such a response to a request for concrete solutions from Chat # 4 (which was rated very high in empathy; see Table 4): Co: “I totally get that you want answers. I’d want them too, in your shoes. It sucks that I can’t give you that right now, but let’s see what we can do to figure out some options.” Ct: “Haha no it’s ok. I get u can’t tell me what to do. Honestly if U did I’d probably not listen anyhow...guess I just want a space to vent.” Co: “thanks for getting that. yeah well you definitely came for the right place for that [venting]…sounds like there’s a lot on your mind.” On a related note, the finding that the concept of “help” was often differentially understood by both parties in the non-rapport-containing chats points to a need for the crisis counsellor to clearly and respectfully delineate the limits of their role and to show empathy and curiousity when exploring what “help” means to a given client (as occurred in the rapport-containing chats). Professional or “scripted” language should perhaps be used with caution in these instances, given the findings of this study; however, these suggestions are tentative based on the scope of the study, as outlined in detail below. 124 Limitations It is recognized that the nature of the data (text transcripts) limits the inferences that can be made regarding the nature of the therapeutic relationship. Specifically, the inability to contact participants (clients and crisis counsellors) for further verification/clarification represents an important limitation, which was mitigated in part through the use of research assistants and experts in discourse analysis who were consulted with respect to the generation of categories and themes. On a related note, the inability to use distal evaluation measures (follow-up with the individuals), prevented the researchers from obtaining potentially valuable information regarding any lasting therapeutic effects from the crisis chat. Additionally, as noted previously, it is recognized that certain demographic information (e.g., age, gender, location) were based on client self-report and may not be accurate. Similarly, the use of gender-neutral counsellor pseudonyms prevents any conclusions related to how counsellor gender impacts the establishment of rapport. It should also be reiterated that the social-constructivist epistemological frame and the qualitative approach to data analysis precludes the generalizability of the findings to contexts beyond the one studied, which is consistent with the aims of qualitative research, which involves an in-depth exploration of a given phenomenon (Grbich, 2011). Another important potential limitation concerns my dual role: I had previously been supervising crisis chat volunteers at the location studied. Therefore, my role as a professional in this context may have impacted my examination of the data in my role as a researcher. However, it must also be emphasized that I would not have access to this rich and important data set were I not connected to the Crisis Centre in this capacity. Therefore, this potential conflict of interest must be viewed in light of the potential gains in knowledge that may be made through this study. 125 Furthermore, the Vancouver Crisis Centre agreed to limit my role to avoid any contact in a supervisory capacity with the assessment or examination of chat transcripts for the duration of the research project – see Appendix D for the letter provided to the ethics board detailing the Crisis Centre’s involvement in this project. Furthermore, as delinated in the section outlining researcher reflexivity, it is acknowledged that since discouse analysis is concerned with looking for exceptions to a given discourse, issues of social power, and social constructions of given constructs, these things would be highlighted in the themes found, and that a different type of analysis may have yielded different themes. Directions for Future Research As noted in the preceding section, one of the limitations of the current study was the inability to follow up with participants and to obtain their view of what transpired in a given chat. My role at the Vancouver Crisis Centre prevented me from conducting follow-up interviews with volunteers, and the anonymity of the service prevented follow-up with the clients who used the online service. White (2012) emphasizes the importance of recruiting the perspectives of those struggling with suicide in order to better understand it from “within.” Therefore, future studies could employ a participatory action research method (PAR; Morrow et al., 2012), perhaps using focus groups with individuals who have used the crisis chat services or have experienced a crisis where they might have used such services in order to explore how the client/counsellor relationship is experienced by the client. Similar investigations could be conducted exploring the perspectives of online crisis counsellors. With respect to the non-rapport containing chats, it would be interesting, in future studies, to examine the feedback provided by clients (e.g., that