Understanding the vicarious trauma and emotional labour of criminological research

Abstract

The psychological cost of exposure to traumatic events is receiving greater recognition in recent years, especially in terms of its impact in helping professions and emergency services. However, the costs to researchers remain relatively unexplored. In this article, we will discuss the nature and impact of vicarious trauma using two criminological research projects as case studies: one a qualitative project engaging with survivors of childhood sexual abuse, and the other, a quantitative analysis of police hate crime reports. In addition to considering the trauma elicited in fieldwork such as interviews, we interrogate the costs imposed on researchers during the coding and analysing processes. We suggest that the cost is potentially greater when the researcher has a personal connection with the issues being researched, but that this personal experience also provides the researcher with important skills for responding to new or compounded trauma. The costs of engagement with trauma may be compensated by the productive outputs and impact on policy and practice that this type of research may elicit. Understanding the impact and costs of engaging with close analyses of trauma is critical in developing more robust and ethical research processes to ensure that this trauma is appropriately managed so as to avert the long-term damage this work can inflict on researchers and participants.

Introduction

Vicarious trauma (VT) is of growing interest to academics and practitioners, particularly as it relates to the work of victim advocates, the helping professions and emergency services (Barrington and Shakespeare-Finch, 2013; Cohen and Collens, 2013; Pack, 2013). Emotional labour is also beginning to be discussed in a growing range of settings (see, for example, Fohring, 2020; Guerzoni, 2020; Jackson et al., 2013). In this article, we discuss the nature and impact of VT using two criminological research projects: one a qualitative project engaging with survivors of childhood sexual abuse, and the other, a quantitative analysis of police hate crime reports. We also discuss the emotional labour required of researchers seeking to enact trauma-informed practice and a witnessing stance in research (Kezelman and Stavropoulos, 2012; Reynolds, 2012). While VT is increasingly addressed in occupational health and safety procedures for workers dealing with trauma (such as social workers, policing and emergency services, and forensic psychologists) (Safe Work Australia, 2019), it remains the white elephant of academic research (Newman et al., 2006). Granted, universities and research centres are more cognisant of researcher safety; however, for the most part, this has not included addressing the VT experienced in undertaking research on topics that are permeated with trauma. This is of particular importance to criminologists, whose work – whether it is with victims, witnesses, offenders or criminal justice practitioners – consistently requires bearing witness to trauma. In addition to considering the trauma elicited in fieldwork such as interviews, we interrogate the costs imposed on researchers during the coding and analysing processes. We suggest that the cost is potentially greater when the researcher has a personal connection with the issues being researched, but that this personal experience also provides the researcher with important skills for meaning-making and developing vicarious resilience. Understanding the impact and costs of engaging with close analyses of trauma is critical in developing more robust and ethical research processes to ensure that this trauma is appropriately managed so as to avert the long-term damage this work can inflict on researchers.

VT, emotional labour and reflexive research

Pearlman and Caringi (2009: 202) define VT as ‘the negative transformation in the helper that results from empathic engagement with trauma survivors and their trauma material, combined with a commitment or responsibility to help them’. Similar to primary trauma, this transformation is often in the form of painful and distressing changes to a person’s core systems of meaning, their relationships, their sense of hope and their measure of safety (Barrington and Shakespeare-Finch, 2013; Hunter and Schofield, 2006; Pack, 2013; Pearlman and Caringi, 2009). VT often manifests as intrusive thoughts and images, nightmares, overwhelming emotional responses and changes in belief systems and behaviour (Barrington and Shakespeare-Finch, 2013; Hunter and Schofield, 2006; Pack, 2013; Pearlman and Caringi, 2009).

VT is considered by some to be an inevitable impact of exposure to trauma material and witnessing the suffering of trauma survivors (Barrington and Shakespeare-Finch, 2013; Pearlman and Caringi, 2009). While there is an overlap between VT, post-traumatic stress disorder, compassion fatigue and burnout, the distinction lies in the nature of the exposure to trauma material (Silver et al., 2015). VT is identified as resulting from secondary exposure to trauma material, most often by hearing or reading trauma survivors’ descriptions of their experiences; an exposure which is undoubtedly present in criminological research. The authors consider VT as distinct from researcher’s experiences of abjection, as described by Thorneycroft (2020), and as extending beyond emotional labour in research discussed by Guerzoni (2020; see also Jackson et al., 2013).

The impact of exposure to trauma material varies depending on the amount, duration and type of exposure, as well as a person’s life history and current circumstances. Pack (2013) found that previous personal experiences of trauma may increase vulnerability to VT; however, the complex interactions between VT, vicarious resilience and personal growth mean that personal experience of trauma is not simply a risk factor, but can also be an asset (see also Barrington and Shakespeare-Finch, 2013). MacRitchie and Leibowitz (2010) suggest that people exposed to trauma material make different meanings about it, influenced by their perception of the trauma survivor’s suffering and their own temperamental and biographical contextualisation. The idea that meaning is made from the unique and subjective position of each interpreter is also a key concept in research, particularly in reflexive qualitative research (Braun and Clarke, 2019). This is an idea that we will return to in the following case studies, reflecting on the particular type of emotional labour necessitated by intersections of VT, lived experience and reflexive research. For this purpose, we draw on aspects of emotional labour as defined by James (1989) and further elucidated by Hochschild (2012), Jewkes (2012) and Stayt (2009). In particular, we focus on the labour of emotion regulation and selective expression during empathic engagement with trauma material in the form of victimisation narratives. In addition, we discuss the emotional labour demands and VT impacts of building genuinely empathic relationships with research participants (Fohring, 2020; Staden, 1998; Stayt, 2009).

