Socioeconomic inequalities of suicide: Sociological and psychological intersections

The curious position of sociology in suicide research

Emile Durkheim’s Le Suicide (1952) has a particular status within sociology for demonstrating the relevance of sociological methods and theory in understanding a phenomenon that had been viewed as individual and psychological (Fincham et al., 2011; Fitzpatrick et al., 2015; Marsh, 2010). However, as Bryant and Garnham (2015) note, it is perhaps surprising, given this early and prominent sociological intervention, that contemporary approaches to understanding suicide tend to be overwhelmingly individualistic. Sociological research addressing suicide in the late twentieth and early twenty-first centuries remains a curiously niche endeavour (Abrutyn and Mueller, 2014: 347; Recker and Moore, 2016; Scourfield et al., 2012). A review of publications in mainstream sociology journals in the US found that only 3% addressed suicide (Wray et al., 2011). Conversely, within suicidology journals – an ostensibly multidisciplinary field – psychiatric and psychological research dominates, quantitative approaches are privileged, and sociological work is in a clear minority (Abrutyn and Mueller, 2014; Hjelmeland, 2015; Hjelmeland and Knizek, 2016; Marsh, 2010).

When sociology does address suicide, there is a great tendency to draw on Durkheim, despite a number of serious critiques being levelled at his theories in the twentieth century (Douglas, 1967; Stack, 1982). This close affinity with Durkheimian sociology may explain in part the propensity towards the development of functionalist models, and reliance on quantitative methods (Gove, 1972; Stack and Wasserman, 2009). Recent work by Abrutyn and Mueller (2014; 2016), and Braswell and Kushner (2012), is notable in attempting to build creatively on Durkheim’s work; with Abrutyn and Mueller in particular advocating for the relevance of Durkheim’s typology to researching the ‘socioemotional foundations’ of suicide.

However, their ‘microsociological view’ retains the functionalist spirit of Durkheim’s original typology, arguing for an extension and elaboration of Durkheim’s original insights. Further, while Abrutyn and Mueller have begun to incorporate interpretive approaches in more recent work (Mueller and Abrutyn, 2016), at times, they maintain a cautious distance from qualitative empirical research – referring, for instance, to the ‘shaky’ nature of qualitative data (Abrutyn and Mueller, 2014: 332), while nonetheless acknowledging the need for qualitative suicide research (p. 345).

As sociologists working on suicide readily acknowledge (Abrutyn, 2017), the study of suicide in the late twentieth and early twenty-first centuries is overwhelmingly dominated by ‘psy’ sciences and by biology – seeking answers to suicidal ‘impulses’ and ‘risk profiles’ in genetics, epigenetics, hormones, neurochemicals and synapses (Hjelmeland and Knizek, 2016). This biological tendency is undoubtedly related to the dominance of psychiatry within the multidisciplinary field of suicidology (Marsh, 2010), and to the concomitant valorization of biological, neurobiological, and genetic research in psychiatry (Insel, 2014; Pickersgill et al., 2013; Rose, 2013). However, as Pickersgill (2011) has noted in relation to psychiatry and personality disorder, psychiatric understandings and responses to distress may incorporate ‘surprising commitments’ to the social and the psychological. In this article, I seek to excavate and critique similarly ‘surprising’ commitments in both sociological and psychological work addressing suicide and socioeconomic inequalities.

Suicide and self-harm are acts which directly engage the social, psychological and biological, unsettling attempts to demarcate distinctions between these (Fitzpatrick et al., 2015). However, despite this, much suicide research continues to operate in disciplinary silos. By focusing on psychological and sociological attempts to understand the relationship between socioeconomic inequalities and suicide, I demonstrate some of the significant promise of more ambitious interdisciplinary endeavours. As such – drawing on Fitzgerald et al.’s (2016) recent call for a ‘revitalization’ of sociology, I argue that a similar ‘revitalization’ of suicidology is required – one which enrols psychology, sociology and other disciplines in working together to understand the ‘mutual entanglements’ of bodies, minds, cognitions, emotions, social and cultural contexts, in which suicides take place, are experienced, understood and given meaning. My arguments can be situated within the emergent field of ‘critical suicide studies’ which works towards complexity, creativity and nuance in suicide research (White et al., 2016).