the counsellor was scripted, 126 robotic, or disinterested) to previously suicidal clients who had received in-person counselling and determine whether these complaints were replicated in in-person sessions. This would help illuminate whether the online environment necessitates a particular stance or particular behaviours on the part of the counsellor in order to facilitate rapport. It is also important to acknowledge that those accessing online crisis counselling (especially those at high risk) may benefit from the increased safety and potential for counsellor intervention provided by in-person or telephone counselling, and, where ethically appropriate, should be encouraged. On a related note, the question remains whether the relational themes observed would have been found in chats that did not include suicide as a presenting concern. Further studies could expand the scope of investigation to include some of the “secondary” concerns experienced by clients in the current study, such as depression, relationship conflicts, and anxiety. Since these concerns appeared to precipate (in conjunction with other life events) the client’s suicidal thoughts in a number of chats analyzed, they seem worthy areas of further investigation in terms of how clients experiencing them can be helped online. With respect to the finding that, at times, risk assessment was missed in favour of rapport-building counsellor interventions (e.g., counsellor self-disclosure; humour), future studies could attempt to shed light on the process by which volunteers make decisions to engage in particular clinical actions; volunteers could be provided with their transcripts and asked to walk researchers through their decision-making process in a given chat. However, it is recognized that this procedure has the potential to evoke fears of being negatively evaluated and/or defensiveness on the part of the volunteer, so such studies would have to be structured with caution. 127 6. CONCLUSION This study used discourse analysis to explore how the client-counsellor relationship was navigated in online chats containing suicide as a presenting concern. The results and their potential impact are viewed with the same social constructionist lens that was used to examine the study’s data sources. Namely, particular discourses or “figured worlds” such as the ones described in this study are viewed by discourse analysts as constantly in flux and amenable to change. Gee (2011) notes that “figured worlds are not static,” recognizing that discourses are subject to changing social times. It is hoped that the findings may be used to potentially shift the dominant discourse of online crisis counselling in a way that is helpful to both clients and counsellors engaging in this important work. The findings and their potential implications are now summarized. Content analysis followed by discourse analysis found that, in the first two phases of the rapport-containing chats, the client tended to express a dialectic of wanting help and simultaneously feeling hopeless, as well as of a person with a story to tell, and the counsellor positioned themselves as an empathic witness to the client’s narrative. The client-counsellor relationship in Phase 1 (Initial Contact) was characterized by themes of informality and equality/mutual respect. Main themes in the client-counsellor relationship in Phase 2 (Suicide Assessment) included client/counsellor collaboration and counsellor vulnerability. In Phase 3 (Termination), main themes in the client-counsellor relationship included humour, counsellor self-disclosure, and client trust. Interestingly, in seven of the rapport-containing chats, the counsellor appeared to initiate a fourth “phase” in which informal topics were discussed; this phase was termed “social conversation” and was hypothesized to convey a sense of mattering to the client. 128 In the non-rapport-containing chats, in the Initial Contact phase, the client positioned themselves as a consumer of services and the counsellor positioned themselves as a service provider. The client-counsellor relationship in this phase was characterized by client frustration and counsellor helplessness with respect to the client’s unmet needs for counsellor directiveness, authenticity, and self-disclosure. In the Suicide Assessment phase, three main themes were found: client dissatisfaction with perceived circularity of the conversation, client feeling misunderstood, and the client feeling unheard. Of note in the Termination phase was client-initiated abrupt termination of the chat, which contributed to the observed theme of counsellor rejection in this phase. Across all phases of all chats, counsellor authenticity and client-experienced mattering (or lack thereof) were major themes. The results found speak to the complexity, ambiguity, and contradiction often seen in the field of suicidology. As astutely observed by Kouri and White (2014), the field of suicidology could benefit from encouraging counsellor flexibility, tolerance of ambuiguity, and willingness to work with