A key factor that requires emotional labour and influences the impact of VT is empathy (Barrington and Shakespeare-Finch, 2013; MacRitchie and Leibowitz, 2010; Woodby et al., 2011). Empathy facilitates authentic and effective engagement in therapeutic and research relationships, but can also amplify the effects of exposure to trauma material, as the hearer connects emotionally with the survivor’s story and emotional state. While relationships between research participants and researchers are different from therapeutic relationships, there are aspects of similarity: researchers often need to build relationships of trust and connection with research participants in order to facilitate research processes (Dickson-Swift et al., 2008). Even in research that does not involve direct contact with participants, researchers can develop a deep empathic connection with data and the stories held in data (Jackson et al., 2013). Researchers are often drawn to research topics that are close to us, through direct personal experience, cultural association or other personally significant connections. Such connections can increase the meaningfulness of our work, but like empathy in therapeutic work may also increase the impacts of exposure to trauma material (Barrington and Shakespeare-Finch, 2013; Hunter and Schofield, 2006; MacRitchie and Leibowitz, 2010; Pack, 2013).

Discussion of VT in criminological research is increasing, especially as it relates to frontline criminal justice staff. Spanning the full criminal justice process, research into the experiences of investigators and police officers (Atkinson-Tovar, 2003; Kunst et al., 2017), victim advocates (Slattery and Goodman, 2009), legal practitioners (Leclerc et al., 2020; Silver et al., 2015), juries (Lonergan et al., 2016; McQuiston et al., 2019) and rehabilitation, detention and correctional staff (Bosworth and Kellezi, 2017; Severson and Pettus -Davis, 2013) clearly identifies the impact VT can have on work in these fields. As noted above, VT is enhanced when there is . . . a commitment or responsibility to help’ primary victims and survivors (Pearlman and Caringi, 2009: 202). Similarly, Stayt (2009) identifies that caring for people in times of distress requires considerable emotional labour and often causes emotional stress in care providers. Criminal justice practitioners are often drawn to this type of work primarily because of their desire to ‘help’ and to do something for their communities. And while many of these criminal justice professions – in particular, policing – have recruitment strategies that sift out workers with a predisposition to negative responses to trauma, ¹ the desire to help, combined with sustained, repeated exposure to others’ trauma, along with inadequate institutional responses to that trauma all but guarantees that these criminal justice actors will experience VT at some point in their careers. The desire to help, and being successful in providing that help, may reduce the impact of VT on frontline staff. In this respect, the desire to help may be both a protective and harmful factor in VT, depending on the success achieved in providing that help. Just as criminal justice practitioners seek to help those who experience primary trauma, the desire to help, to resolve and to find solutions informs why some scholars choose to engage in traumatic research.

Aligned to this increasing interest in the impact of VT on criminal justice practitioners, is the identification of VT in criminological research (Fohring, 2020; Guerzoni, 2020). Dunn (1991) suggests that qualitative research must report not only on the researcher’s observations but also their experiences of undertaking that research. Dunn argues that documenting researcher’s experiences, including their reactions and responses to their participants and the information they collect, is critical as these factors may influence the analysis and interpretation of the data. During the interviews, in reviewing the recorded interviews, in coding and in data analysis, Dunn (1991) experienced VT symptoms in parallel to her participants. Her paper points to the critical need for researcher training and protective strategies for early career researchers, as well as those who have consistently engaged with traumatic research over their careers. The initial engagement with VT affects researchers’ wellbeing, but if addressed appropriately, may be resolved with conventional methods such as debriefing. However, if there is sustained, repeated exposure that embeds the symptoms of VT in everyday life, this is less likely to be addressed or resolved by traditional methods such as peer-to-peer debriefing (Dunn, 1991).

The formalised nature of engagement with people’s stories through research creates a dynamic of witnessing, where stories are heard with a degree of status and significance, and reflected on in the context of broader social issues (Reynolds, 2012). The dialogic transaction that occurs between researchers and research participants is unique: it is distinct from therapeutic relationships where dialogue is often framed as occurring between an expert helper and a person who needs help; and distinct from testimony in the criminal justice system, where victims are often treated with scepticism. Dialogue within research involves both surface and deep acting (Hochschild, 2012) and has the capacity to transform both/all parties involved, via the processes of witnessing and recognition (Benjamin, 2018; Oliver, 2001; Reynolds, 2012). While not discussed in the existing research, as documented in the second of our two case studies, a similar, disconnected witnessing occurs in secondary, quantitative data analysis.