Socioeconomic contexts of suicide and self-harm

Although there are debates on measurement, in the UK, people living in areas of socioeconomic deprivation, who are themselves relatively socioeconomically deprived, are at higher risk of death by suicide, and higher risk of hospitalization following self-harm, than those living in more affluent circumstances (Mok et al., 2012; Platt, 2011; Platt et al., 2005). Historically, rates of suicide have fluctuated, and – following Durkheim’s original arguments – in many cases these can be associated at least superficially with changing economic circumstances (Baudelot and Establet, 2008). While Durkheim originally argued that suicides were concentrated among the more affluent, with poverty being a ‘protective factor’, Baudelot and Establet note that the evidence for this is far from clear (2008: 141). In part, this relates to the limited nature of available historical data. What data is available appears to suggest higher rates of suicide among both the very poor and the very affluent. However, in the latter half of the twentieth century, this changed: ‘Suicide no longer haunts the more affluent section of society, and is now concentrated amongst the poorest and the most fragile’ (Baudelot and Establet, 2008: 147). Epidemiological studies of suicide in the UK suggest that higher rates of suicide among poorer groups can be related to unemployment, job insecurity, housing insecurity, debt, and the impact these experiences have on mental health and substance use (Coope et al., 2014; Gunnell et al., 2003; Haw et al., 2015).

The picture is similar, though perhaps more contested, with self-harm. ² Measuring rates of self-harm is challenging, as most acts of self-harm are not clinically treated, and one of the more prevalent self-harming practices (self-cutting) is generally viewed as ‘hidden’ (Chandler, 2018). Nevertheless, data has been collected in some UK hospitals for several decades, providing longitudinal data regarding rates of hospital-treated self-harm (Geulayov et al., 2016). Such cases are usually overdoses (70–80%), although they do include other acts such as self-cutting, burning, jumping from heights, and hanging. As with suicide, rates of hospital-treated self-harm are higher in poorer areas, and rose following the 2008 recession (Geulayov et al., 2016). A review of data regarding General Practice presentations found a similar pattern, with rates of recorded self-harm higher among patients presenting to practices in areas of deprivation (Carr et al., 2016).

Despite the frequent identification of socioeconomic contexts as a ‘risk factor’ in suicides, mainstream suicide research and prevention efforts have tended not to engage with this (Platt, 2011). Further, in much sociological work, there is a similar – perhaps more curious – disengagement from investigating the role of socioeconomic contexts in suicidal practices. For instance, Wray et al.’s (2011) review of studies of suicide in US sociology journals addresses socioeconomic context or markers only briefly, factors that are ‘controlled’ for in studies of race and suicide. Stack’s comprehensive, two-part review (Stack, 2000a; 2000b) of sociological research on suicide includes a section on ‘economic approaches’ to understanding suicide. Stack includes a number of interesting suggestions regarding how economic conditions, in particular, unemployment, may shape suicidal practices. However, crucially, much of this draws on quantitative research, and involves necessarily making assumptions about how such conditions might impact on individual actions, and the meanings attributed to these:

Unemployment can affect suicide risk directly through eroding the incomes, economic welfare, self-esteem, and other suicidoegenic factors among the unemployed…Furthermore, in times of unemployment, suicide may be increased through a higher level of anxiety among the employed who fear losing their jobs. (Stack, 2000a: 155)

Existing sociological work on suicide and socioeconomic contexts has, then, tended to focus squarely on issues which might comfortably be seen as ‘social’: drawing on theories of strain, examining macro-economic contexts, unemployment rates, and referencing the potential emotional impacts of these, but rarely in any detail. However, this is a very particular form of sociology (functionalist, quantitative) – one that ‘imagines’ how these factors may play out for individuals in terms of their emotions – terror, self-esteem, anxiety – but which fails to engage empirically with this. A rich body of sociological theorization on more intimate contexts – micro sociology, symbolic interaction – and on the emotional tenor of social life (cf. Denzin, 2007; Scheff, 1990; Williams and Bendelow, 1998) is carefully – and curiously – avoided.