ambivalence and contradiction. The results of this study suggest that the online realm is a place where these qualities are also important in building rapport with individuals who most need it. It is hoped that this study has provided a critical, nuanced view of how a youth or adult in suicidal crisis connects with their online counsellor “in the moment” and in doing so, can shift the discourse of online crisis counselling by informing the practice of those who work with suicidal individuals using this modality. Understanding how those accessing services online conceptualize suicidality and the rapport-building process has important implications for further research and practice. The exploratory nature of this study situates its potential results within the discovery rather than the verification stage of research. It is hoped that this research will 129 facilitate a greater understanding of the needs of those in suicidal crisis who seek support online. If helpers of these individuals, (e.g., social workers, mental health clinicians and crisis counsellors) have an idea of “what they [clients] need” in order to build a connection in the moment of crisis, they will be able to more effectively work towards designing and providing appropriate sources of help. 130 Epilogue When representing qualitative results, Grbich (2013) emphasizes the utilization of multiple ways of data representation, citing the differing ways that knowledge can be communicated as greatly increasing the trustworthiness of the findings. Therefore, the following metaphorical representation describing the client/counsellor relationship was arrived at. The relationship between client and counsellor across phases in the rapport-containing chats could be described by the metaphor of two hikers lost in the woods, one of whom has slightly more experience with the terrain, the other of whom is injured. Both are navigating the same rough terrain, but the counsellor is the more experienced hiker who may have access to emergency help and whose non-injured status enables them to facilitate this. However, the isolation of the setting (the online context), renders such help impractical or delayed in arriving. Also, the lack of knowledge about internal injuries/medical training necessitates careful handling by the more experienced hiker (counsellor). However, the advantage that the counsellor may have over other helpers outside this context is that, inhabiting the same terrain (here referring to both the online space and the emotional experience of the client) enables them to assess the situation and react accordingly. With respect to my own journey in this dissertation, I am grateful to those who guided me and for the clients and counsellors who created this data, who have inspired me to actively facilitate a sense of mattering for my clients in many clinical contexts. 131 References Agnew-Davies, R., Stiles, W.B., Hardy, G.E., Barham, M., & Shapiro, D.A. (1998). Alliance structure assessed by the Agnew Relationship Measure (ARM). British Journal of Clinical Psychology, 37(2), 155-172. American Psychiatric Association (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: American Psychiatric Publishing. Anderson, T., Ogles, B.M., & Patterson, 2009. Therapist effects: Facilitative interpersonal skills as a predictor of therapist success. Journal of Clinical Psychology, 65(7), 755-768. Battles, H. (2010). Exploring ethical and methodological issues in internet-based research with adolescents. International Journal of Qualitative Methods, 9(1), 27-37. Berman, A., & Jobes, D.A. (1991). The empirical context. In Adolescent suicide: Assessment and intervention. Washington, DC: American Psychological Association. Brown, S.D., & Lent, R.W. (Eds.) (2008). Handbook of counseling psychology. Hoboken, NJ: John Wiley & Sons. Brown, G.K., Henriques G.R., Sosdjan, D., Beck, A.T., (2004). Suicide intent and accurate expectations of lethality: Predictors of medical lethality of suicide attempts. Journal of Consulting and Clinical Psychology, 72, 1170–1174. Cain, C.L. (2012). Integrating dark humor and compassion: Identities and presentations of self in the front and back regions of hospice. Journal of Contemporary Ethnography 41(6) 668–694. 132 Canadian Mental Health Association: Statistics (n.d.). Retrieved March 14, 2013 from http://www.cmha.ca/bins/content_page.asp?cid=284-285 -1258-1404&lang=1. Cheung, A.H., Dewa, C.S. (2007). Mental health service use among adolescents and young adults with major depressive disorder and suicidality. Canadian Journal of Psychiatry, 52 (4), 228-232. Clifford. A.C., Doran, C. M., & Tsey, K. (2013). A systematic review of suicide prevention interventions targeting indigenous populations in Australia, United States, Canada and New Zealand. BMC Public Health, 13(463), 1471-2458. Cook, J.E. & Doyle, C. (2002). Working alliance in online therapy as compared to face-to-face therapy: Preliminary results. Cyberpsychology and Behavior, 5(2), 95-105. Corbeire, M., & Amundson, N.E. (2007). Perceptions of the ways of mattering by people with mental illness. Career Development Quarterly, 56, 141-149. Cowan, S. & McLeod, J. (2004). Research methods: Discourse analysis. Counselling and Psychotherapy Research, 4(1), 102-103. Creswell, J. (2009). Research design: Qualitative, quantitative and mixed methods. New York: Sage. Curtis, C. (2010). Youth perceptions of suicide and help-seeking: “They’d think I was weak or ‘mental.’” Journal of Youth Studies, 13(6), 699-715. Cutcliffe, J.R., Santos, J., Links, P.S., Zaheer, J., Harder, H.G., Campbell, F., McCormick, R., Harder, K., Bergmans, Y., & Eynan, R. (Eds.). (2014). International Handbook of Clinical Suicide Research. New York, NY: Routledge. 