Testimony in the context of research participation can afford the participant status as an expert contributor to a process that is culturally and politically located, and is intended to lead to a meaningful outcome. In these moments, the researcher, and the research processes and outputs, function as ‘the third’, referred to by Benjamin (2018) as ‘a vantage point outside the two’ (p. 23). Benjamin (2018) argues that there is a transformative effect occurring in processes of witnessing and recognition, which utilise a third, shifting meaning-making beyond the conceptual limitations of ‘doer and done to’ (p. 24).

For researchers dedicated to trauma-informed practice, and the facilitation of meaningful participant experiences, the potential for healing that lies in the act of witnessing becomes a compelling ethical (and moral) positioning. Trauma-informed practice guidelines (Kezelman and Stavropoulos, 2012) emphasise the importance of relational conditions that avoid re-traumatisation and that foster opportunities for healing. Key priorities in trauma-informed practice are safety, trustworthiness and the clear communication of belief in victim disclosures. Belief cannot be performed (surface acting); it must be authentic and embodied (deep acting) (Hochschild, 2012). Similarly, as Reynolds (2012) describes, justice-doing through witnessing demands that the witness be deeply present, leaning into the victim’s experiences and their personal and political meanings. Emotional labour in this context is complex and demanding; it is not simply a matter of performing emotion and connection in the desired way (Hochschild, 2012). The witness must journey empathically (and therefore feel deeply) with the victim while paying careful attention to the authenticity and appropriateness of the emotional responses they allow themselves to show (James, 1989; Jewkes, 2012; Staden, 1998; Stayt, 2009). This kind of presence, engagement with suffering and the emotional labour of witnessing cannot help but transform the witness (Sampson et al., 2008); however, whether the legacy of this work is VT or vicarious resilience depends on a variety of factors.

The emotional impact on researchers from conducting research on distressing topics is well recognised in the methodology literature relating to qualitative research and primary data (Sampson et al., 2008, 2010; Dickson-Swift et al., 2008; Jackson et al., 2013; Woodby et al., 2011). As Sampson et al. (2008) identified, the risk of emotional harm is amplified when researchers value reflexivity and research relationships, and care deeply about participants’ interests, as is common in feminist, embodied methodologies. However, there is little recognition, in scholarly work or policy, of the potential for emotional distress and VT in researchers working with secondary data or quantitative methods (see Guerzoni, 2020). Jackson et al. (2013) describe the ways that researchers immersed in secondary data containing children’s accounts of abuse were haunted by their imaginings, hearing the children’s voices in their minds and picturing the abuse they were reading about. This challenges the assumption that secondary analysis does not create the interpersonal connections and emotional responses (Sampson et al., 2008) that MacRitchie and Leibowitz (2010) have identified as a causal factor in VT and secondary traumatic stress (STS). Compulsive rumination and imagining is both an indicator and a cause of VT (Pack, 2004). However, imagining is a natural step for an empathically engaged researcher, and VT training to discourage mentally replaying trauma material is rarely provided for researchers.

In practice settings, a broad range of services are now investing in VT training, although many still do not have established protocols for VT prevention or treatment beyond generic employee assistance pathways and workplace mental health prevention programmes (see, for example, the Queensland Department of Child Safety, Youth and Women (QDCSYW), 2017, guidelines on VT). Recognition of VT as a risk factor in academic work is also variable. While many universities and their ethics committees address researcher wellbeing and safety in research training and ethics approval processes, again, this is often generic, without specific examination of VT (Bloor et al., 2010; Jackson et al., 2013).

Suggested strategies for managing VT include training and preparation, formal and informal debriefing, accessing supervision, reducing the amount of exposure to trauma material and prioritising self-care (Barrington and Shakespeare-Finch, 2013; Pack, 2004, 2013; Pearlman and Caringi, 2009; QDCSYW, 2017; Severson and Pettus-Davis, 2013; Trippany et al., 2004). Interestingly, Pack (2004, 2013; see also Hunter and Schofield, 2006) found that social workers supporting sexual abuse victims were able to mitigate some VT impacts by adjusting their philosophical approaches to their work and their frameworks for understanding sexual violence, and by engaging in advocacy and activism work. The reparative effect of advocacy work makes sense if we consider that meaning-making and a sense of agency are central in both primary trauma and VT: ‘helping’ or doing something assists in restoring a hopeful and optimistic outlook, and a person’s sense of being able to effect change in the world (Brown, 2008; Herman, 2015; Hunter and Schofield, 2006; Pack, 2013).

Dickson-Swift et al. (2008) note that researchers frequently use informal debriefing and peer support as a method for managing emotional distress, and raise concerns that while valuable, informal support from peers untrained in counselling may not always be sufficient or helpful. They suggest a number of strategies for managing the emotional risks of sensitive research, including specialist training, professional supervision and access to skilled counselling support. Furthermore, they argue that emotional risk is as important a consideration as physical risk, and that it is the responsibility of institutions, as employers, to adequately attend to this risk. In the context of sexual abuse counselling, Pack (2013) has argued that broader social discourses and workplace cultures that stigmatise and isolate victims and those supporting them increase the effects of VT. In contrast, creating cultures in which VT is openly discussed aids in the development of vicarious resilience. There is a small but persistent call for comparative efforts towards culture change in research institutions (Dickson-Swift et al., 2008; Jackson et al., 2013; Woodby et al., 2011).