Dying from inequalities: sociological and psychological contributions

The inspiration and material for this article are drawn from two chapters of a report (Chandler, 2017; Kirtley and O’Connor, 2017), commissioned by Samaritans (2017). As noted above, the report drew on a range of disciplinary expertise to enhance understandings of the mechanisms that might explain the relationship between socioeconomic disadvantage and higher rates of suicide and self-harm in the UK. Contrasting the findings of the two chapters, and of the papers that were included in the reviews, provides useful insights into some of the similarities and differences between sociology and health psychology in how the issue of socioeconomic disadvantage and suicide is tackled and understood. This approach is not intended to provide an exhaustive picture of sociological and psychological approaches to this issue, but rather serves to illustrate important resonances and disjunctures between and within the bodies of work reviewed in the chapters.

My chapter comprised a rapid literature review (conducted in 2016) and synthesis of qualitative research reporting the accounts of people who had self-harmed, or died by suicide, which acknowledged or engaged with socioeconomic disadvantage. Although 35 papers were included in the review, only seven focused explicitly on socioeconomic disadvantage (Abrams and Gordon, 2003; Barnes et al., 2016; Elliott et al., 2014; Huey et al., 2014; Kidd, 2004; Redley, 2003; Stack and Wasserman, 2007). Indeed, an ancillary finding of the review was that studies tended not to report anything regarding socioeconomic contexts in qualitative studies of self-harm and suicide, much less to take socioeconomic conditions as a focus of analysis. The chapter presented a qualitative synthesis (as set out by Rhodes and Treloar, 2008) of both the findings and analyses present in the reports, informed by an interpretive, interactionist sociological perspective.

Kirtley and O’Connor (2017) reviewed psychological research, examining papers which addressed the psychological ‘mediators and moderators’ of the relationship between socioeconomic disadvantage and suicide. Kirtley and O’Connor also found that research engaging directly with this relationship was scarce. Their review included 49 papers which were synthesized, and analysed in light of O’Connor’s Integrated Motivational-Volitional (IMV) model of suicidal behaviour (O’Connor, 2011). The IMV model has broad similarities with other key ideation-to-action psychological models which seek to explain suicide from a psychological perspective (Klonsky et al., 2017). These tend to invoke a linear accumulation of background stressors (environment, which can include socioeconomic context, and life events, such as childhood trauma), specific feelings and personality traits (e.g. feelings of burdensomeness, tendencies towards rumination, memory deficits), and acute behavioural features (including suicidal behaviour or self-harm, but also addressing access to means, and suicide planning).

Despite the different bodies of research examined, and the different approaches taken to interpreting the data, there were important resonances between the two chapters. In the following sections I draw out these resonances, and in so doing, point to significant intersections between, as well as disconnects within the findings in each chapter.

Cumulative disadvantage

The theme of cumulative disadvantage featured in both the psychological literature reviewed by Kirtley and O’Connor, and in the qualitative research. In the psychological literature, this theme took the form of ‘adverse childhood experiences’ (ACEs). Numerous studies indicated that experiencing a greater number of ACEs (such as neglect or abuse) was correlated with suicidal thoughts, self-harm and completed suicide in a range of different populations. ACEs were more commonly reported by those whose parents had lower levels of education (a proxy for socioeconomic status). A crucial additional argument made by Kirtley and O’Connor concerned the role of ACEs in also increasing the risk of experiencing socioeconomic disadvantage as an adult. Thus, ACEs were framed as doubly contributing (although not neatly) to suicide risk in adulthood, both through independent correlation of ACEs with suicide attempts, and through increasing the chance of living with socioeconomic disadvantage as an adult, which is also independently correlated with suicide. Kirtley and O’Connor’s findings suggest a significant amount of complexity, which is sketched out via quantitative findings, but clearly requires additional data and theory to better understand how these data points result in greater suicide risk.