133 Darchuck, A., Wang, V., Weibel, D., Fende, J., Anderson, T., & Horvath, A. (2000). Manual for the working alliance inventory observer form (WAI-O). Department of Psychology, Ohio University. DeCuir-Gunby, J.T., Marshall, P.L., & McCulloch, A.W. (2011). Developing and using a codebook for the analysis of interview data: An example from a professional development research project. Field Methods, 23(2), 136-155. Dixon Rayle, A. (2006). Mattering to others: Implications for the counseling relationship. Journal of Counseling and Development, 84, 483-487. Dowell, N., & Berman, J.S. (2013). Therapist nonverbal behavior and perceptions of empathy, alliance, and treatment credibility. Journal of Psychotherapy Integration, 23(2), 158- 165. Durkheim, E. (1951). Suicide: A study in sociology. Chicago, Illinois: The Free Press. Efstation, J.F., Patton, M.J., & Kardash, C.M. (1990). Measuring the working alliance in counselor supervision. Journal of Counseling Psychology, 37(3), 322-329. Elliott, G.C., Colangelo, M.F., & Gelles, R. (2005). Mattering and suicide ideation: Establishing and elaborating a relationship. Social Psychology Quarterly, 68(3), 223- 238. Feng, B., Li, S., & Li, N. (2016). Is a profile worth a thousand words? How online support- seeker’s profile features may influence the quality of received support messages. Communication Research, 43(2), 253-276. 134 Fletcher-Tomenius, L. J., & Vossler, A. (2009). Trust in online therapeutic relationships: The therapist’s experience. Counselling Psychology Review 24(2), 24-33. Flick, U. (2006). An introduction to qualitative research. Thousand Oaks, CA: Sage. Foucault, M. (1973). The birth of the clinic: An archaeology of medical perception. New York, NY: Vintage Books. Foucault, M. (1965). Madness and civilization: A history of insanity in the age of reason. New York, NY: Pantheon Books. Fullagar, S. (2003). Wasted lives: The social dynamics of shame and youth suicide. Journal of Sociology, 39(3), 291-307. Garb, H. N. (1998). Studying the clinician: Judgment research and psychological assessment. Washington, DC: American Psychological Association. Gee, J.P (2011). How to do discourse analysis: A toolkit. New York, NY: Routledge. Glaser, B., & Stauss, A. (1967). The discovery of grounded theory: Strategies for qualitative research. New York: Aldine Publishing. Granello, D. (2010). A suicide crisis intervention model with 25 practical strategies for implementation. Journal of Mental Health Counseling, 32(3), 218-235. Greidanus, E., & Everall, R.D. (2010). Helper therapy in an online suicide prevention community. British Journal of Guidance and Counselling, 38(2), 191-204. Groth-Marnat, G. (1997). Handbook of psychological assessment. New York: John Wiley & Sons. 135 Government of British Columbia: Representative of children and youth act. Retrieved March 9, 2016 from http://www.bclaws.ca/civix/document/id/complete/statreg/06029_01 Grbich, C. (2013). Qualitative data analysis, Second edition. Thousand Oaks, CA: Sage. Gordon, M., & Melvin, G.A. (2014). Prescribing for depressed adolescents: Office decision- making in the face of limited research evidence. Journal of Pediatrics and Child Health, 50, 498-503. Gortner, E-M., Rude, S. S., & Pennebaker, J.W. (2006). Benefits of expressive writing in lowering rumination and depressive symptoms. Behavior Therapy, 37, 292-303. Gould, M.S., Kalafat, J., Harris-Munfakh, J.L., & Kleinman, M. (2007). An evaluation of crisis hotline outcomes part 2: Suicidal callers. Suicide and Life-Threatening Behaviour, 37(3), 338-352. Jobes, D.A., Eyman, J.R. Yufit., R.I. (1995). How clinicians assess suicide risk in adolescents and adults. Crisis Intervention and Time-Limited Treatment, 2, 1-2. Hackney, H.L. & Cormier, S. (2009). The professional counsellor. Columbus, Ohio: Pearson Education. Hanley, T. (2006). Understanding the online therapeutic alliance through the eyes of adolescent service users. Counselling and Psychotherapy Research, 12(1), 35-43. Halderman, B. L., Eyman, J.R., Kerner, L., & Schlacks, B. (2010). A paradigm for the telephonic assessment of suicidal ideation. Suicide & Life-Threatening Behavior, 39(6), 639-647. 