Vicarious resilience and post-traumatic growth are relatively new terms used to describe beneficial impacts of working with trauma, such as increased gratitude, growth in self-awareness and development of new skills useful in professional and personal contexts (Barrington and Shakespeare-Finch, 2013; Kunst et al., 2017; Pack, 2013). Such growth often occurs in the same domains that bear the damaging emotional, professional, existential, social/relational and spiritual impacts of VT (Barrington and Shakespeare-Finch, 2013; Pack, 2013). It is also important to acknowledge that, as researchers, many of us feel extraordinarily privileged to do the work that we do and recognise the sacredness of bearing witness to people’s suffering, survival, courage, healing and growth (Jackson et al., 2013).

Reflecting on VT

This reflexive article stems from ongoing discussions between the authors over the last 10 years. When we embarked on our first collaborative project in 2009, we felt alone in the task of coding and analysing traumatic material, and in the absence of formal institutional support, we relied on informal peer support from each other (Dickson-Swift et al., 2008). The choice of these two studies is therefore exploratory, purposive and intrinsic (Tellis, 1997). Our goals are simply to reflect on our experiences of engaging with trauma in research and identify the critical tipping points when VT and emotional labour arose in our research processes. Through this discussion, we hope to provide other scholars with signposts to the symptoms of VT and possible ways to address the inevitable VT and emotional labour involved in this type of scholarship.

The research studies presented below share only the VT produced in undertaking this research. They are theoretically, methodologically and conceptually different. One is qualitative, the other quantitative. One is framed by psycho-social trauma theory, the other, the core principles of forensic linguistics. One was undertaken with primary data, the other secondary. Yet, despite these differences, they share the common characteristic of generating VT in the researchers. Importantly, common to both is the involvement of both authors in each project. In the first, Rebecca was the primary researcher and Nicole provided peer-to-peer support and informal supervision, and in the second, Nicole was primary researcher and Rebecca was her Research Assistant. In each study, we outline the projects and what was involved in primary and secondary fieldwork and data analysis, with a particular focus on the VT exposures, impacts, how we managed them, and the gaps in managing VT exposures and impacts.

Qualitative child sexual abuse research

Rebecca’s PhD research explores the experiences of adult survivors of child sexual abuse (CSA) who made submissions to the Australian Royal Commission into Institutional Responses to Child Sexual Abuse (2019). Data collection consisted of in-depth semi-structured interviews with 26 survivors, with interviews then transcribed and analysed using reflexive thematic analysis (Braun and Clarke, 2019). While interviews did not include questions about participants’ experiences of abuse, many participants nonetheless shared information about abuse experiences, and all participants spoke at length about the devastating impacts of trauma and abuse. Rebecca has her own history of trauma and abuse, as do members of her family. A number of risk factors for VT are therefore present: prolonged and repeated exposure to trauma material and survivor distress, with potential exacerbations due to previous experiences of trauma and close personal connection to the research topic (Barrington and Shakespeare-Finch, 2013; MacRitchie and Leibowitz, 2010; Pack, 2013; Pearlman and Caringi, 2009).

Rebecca also has experience working with trauma survivors in the community, teaches in social work and works as a trainer teaching on trauma-informed practice and VT in government and community services. This experience is relevant as both an additional source of exposure to trauma material and an indicator of skill and experience in managing VT risk. Rebecca’s research training did not cover VT as a risk, and it was not a requirement for securing ethics approval. However, remembering the VT impacts of her master’s research and her work with Nicole on the hate crimes project (see Study 2 below), Rebecca did make plans for attending to the risk of VT, primarily through selection of research questions and a project design which (theoretically) would limit exposure to traumatic content. As described above, this plan did not entirely work, as participants shared abuse experiences without any prompting. Noting the early signs of VT, Rebecca outsourced some transcribing; again, in an effort to reduce the quantity and duration of exposure to trauma material, Rebecca also found ways to debrief informally with supervisors and peers.

The VT impacts Rebecca encountered were, as MacRitchie and Leibowitz (2010) and Barrington and Shakespeare-Finch (2013) argued, often related to empathy, personal experience and perceptions of participant’s suffering. The most compelling example is the difficulty Rebecca had conducting, transcribing and coding an interview with a woman (we call her ‘Jasmine’) who described herself as having Dissociative Identity Disorder, and who ‘switched’ between parts many times during the video interview. Jasmine sometimes appeared masculine and aggressive, and at other times childlike and vulnerable, requiring Rebecca to work hard emotionally in order to remain connected with Jasmine through these changing states. The abuse Jasmine has endured, and described in detail, was extreme, prolonged and horrifying. Witnessing Jasmine’s distress and engaging empathically with her experiences demand emotion regulation for self-preservation (Stayt, 2009), while the need to hold safe space for her required thoughtful communication of Rebecca’s own emotional responses (James, 1989; Jewkes, 2012; Stayt, 2009). While other participants described comparable abuse experiences, Rebecca’s sense of relating to Jasmine, particularly with regard to the lasting emotional and existential impacts of her abuse experiences, made this interview data particularly traumatic to engage with and tested her emotion regulation skills.