The qualitative studies reviewed also pointed to the importance of attending to early experiences of difficulty – particularly trauma (e.g. childhood sexual, physical or emotional abuse or neglect) and bereavement or loss. Accounts of participants reported in the studies frequently identified bereavement as contributing to an act of self-harm, or desire for death. As Redley (2003) pointed out, participants were telling him (in interviews) about the problems they had faced in life, as a way of explaining why they had self-harmed. Padgett and colleagues (2012) noted that participants in their study suggested that early experiences of trauma were most significant in explaining suicidal feelings, but that this could be overlooked among groups who faced acute and recent challenges, such as homelessness or severe mental illness.

Curiously, two of the sociologically informed qualitative studies displayed a tendency to draw on qualitative data in order to quantify the impact of such experiences (Huey et al., 2014; Stack and Wasserman, 2007). As such, findings and arguments resonated with those identified in the review of psychological literature both in terms of themes, and in terms of the analytic approach to the data – despite the data being of a very different form (structured, validated psychological surveys and qualitative data from interviews, or reviews of coroner reports). This emerged as a clear cross-cutting theme – an intersection – in how papers written from a psychological and sociological perspective regarded the relationship between socioeconomic disadvantage, ACEs and later suicide or self-harm. These analyses tended to treat the accumulation of disadvantage in a straightforward, additive manner. Even where theoretical models (e.g. strain theory, drawn on by Huey et al., 2014, and by Stack and Wasserman, 2007) were enrolled to better understand the relationship between accumulations of disadvantage, and self-harm and suicide, these did not go much beyond a fairly mechanical imaginary, whereby a person experiences a certain number of negative events, resulting in the outcome of self-harm and suicide. The emotional, social and cultural meanings of these events, how they were narrated, experienced, or understood by individuals in particular social and cultural contexts were absent, or little remarked upon. This resulted in analyses which ultimately stopped short of explaining how and why such accumulations of disadvantage result in greater chances of self-harm and suicide in a given context.

Analyses which drew on the notion of cumulative disadvantage also failed to explain why individuals who experienced a particular number, amount or type of ACEs might not experience suicidal thoughts or enact self-harm or suicide. In contrast, one of the qualitative studies reviewed noted a surprising ‘lack of association between suicide attempts and past histories of abuse, abandonment and neglect’ among a group of male prisoners (Byng et al., 2015: 940, emphasis added). The disjuncture between qualitative findings within a particular population (men leaving prison) and quantitative studies of risk factors in more general populations indicates both the need for, and the challenge of doing, suicide research which moves between the specific and the general. Thus, while a quantitative study of a general population sample may find that, on the whole, ACEs are correlated with later suicidal thoughts or behaviour, this level of generality may not work well with specific groups – particularly those who already embody other risk factors for suicide (e.g. in terms of gender, age and socioeconomic background). In the following two sections, I address two further themes which illustrate other means through which to understand the accumulation of suicide risk among people living with socioeconomic disadvantage.

Emotions and inequalities

The role of emotions in mediating or shaping the relationship between socioeconomic disadvantage, self-harm, and suicide was another clear intersection between the two chapters. Further, the way in which emotion was conceptualized within the studies reviewed provides some instructive points regarding the close relationship between sociological and psychological thought – particularly as this pertains to what is often perceived as ‘inner’ and as unproblematically the realm of the ‘psychological’ (Burkitt, 2014; Williams, 2001; Williams and Bendelow, 1998). Indeed, while emotions were clearly present, and the focus of analysis in some qualitative sociological studies (Fullagar, 2003), in others, emotions were either absent, or treated superficially and descriptively (Huey et al., 2014). This reflects sociological work on suicide more generally, which has until recently (see Abrutyn and Mueller, 2014) almost entirely neglected a rich body of critical work on the sociology of emotions.

The psychological literature reviewed by Kirtley and O’Connor (2017) highlighted the role of specific emotions or feeling-states, namely, defeat, entrapment, humiliation and shame, though they were forced to extrapolate findings somewhat, since they were unable to identify psychological studies which addressed these factors in relation to socioeconomic disadvantage. Reviewed studies found that individuals scoring more highly on scales measuring defeat, entrapment and humiliation tend to report greater levels of emotional distress and in some cases suicidal ideation. Kirtley and O’Connor suggested that experiencing socioeconomic disadvantage ‘may induce feelings of entrapment and defeat’ (2017: 113), thus providing a theoretical connection between socioeconomic disadvantage and suicidal outcomes. Additionally, they note the potential role of stigma for those living with socioeconomic disadvantage, also highlighting evidence that feelings of low self-worth endured even when personal finances improved. Their findings point to an important role for emotions and feelings – for affective evaluations of the self, and how these relate to social contexts.