136 Henretty, J. R., Currier, J.M., Berman, J.S., & Levitt, H.M (2014). The impact of counselor self- dislosure on clients: A meta-analytic review of experimental and quasi-experimental research. Journal of Counseling Psychology, 61(2).191-207. Hilsenroth, M.J., Cromer, T., & Ackerman, S. (2012). How to make practical use of therapeutic alliance research in your clinical work. In Levy, R. A., Ablon, J. S., & Kaechele, H. (Eds.). Psychodynamic Psychotherapy Research: Evidence-Based Practice and Practice- Based Evidence. (pp. 361-380.). NY, NY: Springer Press. Horvath, A., & Greenberg, L.S. (1989). Development and validation of the working alliance inventory. Journal of Counseling Psychology, 36(2), 223-333. Horvath, A., & Luborsky, L. (1993). The role of the therapeutic alliance in psychotherapy. Journal of Consulting and Clinical Psychology, 61(4), 561-573. Isacsson, G., & Ahlner, J. (2014). Antidepressants and the risk of suicide in young persons – prescription trends and toxicological analyses. Acta Psychiatrica Scandinavia, 129(4), 296-302. Ishiyama, F. I. (1995). Use of validationgram in counselling: Exploring sources of self- validation and impact of personal transition. Canadian Journal of Counselling, 29(2), 134-146. Joiner, T.E. (2005). Why people die by suicide. Cambridge, MA: Harvard University Press. Kaviani, H., & Hamedi, R. (2011). A quantitative/qualitative study on metaphors used by Persian depressed patients. Archives of Psychiatry & Psychotherapy, 4, 5-13. 137 Knaevelsrud, C., & Maerker, A. (2007). Internet-based treatment for PTSD reduces distress and facilitates the development of a strong therapeutic alliance: a randomized controlled clinical trial. BMC Psychiatry, 7(13), 1-10. Korzybski, A. (1990). Collected Writings. New York: Institute of General Semantics. Kouri, S., & White, J. (2014). Thinking the other side of youth suicide: Engagements with life. Internatonal Journal of Child, Youth, and Family Studues, 5(1), 180-203. Lalvani, P., & Douglas, L.P. (2013). Historical perspectives on studying families of children with disabilities: A case for critical research. Disability Studies Quarterly, 33(3), 1-9. Lambert, M.J. & Barley, D.E. (2001). Research summary on the therapeutic relationship and psychotherapy outcome. Psychotherapy, 38(4), 357-361. Langer, M., Scourfield, J., & Fincham, B. (2008). Documenting the quick and the dead: a study of suicide case files in a coroner’s office. The Sociological Review, 56(2), 294-308. Larsson, P., Loewenthal, D., & Brooks, O. (2012). Counselling psychology and schizophrenia: A critical discursive account. Counselling Psychology Quarterly, 25(1), 31-47. Lapidot-Lefler, N., & Barak, A. (2015). The benign online disinhibition effect: Could situational factors induce self-disclosure and prosocial behaviors? Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 9(2), 1-19. Linehan, M.M., Tutek, D.A., Heard, H.L., & Armstrong, H.E. (1994). Interpersonal outcome of cognitive behavioral treatment with chronically suicidal borderline patients. The American Journal of Psychiatry, 151(12), 1771-1776. 138 MacGeorge, E. L., Feng, B., & Butler, G. L. (2003). Gender differences in the communication values of mature adults. Communication Research Reports, 20, 1-9. Makinen, I. (1997). Are there social correlates to suicide? Social Science and Medicine, 44(12), 1919-29. Martin, J. & Stuart, C. (2011). Working with cyberspace in the life-space. Relational Child and Youth Care Practice, 24 (1), 56-66. Menna, R. & Ruck, M.D. (2005). Adolescent help-seeking behavior: How can we encourage it? Guidance and Counseling, 19(4), 176-183. Milbury, K., Spelman, A., Wood, W., Matin, S., Tannir, N., Jonasch, E., Pisters, L., Wei, Q., & Cohen, L. (2014). Randomized controlled trial of expressive writing for patients with renal cell carcinoma. Journal of Clinical Oncology, 32, 663-670. Mishara, B.L., Chagnon, F., Daigle, M., Balan, C., Campbell, J., & Berman, A. (2007). Which behaviors and intervention styles are related to better short-term outcomes in telephone crisis intervention? Results from a silent monitoring study of calls to the U.S. 1-800- SUICIDE network. Suicide and Life-Threatening Behavior, 37(3), 308-321. Morrow, S. L. (2005). Quality and trustworthiness in qualitative research in counseling psychology. Journal of Counseling Psychology, 52, 250-260. Mudry, T.E. & Strong, T. (2012). Doing recovery online. Qualitative Health Research, 23(3), 313-325. Neimeyer, R. A, & Bonelle, K. (1997). The suicide intervention response inventory: A revision and validation. Death Studies, 21(1), 59-81. 