The nuances of empathy development are slippery and somewhat intangible; it is hard to say exactly what creates a stronger sense of connection with one participant over another. Nussbaum (2013) argues that one factor in the development of empathy is the conceivable idea that ‘this could have been me’. This idea certainly resonates for Rebecca when reflecting on her experiences with Jasmine, as does Lifton’s (2012) description of ‘survivor guilt’. However, Jasmine’s suffering is so clear and immediate in her presentation that some VT impacts would be likely for anyone witnessing her distress and hearing her story, regardless of personal connections or experiences. Jasmine is intelligent, humorous and very likable; perhaps this affection also increases the emotional impact of recognising her pain.

The way Jasmine described her abuse experiences and their impacts elicited an immediate and lasting emotional response in Rebecca (Woodby et al., 2011). Jasmine often seemed to be pleading with Rebecca to help her make sense of these experiences – to help her understand and manage the existential trauma (Brown, 2008; Herman, 2015; Lifton, 2012) that comes from unrelenting exposure to the banality of evil (Arendt, 1963). Rebecca carried Jasmine’s bewilderment and grief away from the interview and added it to her own reservoir of existential trauma.

The scene from Jasmine’s story which still haunts Rebecca the most is two perpetrators of organised sexual abuse laughing, while two children screamed in fear for their lives: perpetrators inducing terror as a joke, for their own entertainment. This is an imagining made more painful for Rebecca by her role as a mother to a young child, as well as her own memories of childhood abuse and dehumanisation. Jasmine’s reliving of this experience contained complex emotional currents: terror still raw; guilt over her own actions to survive; grief; and a childlike confusion that hints at an unspoken question: ‘why did they do this to us?’ As an emotionally engaged interviewer, Rebecca feels all of these with Jasmine, as well as noting her own connections to such feelings.

In addition, Jasmine’s vulnerability and distress enlists Rebecca as a helper of sorts; albeit, in the brief and somewhat distanced context of a researcher–participant relationship, and a video interview. As in Pearlman and Caringi’s (2009: 202) definition, all the elements for VT are present: empathic engagement with trauma material and a sense of responsibility to somehow provide help. ‘Help’ in this instance can only be in the form of witnessing, understanding and validating Jasmine’s experience and suffering, all of which require emotional labour to regulate and express appropriately (Jewkes, 2012; Reynolds, 2012). As Woodby et al. (2011) observe, in transcription and coding there is even less opportunity to help – a powerlessness more difficult to sit with when victim participants are still not beyond the reach of perpetrators.

VT manifested for Rebecca in a number of ways throughout her engagement with Jasmine’s interview data. During the live interview, while Rebecca remained primarily focussed on Jasmine (her story and her emotional and psychological wellbeing), Rebecca’s distress level was fairly low. However, once the interview and follow-up support were complete, and Rebecca was left alone with the data, significant distress began to emerge. Facing the data for transcription and coding felt unbearable without the distraction of focussing on Jasmine’s wellbeing. For many weeks after the interview, and for days after any attempt to complete transcription or coding, Rebecca experienced intrusive thoughts, strong emotional responses (grief, shame, a sense of hopelessness and despair), insomnia and an urge to withdraw and isolate herself.

Rebecca was reluctant to seek out support or to debrief with anyone, due to the desire for isolation and avoidance of other people, but also out of fear that talking to anyone about Jasmine’s experiences and the impact they were having on her would simply be spreading VT further; a kind of contagion. Instead, Rebecca utilised skills developed through surviving her own trauma experiences to navigate this distressing and isolating period and by managing her response to Jasmine’s story as she would any other resurgence of trauma impacts. This included allowing and reflecting on difficult feelings, being aware of temptation to revert to harmful coping strategies and investing in activities to increase safety and wellbeing, such as exercise, meditation, journaling and time spent away from trauma material. Fortunately, Rebecca’s primary supervisor had, coincidentally, previously interviewed Jasmine and was familiar with her presentation and her abuse history, so Rebecca was eventually able to talk about the impact and work out a plan for transcribing and analysing this distressing interview. Rebecca still has not talked about the content of Jasmine’s interview, but has benefitted from the informal support of discussing the VT impacts with her supervisor and with Nicole.

It took many months to complete transcription, working through small sections at a time, and some content was left un-transcribed and not included in analysis. It would not have been appropriate to outsource transcription in this instance: Jasmine had asked that some parts of her interview be kept private, and the likelihood of a significant VT impact on transcription service staff (who often do not have access to VT support) posed an unacceptable risk. The un-transcribed sections, while providing valuable contextual information, were not crucial to addressing the research questions, and Rebecca and her supervisor decided that it was best to summarise rather than transcribe these particularly graphic parts of Jasmine’s story.