Emotions were a key theme in the findings and analyses of qualitative papers reviewed. Perhaps unsurprisingly, the way in which emotions emerged in these papers diverged significantly from the quantitative studies, although there were also important divergences among the qualitative studies. In part, this can be clearly related to researchers working in a range of disciplinary traditions. One of the papers reviewed reported on a qualitative psychology study of young people’s experiences with suicidal thoughts (Everall et al., 2006). In contrast to the quantitative psychology studies reviewed by Kirtley and O’Connor, Everall et al. argued that ‘[r]ather than describing emotional traits, participants’ experiences included complex interactions of negative emotions characteristic of overwhelming despair’ (2006: 385–6). This highlights a rather obvious distinction between the data generated by quantitative and qualitative work. While quantitative psychological studies might seek to identify the relative importance of specific, measurable traits (including emotional states); qualitative psychological studies demonstrate the ‘messier’ ways that such traits emerge in individual accounts.

However, a further divergence can be identified among the qualitative studies reviewed. While most (though certainly not all) emphasized complexity in the way in which emotions were articulated by participants, the manner in which emotions were conceptualized varied. For instance, Everall et al. interpreted participants’ accounts of emotional problems in light of their individual life histories, focusing on early care-giving, and the role of parents in validating emotions, or failing to model appropriately healthy emotions. They noted that ‘participants tended to interpret difficulties with emotional regulation and expression as signifying personal inadequacy’ (2006: 388). This finding goes unqueried. In stark contrast, Fullagar’s (2003) and Kidd’s (2005) analyses of similar accounts portray the role and nature of emotions very differently. Fullagar argues that ‘individualizing discourses of neo-liberalism are implicated in perpetuating shame’ (2003: 300), as well as in encouraging youth experiencing suicidal thoughts to feel personally responsible (and thus doubly shamed) for their emotions (see also McDermott and Roen, 2016). Kidd, in analysing accounts of street-involved young people, also links their experiences of shame, and entrapment, with social stigma, supported by social policies which marginalize and oppress. This points to a difference in focus, with more sociologically informed research addressing the impact of wider ideologies or policies on individual emotional experiences, while those writing in a more psychologically oriented manner appear to limit ‘context’ to individual life-histories and close family relationships.

The relevance of shame

The qualitative review identified two emotions as important in both researchers’ analysis of suicide and self-harm in the context of socioeconomic context, and in the accounts of participants: anger and shame. While anger was not highlighted specifically in Kirtley and O’Connor’s review, shame was. Further, shame is an emotion which has special relevance to the study of socioeconomic inequalities and suicide (Walker, 2014). Indeed, Scheff has made some initial suggestions regarding this. Scheff’s analysis of The Sorrows of Young Werther (Goethe, 1774, cited in Scheff, 1990), a fictional account of the life and suicide of a young man, addresses the emotional basis – and potentially fatal implications – of false consciousness. Scheff draws here on Marx’s notion of false class consciousness, proposing that attending to emotion can help explain why members of subordinate classes tolerate their oppression (1990: 120–1). This analysis is situated in a larger body of work by Scheff (with Retzinger) which argues that shame is ‘ubiquitous’ in modern societies, but that it is also often unacknowledged (Scheff, 1990; 2000; 2014; Scheff and Retzinger, 2001). Unacknowledged shame, it is argued, can lead to ‘rage-shame’ spirals which may culminate in suicide. With respect to the story of Werther, Scheff’s analysis explicitly draws out the relevance of social structures, and inequalities in wealth and status. Scheff’s work is limited, however, in that his theoretical approach has not been empirically tested among people who have self-harmed or died by suicide.