139 Neimeyer, R. A. & Pfeiffer, A. M. (1994). Evaluation of suicide intervention effectiveness. Death Studies, 18(2), 131-166. Nezu, A., Nezu, C.M., & Blissett, S.E. (1988). Sense of humor as a moderator of the relation between stressful events and psychological distress: A prospective analysis. Journal of Personality and Social Psychology, 54(3), 520-525. Nye, C. (1998). Power and authority in clinical practice: A discourse analysis approach to narrative process. Clinical Social Work Journal, 26(3), 271-280. O’Carroll, P., Berman, A.L., Maris, R.W., Mosckiki, E.K., Tanney, B.L., & Silverman, M.M. (1996). Beyond the tower of babel: A nomenclature for suicidology. Suicide and Life- Threatening Behaviour, 26(3), 237-252. Overholser, J., Braden, A., & Dieter, L. (2011). Understanding suicide risk: Identification of high-risk groups during high risk times. Journal of Clinical Psychology, 68(3), 334-348. Parker, I. (1992) Discourse dynamics: Critical analysis for social and individual psychology. London and New York: Routledge. Phelan, S.K., Wright, V., & Gibson, B.E. (2014). Representations of disability and normality in rehabilitation technology promotional materials. Disability and Rehabilitation, 36(24), 2072-2079. Potter, J., & Wetherall, M. (1987). Discourse and social psychology: Beyond attitudes and behaviour. Loughborough: Sage Publications. Priester, M.J. & Clum, G. (1993). Perceived problem-solving ability as a predictor of depression. Journal of Counseling Psychology, 40(1), 79-85. 140 Reeves, A., Bowl, R., Wheeler, S., & Guthrie, E. (2004). The hardest words: Exploring the dialogue of suicide in the counselling process: A discourse analysis. Counselling and Psychotherapy, 4(1), 62-70. Reynolds, D.J., Stiles, W.B., Bailer, J., & Hughes, M.R. (2013). Impact of exchanges and client- therapist alliance in online-text psychotherapy. Cyberpsychology, behavior, and social networking, 16(5), 370-377. Richards, D., & Vigano, N. (2013). Online counseling: A narrative and clinical review of the literature. Journal of Clinical Psychology, 69(9), 994-1011. Rittell, H.W.J., & Weber, M.M. (1973). Dilemmas in a general theory of planning. Policy Sciences, 4, 155-169. Roen, K., Scourfield, J., & McDermott, E. (2008). Making sense of suicide: a discourse analysis of young people’s talk about suicidal subjecthood. Social Science and Medicine, 67(12), 2089-2097. Rogers, C.R. (1957). The necessary and sufficient conditions of therapeutic personality change. Journal of Consulting and Clinical Psychology, 21, 95-103. Rogers, J.R., & Lester, D. (2009). Understanding suicide: Why we don’t and how we might. Cambridge, MA: Hogfree. Rosenberg, M., & McCullough, B.C. (1981). Mattering: inferred significance and mental health among adolescents. Research in Community Mental Health, 2, 163-182. 141 Roy, T., & Gillet, H. (2008). Email: A new technique for forming a therapeutic alliance with high-risk young people failing to engage with mental health services? A case study. Clinical Child Psychology & Psychiatry, 13(1), 95-103. Rudd, M.D., Berman, A.L., Joiner, T.E., Nock, M.K., Silverman, M.M., Mandrusiak, M., Van Orden, K., Witte, T. (2006). Warning signs for suicide: Theory, research, and clinical applications. Suicide and Life-Threatening Behavior, 36(3), 255-262. Ryan, G. & Bernard, H.R. (2003). Techniques to identify themes. Field Methods, 15(1), 85- 109. Sade-Beck, L. (2004). Internet ethnography: Online and offline. International Journal of Qualitative Methods, 3(2). Schlossberg, N.K., Lynch, A.Q. & Chickering, AW. (1989). Improving higher education environments for adults. San Francisco: Jossey-Bass. Shea, S.C. (2008). The practical art of suicide assessment: A guide for mental health professionals and substance abuse counselors. Hoboken, NJ: Wiley. Shneidman, E. (1998). The suicidal mind. Oxford, England: Oxford University Press. Silverman, M.M., Berman, A., Sandal, N.D., O’Carroll, P. W., & Joiner, T.E. (2007). Rebuilding the Tower of Babel: A revised nomenclature for the study of suicide and suicidal behaviours part 1: Background, rationale, and methodology. Suicide and Life- Threatening Behaviour, 37(3), 248-263. Sparks, J.A. & Barry, D. (2013). Outside the black box: Re-assessing pediatric antidepressant prescription. Child and Adolescent Psychiatry, 22(3), 240-246. 142 Spong, S. (2010). Discourse analysis: Rich pickings for counsellors and therapists. Counselling and Psychotherapy Research, 10(1), 67-74. Stake, R. (1995). The art of case study research. Thousand Oaks, CA: Sage. Stiles, W.B., Gordon, L.E., & Lani, J.A. (2002). Session evaluation and the Session Evaluation Questionnaire. In G. S. Tryon (Ed.), Counseling based on process research: Applying what we know (pp. 325-343). Boston, MA: Allyn & Bacon Suler, J. (2005). The online disinhibition effect. International Journal of Applied Psychoanalytic Studies, 2, 184-188. Tannen, D. (1990). You just don’t understand: Women and men in conversation. New York, NY: William Morrow. Taylor, L.D. (2011). Avatars and emotional engagement in synchronous online communication. Cyberpsychology, Behavior, and Social Networking, 14, 207-214. Timm, M. (2011). Crisis counselling online: Building rapport with suicidal youth. Unpublished Master’s thesis. University of British Columbia, Vancouver, BC. Timulak, L. (2007). Identifying core categories of client-identified impact of helpful events in psychotherapy: A qualitative meta-analysis. Psychotherapy Research, 17, 305-314. Tobin, G.A., & Begley, C.M. (2004). Methodological rigour within a qualitative framework. Journal of Advanced Nursing, 48(4), 388-396. Valach, L., Young, R., & Michel, K. (2011). Understanding suicide as an action. In Michel, K., & Jobes, D.A. (Eds.). Building a Therapeutic Alliance with the Suicidal Patient (pp.129-148). Washington, DC: American Psychological Association. 