Rebecca experienced VT impacts from other aspects of her research too, but none so severe as those related to Jasmine’s interview. Over many months, the impacts eased, although, as Lifton (2012) reflects, the imprint of immersion in horror never really leaves us. However, when we can incorporate comprehension of horror into our worldview and still find ways to feel hope and happiness, the impacts of VT are no longer debilitating – particularly if we can find ways to make value and meaning from the experience (Brown, 2008; Herman, 2015; Lifton, 1975). Rebecca was able to build on the meaning-making work already done in her own trauma healing and in previous episodes of VT to assimilate this new horror fairly quickly; turning it from paralysing despair into compassion and determination.

Quantitative hate crime research

After working at the Lesbian and Gay Anti-Violence Project (AVP) in the 1990s as a Client Advocate for victims of hate crime, Nicole made the decision when she returned to postgraduate studies that she was no longer going to undertake primary research with victims, nor investigate hate crime experiences. Unknown to her at the time, these were adaptation and coping strategies (Hunter and Schofield, 2006). In their place, she pursued secondary research into the role of hate speech in hate crimes. This decision was not taken lightly. While there continues to be a dearth of research on the impact of hate crime on individuals and communities, the individual costs of undertaking this type of research are significant for both the researcher and the participant, especially when both share community and experiences of violence (Barrington and Shakespeare-Finch, 2013; Hunter and Schofield, 2006; MacRitchie and Leibowitz, 2010; Pack, 2013).

At the time of embarking on postgraduate studies, Nicole was still managing the VT acquired during her term as a Client Advocate, during which limited psychological support was provided to workers. As part of her work at the AVP, she was required to complete a Report of Violence for each client and thus knew that secondary hate crime data existed, albeit held by a small number of community organisations and some policing organisations. It was from this insider knowledge that she developed her first major study into the role of verbal and textual hostility (aka, hate speech) in reported hate crime. Acknowledging the limits imposed in utilising existing data collected by others, for other purposes, she sought to create a typology of hate speech reported by victims to community and policing organisations. This was achieved through two separate, but related projects using Critical Discourse Analysis (CDA; Wodak, 2001). CDA is not a method, per se, but rather a methodology that deploys qualitative, quantitative and mixed-methods approaches to examine the individual, social and institutional contexts of text/speech.

Her doctoral research analysed 1227 complaints of heterosexist and antisemitic hate violence reported to the AVP, Executive Council of Australian Jewry (ECAJ) and the New South Wales Anti-Discrimination Board (ADB) between January 1995 and December 1999. The AVP was unique among these organisations in that their Report of Violence included a specific field to narrate the hate speech used during the incident, which enabled Nicole to avoid re-reading the whole case file. However, in these cases, as well as those reported to the ECAJ and the ADB, it was only in reading the full transcript of the case that the contextual factors (type of violence, place of incident, presence of witnesses/other victims/bystanders) could be identified.

The analysis of these data provided sufficient information to create the first typology of verbal and textual hostility in hate crime (Asquith, 2004, 2007). She then tested this typology using 99,727 cases reported to the London Metropolitan Police Service (MPS) between January 2003 and December 2007 (Asquith, 2013). Of these cases, 2 years of data (2003 and 2007), consisting of 27,164 cases, were analysed for the presence and type of verbal and textual hostility. While 20,756 (76%) of these cases included reports of racist, heterosexist and faith-based hate speech, only 5584 cases (21%) reported that speech verbatim.

This verbatim hate speech was not recorded in an individual field, and could only be identified by reading the abridged (250-character) narrative of the incident recorded by the police officer. These abridged narratives included a summary of the incidents, and in 21% of cases, the recording of the hate speech used in the incident. It was only in reading every one of the 27,164 abridged narratives that the hate speech and contextual factors could be identified. Knowing from experience that even secondary data analysis could elicit VT, Nicole sought funding to employ Research Assistants to undertake this initial coding of the data. Rebecca and another casual Research Assistant undertook the majority of the data coding for this second study into the role of hate speech in hate crime. The decision to delegate this difficult work to casual staff was instrumentalist and self-serving. Dispersing the workload of coding was a preventive measure as it shared the VT between researchers and limited the exposure of any one person (Hunter and Schofield, 2006; Woodby et al., 2011).

While the goal in utilising secondary data was to avoid the VT elicited in primary fieldwork (such as interviews), this was not achieved in either study. Apart from the generalised VT produced from reading the secondary narratives of victims’ experiences collated by these organisations, in her doctoral research, Nicole was re-reading all of the cases she was responsible for as the Client Advocate at the AVP. This analysis created a new layer of VT, as well as reminders of the initial VT produced at the time of reporting.

Reading, reviewing and coding incidents of hate violence is traumatic whether it is a single incident or, as in the MPS research, 27,164 cases. For hate crime victims, some experiences are so impactful that they inform life decisions from that point on. From her experience as an advocate, Nicole also knows that what creates the conditions for ongoing trauma and behaviour modification is not necessarily the level of violence or physical harms caused. For some victims, it can be simply the wrong thing said at a particular moment that undermines cherished ideals about human nature. Likewise, for coders, words wound (Matsuda, 1993). We procrastinate, and when forced to return to the task, we flinch away from the computer screen. We scan the narrative for any explicit sign (such as quotation marks) that verbatim text is recorded so as to avoid having to digest information that may not be useful to the study, such as detailed accounts of horrendous physical and sexual abuse that do not include verbatim hate speech. And yet still generalised anxiety and VT returns over and over again.