In contrast, McDermott and Roen (2016) have drawn on alternative literatures (e.g. Ahmed, 2014; Sedgwick, 2003) to theorize the embodied, affective experiences of shame and self-harm among young, queer youth from lower socioeconomic backgrounds. Their analysis somewhat contradicts Scheff’s claim that shame is ‘ubiquitous and hidden’, highlighting the ways that shame is unequally distributed, or perhaps more pointedly, how resources through which shame might be resisted are unequally distributed. Further, while they also frame shame (and self-harm) as hidden and secret, they suggest that when shame becomes visible – when it is acknowledged – it can be felt more intensely. Like Fullagar (2003), McDermott and Roen note the importance of wider social and cultural contexts in shaping experiences of shame, and in understanding how such affective feelings might relate to self-harm. Further, their analysis addresses the ways in which some young people (particularly those with greater material and educational resources) might be able to navigate a life away from shaming contexts (such as school).

Agency and the status of accounts

The themes addressed thus far are each liable to result in structurally oriented accounts of the role of socioeconomic disadvantage, which tend to minimize the agency of the person self-harming or dying by suicide. Socioeconomic disadvantage either accumulates in a person’s life ‘resulting’ in suicide or self-harm; or else it contributes to emotional responses, which similarly ‘result’ in suicide or self-harm. This is illustrated in a paper by Barnes and colleagues, in which they analysed interviews with people who identified ‘economic hardship’ as a ‘precipitating feature’ in an act of self-harm. In their analysis the authors repeat a particular phrase, with participants’ accounts of a given event/circumstance said to have ‘led to self-harm’ (2016: 4). Self-harm is framed in the analysis as leading almost inevitably from a series of events, and this is a feature of many of the other qualitative papers reviewed in the chapter.

The act of self-harming or suicide is often unexamined, as is the agency of the person carrying out this act. In part, this can be understood in the broader context of suicidology which is – as discussed above – dominated by approaches which are wedded to an understanding of suicide and self-harm as pathological, and relating to mental illness (Marsh, 2010). As such, the agency of the individual at the centre of the suicidal event is portrayed as perhaps inherently untrustworthy, or ultimately irrelevant (since it is shaped by pathological factors outside of the individual’s control – genetics, neurochemistry, childhood trauma, and so on). Jaworski (2014), however, has argued that the ascription of agency in suicide and self-harm is strongly gendered. Suicide, associated with masculinity and the ‘active male’, is more likely, she suggests, to be framed as agentic. In contrast, self-harm is associated with femininity, passivity and an absence of agency. It is possible that the overall picture that emerged in the review of qualitative accounts was shaped by this: the majority of studies were of people (often, but not exclusively women) who had self-harmed. Thus, the tendency in many cases towards framing self-harm as a passive outcome of circumstances may be related to what Jaworksi terms the ‘gender of suicide’.

Attending to qualitative accounts of self-harm and suicide, and examining how these might contribute to an understanding of the relationship between suicide and socioeconomic inequalities, necessitate a critical examination of agency, and with this, the status of accounts about self-harm and suicide. Such an approach is vital in order to avoid ‘top-down’ theorization which can result in a failure to explain why someone with accumulated disadvantage, who is shamed by their social status, who experiences further traumatic events which may be tied to socioeconomic circumstance, does not self-harm, or die by suicide (Byng et al., 2015; Redley, 2003). To suggest such occurrences are explained by a particular set of ‘risk factors’ embodied in an individual, runs the risk of all kinds of reductionism – simplifying a process which is agentic: as Byng and colleagues suggest, suicide could be understood as ‘one of life’s ultimate acts’ (2015: 937). ³