143 Van Orden, K., Witte, T., Cukrowitz, K., Braithwaite, S.R., Selby, E.A., & Joiner, T.E. (2010). The interpersonal theory of suicide. Psychological Review, 117(2), 575-600. Williams, R., Bambling, M., King, R., & Abbott, Q. (2009). In-session processes in online counselling with young people: an exploratory approach. Counselling and Psychotherapy Research: Linking Research with Practice, 9(2), 93-100. Westefeld, J.S., Range, L., Rogers, J.R., & Hill, J.M. (2010). Suicide prevention. In Woolfe, R., Strawbridge, S., Douglas, B., Dryden, W. (Eds.). Handbook of Counselling Psychology. Oxford, UK: Sage. Westwood, M.J., & Ishiyama, F.I. (1990). The communication process as a critical intervention for client change in cross-cultural counselling. Journal of Multicultural Counseling and Development, 18, 163-171. White, J. (2015). Shaking up suicidology. Social Epistemology Review and Reply Collective, 4(6), 1-4. White, J. (2012). Youth suicide as a “wild” problem: Implications for prevention practice. Suicidology Online, 3, 42-50. Wilson, C.J., & Deane, F.P. (2010). Help-negation and suicidal ideation: The role of depression, anxiety and hopelessness. Journal of Youth and Adolescence, 39, 291-305. Winter, D., Bradshaw, Bunn, F., & Wellsted, D. (2013). A systematic review of the literature on counselling and psychotherapy for the prevention of suicide: 1. Quantitative outcome and process studies. Counselling and Psychotherapy Research, 13(3), 164-183. 144 Witte, T.K., Gordon, K.H., Smith, P.N., & Van Orden, P.A. Stoicism and sensation-seeking: Male vulnerabilities for the acquired capacity for suicide. Journal of Research in Personality, 46(4), 384-392. Wong, L.C.J., Wong, P.T.P, & Ishiyama, F.I. (2013). What helps and what hinders in cross- cultural clinical supervision: A critical incident study. The Counseling Psychologist, 4(1), 66-85. Wongpakaran, N., Wongpakaran, T., Wedding, D., & Gewt, K.L. (2013). A comparision of Cohen’s Kappa and Gwet’s AC1 when calculating inter-rater reliability coefficients: A study conducted with personality disorder samples. BMC Medical Research Methodology, 13(61), 1-7. Wood, J.T. (1997). Gendered lives: Communuication, gender, and culture (2nd ed.). Belmont, CA: Wadsworth. Wrzesien, M., Burkhardt, J-M., Botella, C., & Alcaniz, M. (2014). Towards a virtual-reality and augmented reality-mediated therapeutic process model: A theoretical revision of clinical issues and HCI issues. Theoretical Issues in Ergonomics Science, 16(2), 124-153. Woolsey, L.K. (1986). The critical incident technique: An innovative qualitative method of research. Canadian Journal of Counselling, 20(4), 242-254. 145 Appendix A: Volunteer Consent Form Volunteer Consent Form “Are you a robot?” A discourse analysis of online crisis chats Principal Investigator: Dr. Ishu Ishiyama, UBC Department of Educational and Counselling Psychology and Special Education Co-Investigators: Maria Timm, Doctoral Student, PhD Program, UBC Department of Educational and Counselling Psychology and Special Education Purpose: You have been invited to take part in this research because we are interested in examining if certain ways of communicating are more effective than others in helping individuals who use our online services (www.youthinbc.com and www.crisiscentrechat.ca). The results of this study may be used to improve the existing training procedures to better reflect the concerns and counselling needs of those who access them, as well as potentially improve other online programs. In partnership with the Vancouver Crisis Centre and UBC, and on behalf of those in distress, the researchers appreciate your help! Study Procedures: If you decide to volunteer for this study, we will be reviewing transcripts of your chats with chatters using the wwwyouthinbc.com site or the www.crisiscentrechat.ca site. The chats will be stripped of all information that identifies you as the volunteer before it is seem by the researchers. You will remain anonymous and this research will not impact your volunteer duties at the Vancouver Crisis Centre. This research is not evaluative in nature. Potential Risks: Knowing your transcripts are being examined may produce some anxiety. There are no other known risks to participation in this project. Potential Benefits: By contributing to a fuller understanding of what constitutes effective chats, your participation in this study will help the many clients who access our online chat services receive better service and support. You will have the opportunity to request a copy of the completed study and/or the study results. 