Although a single narrative may cause harm, it is in the cumulative effect of 27,164 cases that we can begin to see to dangers of VT in research. Months of reviewing and coding violent data begins to seep into our worlds and colours even our most precious relationships. We need to debrief but do not want to share the trauma and its effects. So we debrief with each other, and in the process, embed the trauma in our shared knowledge. We talk of this, or that incident as if it is qualitatively different from all others, when in most cases, they are simply the outliers that we cannot forget. In the midst of her work on the MPS data set, Rebecca called Nicole just to talk – not debrief, as neither of us were cognisant of the necessity to do so, nor did either of us have the skills to do so at the time. Ten years later, we still remember the case that had hit both of us hardest that day. It was not extraordinary; there were only a few minor injuries. It was not a case that either of us, as White women, could imagine experiencing. Nor did it have any salience or link to the interpersonal trauma that both of us have experienced as children and adults. It was a case of a young Black man, who had been detained in a cage while his perpetrators called him a monkey and threw bananas at him. As with Woodby et al. (2011: 831), there was a deep level of ‘affective permeability’ that came from sharing this trauma narrative, including the emotional labour involved in managing each other’s traumatic responses to that data, and Nicole’s responsibility as supervisor of Rebecca’s work. There was significant emotional labour deployed in the caring and empathy work required to address each other’s VT, as well as the labour required to ensure that as an employee Rebecca was supported in their work. It was the banality of the evil we parsed (Arendt, 1963), we believe, that shaped – and continues to shape – our responses to the hatred we have coded.

As noted in the existing literature on VT in research practice, much attention is focussed on the rich, thick accounts of trauma collected in qualitative research. The two projects discussed in this case study were primarily quantitative; numbers crunched rather than emotions reflexively coded and analysed. Yet, VT remains a lasting artefact of this work for both Nicole and Rebecca. The trauma may be reduced to numbers, but in order to crunch those numbers, the text must be coded, and in the process, the disembodied text of police reports is reimagined, gaps are filled and these stories are embedded in the everyday in ways not dissimilar to that which occurs in coding primary qualitative data such as that collected by Rebecca.

Witnessing, reflexivity and emotional labour in research

VT is an ‘absent presence’ (Derrida, 1997) in criminological research. It permeates the work of criminal justice practitioners and the victims, witnesses and offenders they assist; yet, it remains a hidden trauma largely unaddressed in academic research design, training and ethics protocols. While the Australian system for ethics approval has developed significantly, and now requires researchers to account for the harms caused to both participants and researchers, there is still a belief that secondary data analysis poses fewer risks than primary fieldwork (Jackson et al., 2013). Furthermore, too often, secondary data analyses are still considered under the low- or no-risk protocol, especially if the data collected have already been de-identified. This leaves too much of the trauma of criminological research unattended and abrogates the responsibility for addressing VT to the individual researcher. As our research studies demonstrate, all that was available to us at the time was to design our research in ways that sought to reduce VT (unsuccessful in both cases) and to call on informal supports when we found ourselves affected by VT.

A critical factor in data coding is the nature of the ‘text’ to be analysed. In the studies we discuss in this article, the coded data come in two forms: thick, rich life-history interviews with survivors of violence and detached, abbreviated narratives of reported hate crimes. While an assumption can easily be made that the former may elicit increased VT due to its extended and emotional engagement with traumatised participants (Jackson et al., 2013), we suggest that VT is also created by the dehumanising language of forensic investigation. Reduced to policing terms, and with the trauma stripped of its context, police reports and coronial enquires are VT minefields. The reduction of life-changing events to the bare basic information required to complete the 250-character field of police incident reports affords no space for the trauma, and when read by others out of context – such as researchers in data coding – produces a sanitised account of the damage incurred in seeking justice. The disembodied text of police narratives, such as IC4 and VIW1, ² strips these accounts of the full context and leaves the coder with a blank slate on which to map the imagined back-story to each of these narratives.

Our research philosophy is foregrounded by a commitment to trauma-informed practice, which means a commitment to doing research with survivors in a way that could be meaningful and rewarding for participants. The key to trauma-informed practice is relationships that prioritise safe and trustworthy connections, shared power, belief and recognition of the impacts of trauma (Kezelman and Stavropoulos, 2012). Obviously, this is easier to put in practice when conducting qualitative research; however, in a limited way, it can also frame how we engage with, code and analyse quantitative data, especially as it relates to the outcomes and implications of this research. As Reynolds (2012) describes, witnessing is, in itself, a form of justice-doing: witnessing someone’s story of trauma, while carrying an awareness of the historical and political context of their suffering, can be transformative for both parties, and may contribute to the development of post-traumatic growth (Cohen and Collens, 2013; Kunst et al., 2017). A witnessing stance extends into data analysis, as does trauma-informed practice, as it requires the researcher to hold an emotional connection to each participant throughout coding, analysis and discussion. This does not mean shying away from rigorous analysis, but it does foster a reluctance to break up data unnecessarily or obscure participants’ voices.

Braun and Clarke (2019; see also Jackson et al., 2013) argue for the importance of researcher reflexivity in qualitative research. Understanding the influence of our own experiences, values, meaning systems, aversions and vulnerabilities is essential in rigorous analysis – and this includes an awareness of what drives us to study a particular topic, and how our own experiences might blind us, or sharpen our focus. When VT is added to the mix, reflexivity becomes more difficult and urgent. A defining feature of trauma responses is a desire to avoid unbearable stimuli (Herman, 2015); and a key feature of reflexivity is to look directly at what we are driven to avoid. This highlights the importance of attending to VT in research, as unaddressed VT can create a compelling reluctance to examine ourselves, or our data. Such blind spots undermine the researcher’s capacity for insightful analysis (Braun and Clarke, 2019; Jackson et al., 2013). As Woodby et al. (2011) describe, there is an expectation that researchers are skilled in emotional reflexivity and able to engage in the emotional labour of reflexive coding, yet practical training to develop these skills is rarely provided to researchers.

Having discussed the difficulties, risks and costs of doing criminological research that is at times traumatic, the reasons we keep doing it and the value it holds for us need to be considered. Lifton (1975) argues that ‘. . . the overall struggle of the survivor is to give form, significance, and meaning to the death immersion, in order to move forward in one’s postholocaust existence’ (p. 182). Without claiming that ‘researcher’ and ‘survivor’ are interchangeable – although they are, at times, overlapping identities – we suspect that something similar is true for researchers who chose to confront psychologically challenging issues through their work. Once we know of an atrocity, it is difficult to suppress and ignore that knowledge. There is a point of no return, where you cannot un-know or un-see injustice; which leaves only the option of trying to create meaningful, valuable and significant insights from the unbearable. While a circular argument, we suggest that the VT that comes from research (and the trauma of lived experiences) becomes the source of the need for meaning-making, which in turn becomes the sustaining force in research; and, as we know, this is likely to cause more VT. As such, there is no escaping the probability of VT for those who engage with traumatic research. Instead, we need to consider strategies that both prevent and respond to VT in research, extending beyond the informal supports that are often the only available option.

Pack (2013) identifies a related paradox, in that it is precisely the processes of navigating VT that leads to vicarious resilience. Valuable meaning is created from trauma learning, which enriches our lives and builds personal and professional competencies. There is also meaning to be found in the bonds that develop through shared knowledge, and the relational skills researchers can develop through supporting each other. Increased knowledge of self is rewarding; as is the recognition we can find in knowing that we have shared a profound, if painful, experience. As Newman et al. (2006) suggest, and as our experience has shown, there is some protection against VT to be found in peer support, and the development of working cultures that acknowledge the significance of emotional responses in qualitative and quantitative research (see also Dickson-Swift et al., 2008; Pack, 2013; Woodby et al., 2011). In this way, we can extend the therapeutic value of witnessing and recognition beyond researcher–participant relationships to include our fellow researchers. However, as Dickson-Swift et al. (2008) identify, research institutions also have a role to play in responding to VT, through the provision of appropriate training and formalised support, and the facilitation of workplace cultures that openly recognise and discuss the emotional impacts of doing research.

Conclusion

Comprehending the senselessness of profound cruelty is existentially traumatic. Awareness of the pervasiveness and prevalence of abuse, combined with looking into the abyss and seeing the unbearable ordinariness of those who perpetrate atrocities, exposes the world as dangerous and cruel without reason. Walking through shopping centres, driving to work or playing with your own child takes on a different quality when you are cognitively and emotionally preoccupied with everyday atrocities and their devastating impacts. And yet, we remain drawn to this work, and grateful for the learning – academic, personal and existential – we find in this space. We value the opportunities for meaning-making and growth found in engaging with trauma in this way, but remain concerned that VT is not currently well recognised as a health risk in research – particularly research using quantitative methods and secondary data – and that this lack of recognition creates isolation and stigma.

We suggest that reflexivity in coding and analysing data is critical to both qualitative and quantitative research, and that without this deep engagement with our positionalities in relation to our participants’ stories, the trauma imposed through this work is difficult to resolve. We also suggest that there is a bi-directional relationship between emotional labour and VT, in that the emotional labour required to undertake this work may exacerbate the VT experienced. However, not all emotional labour necessarily results in VT. VT arises only when there is a deep and extended engagement with trauma material, and that this VT requires emotional labour in order to manage the resulting effects (such as anxiety etc.). The case studies we have discussed here are not unique; similar methodologies and analytical frameworks are used throughout criminological research. As such, we hope our discussion of emotional labour and VT in collecting, coding and analysing traumatic material provides the necessary insights required to address VT in criminological research and assist in the development of a scholarly culture where the emotional challenges of research are openly discussed. Most of all, we hope that our transparent reflection on our experiences is of value to other researchers impacted by VT.