146 Confidentiality: Your identity will be kept strictly confidential. All data will be identified only by code number and kept in a locked cabinet. Neither you nor the chatters accessing the service will ever be identified by name or your volunteer identity in any reports of the study. Contact for information about the study: If you have any questions or desire further information with respect to this study, you may contact Dr. Ishu Ishiyama at xxx-xxx-xxxx. Contact for concerns about the rights of research subjects: If you have any concerns about your treatment or rights as a research subject, you may contact the Research Subject Information Line in the UBC Office of Research Services at 604-822-8598, or, if long distance, e-mail to RSIL@ors.ubc.ca. Consent: Your consent is entirely voluntary and you may refuse to participate or withdraw from the study at any time without negative consequences. If you choose to withdraw their consent, you can do so at any time by contacting the Online Services Coordinator, who will inform the researchers by referring to your assigned anonymized pseudonyms. Your signature indicates that you have received a copy of this consent form for your own records. Signed: Print name: Date: 147 Appendix B: Satisfaction Scale 1 2 3 4 5 N/A Caller Says Not Helpful Caller Implies Not Helpful Caller Gives No Indication Caller Implies Helpful/Thanks Caller Says Helpful/Thanks Not Applicable \"You're not helpful.\" \"I'll phone someone else. Sorry to bother you.\" \"I'm no further ahead. This isn't helping.\" \"You don't understand.\" Caller becomes angry, annoyed, irritated with your responses. Caller is sarcastic to your responses. Caller hangs up abruptly. Caller makes no comment about accuracy of your empathic statements. Caller sounds same through the call. Caller just responds to you. ‘That’s right” “That’s it.” Caller sounds calmer with responses Caller responds positively to your statements. \"Thank you.\" \"I feel better.\" \"You've been helpful.\" Prank Call Incomplete Call Sex Caller Caller just swears and hangs up 148 Appendix C: Disclaimer Placed on Online Sites (Additions to the website on December 24th 2009 are noted in bold) PRIVACY & LEGAL How do we use and disclose personal information? We use and disclose client personal information only for the purpose for which the information was collected, except as authorized by law. The Crisis Centre does not rent, sell or trade any personal information with third parties. From time to time, the Crisis Centre participates in research programs to help ensure the service we provide to our clients remains relevant. All data provided for research is edited to ensure complete anonymity of our clients. (Obtained from http://crisiscentre.bc.ca/privacy/). 149 Appendix D: Letter of Consent from Vancouver Crisis Centre To: UBC Behavioural Research Ethics Board (BREB) Dr. Ishu Ishiyama (Maria Timm's dissertation research supervisor at UBC) This is to verify that Maria Timm and her research supervisor, Dr. Ishu Ishiyama, have the consent of the Crisis Intervention and Suicide Prevention Centre of BC to conduct research as part of the requirements for the completion of her PhD in Counselling Psychology degree at the University of British Columbia. The title of the proposed research is “Are you a robot?” A discourse analysis of rapport-building in online crisis chats from a suicide prevention centre. I am aware that Maria Timm and Dr. Ishu Ishiyama will be using transcripts of chats obtained from the youthinbc.com and crisiscentrechat.ca online databases. I am also aware that precautions will be taken to ensure that the anonymity of participants (users of the youth and adult crisis chat services) is preserved and that no identifying information will be collected. I understand that Maria Timm and Dr. Ishiyama will not have access to volunteer names and that volunteer pseudonyms will be numerically coded to prevent this. The Crisis Centre will remove uniquely identifying chatter information from the transcript, such as IP address, email addresses, schools’ names, etc. before providing the researchers with the transcripts. Transcripts will be selected by Crisis Centre staff using the inclusion criteria provided by Maria Timm. The Crisis Centre has placed a disclaimer on the youthinbc.com and www.crisiscentrechat.ca websites indicating to users that their (non-identifying) information may be used for research purposes. Only transcripts from chats taken after these disclaimers were placed will be examined. The Crisis Centre considers that the proposed research is very important in developing effective online services to individuals considering suicide. In our view this research involves virtually no risk to clients or potential clients of the Crisis Centre. 150 Appendix E: Sample Wordle Obtained from Counsellor Responses in Rapport-Containing Chats "@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "2017-09"@en ; edm:isShownAt "10.14288/1.0307300"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Counselling Psychology"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "Attribution-NonCommercial-NoDerivatives 4.0 International"@* ; ns0:rightsURI "http://creativecommons.org/licenses/by-nc-nd/4.0/"@* ; ns0:scholarLevel "Graduate"@en ; dcterms:title "\"Are you a robot?\" A discourse analysis of rapport-building in online crisis chats from a suicide prevention centre"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/58652"@en .