Sociological theory and research continue to grapple with the concept of agency (Burkitt, 2016; Emirbayer and Mische, 1998; McNay, 2008), and reflection on the position of qualitative accounts of suicide and self-harm within suicidology highlights the enduring importance of these debates. Agency and the ascription of motivation are central to long-standing deliberations regarding the definition of self-harm and suicide (Chaney, 2017; Jaworski, 2014; Millard, 2015). This issue has recently been played out in discussions regarding the DSM-5 (American Psychiatric Association, 2013) and the inclusion of proposed new diagnoses: Suicidal Behaviour Disorder, and Non-Suicidal Self-Injury (Chandler, 2016). These diagnoses rest, ultimately, on the ascription of motive to a particular act (self-harm), which in turn relies (perversely, given the subordinate position of qualitative research in suicidology) on the accounts of those who self-harm. ⁴ Both psychological and sociological perspectives can be enrolled to trouble such an ascription. Within suicidology, psychiatrists and psychologists alike have noted that motives – and methods – for self-harm may change over time, be multiple and contradictory, or reflect deep ambivalence about life or death (Arensman and Keeley, 2012; De Leo, 2011; Kapur et al., 2013). Sociologically, the status of accounts have long been an issue of importance (Mills, 1940; Scott and Lyman, 1968), and these theoretical and methodological discussions are directly relevant to understanding debates about motives, and the status of accounts about self-harm and suicide (Fincham et al., 2011).

Enrolling agency in suicide research

Two studies reviewed in the qualitative chapter stand out in placing agency at the centre of analysis while also attending to the role of socioeconomic circumstances. Given Jaworski’s comments on the importance of gender in considerations of agency in suicide, it may be notable that one of these studies included a balanced sample of 25 men and 25 women, while the other interviewed men only. Redley (2003) sought to examine ‘how respondents interpret their lives’ as well as ‘what they report’ (p. 351), in interviews with 50 people who lived in an area of multiple deprivation and had been treated in hospital for self-harm. Such an approach was surprisingly absent across many of the other qualitative papers reviewed, which focused more often on ‘what’ accounts of suicide and self-harm contained rather than ‘how’ they were narrated (Barnes et al., 2016; Elliott et al., 2014; Neale, 2000). In part, this may relate to the paucity of qualitative approaches in suicide research – meaning that researchers are of necessity addressing more basic questions in the first instance.

A more recent paper, by Byng and colleagues (2015) significantly pushes forward a sociological agenda in interpreting accounts of suicide and self-harm, drawing on Emirbayer and Mische’s (1998) comprehensive examination of sociological approaches to the concept of agency. Their study examined accounts of men before and after release from prison, some of whom had a history of self-harm and suicide attempts, some who did not. As such, Byng and colleagues were able to interrogate the question of ‘why not?’ self-harm/suicide raised by Redley’s analysis (2003: 398).

A significant aspect of Byng et al.’s analysis is the role of mastery, and their enrolment of both sociological and psychological theory in seeking to understand the variations between participants’ reported histories of suicidal thoughts, self-harm and adversity. They draw on the notion of mastery as a ‘pre-existing character trait’ (2015: 949) which may help explain why some individuals might ‘succeed’ ⁵ in carrying out suicide, where others embody more ambivalence, and less mastery, resulting in cycles of ‘chaotic’ non-fatal self-harm and substance use. Importantly, Byng et al. note that much of the focus of existing suicide prevention focuses on those who present with repeated self-harm. While presenting at hospital with self-harm is associated with an increased risk of death by suicide, many suicide deaths are enacted by people with no history of self-harm. This argument underlines the relevance of theoretically informed analyses which draw on both sociological and psychological traditions, in order to inform policy and practice in suicide prevention.

Byng et al. identify a group of individuals who engage in repeated acts of self-harm, and who describe ‘chaotic’ lives in which they feel powerless, categorizing these as embodying ‘restricted agency’, which is primarily iterative (following Emirbayer and Mische, 1998) or akin to Margaret Archer’s ‘fractured reflexives’ (Byng et al., 2015: 948). This analysis resonates strongly with other qualitative accounts, but also with findings in the health psychology literature. For instance, Kirtley and O’Connor’s (2017) discussion of defeat and entrapment indicates significant parallels with the notion of ‘restricted agency’. Further, in both analyses, a link is made with the role of life histories. Emirbayer and Mische’s concept of iterative agency relates to an individual reflecting and drawing upon a remembered history, while Kirtley and O’Connor highlight the role of defeat and entrapment in inhibiting autobiographical memory and limiting an individual’s ‘social problem-solving ability’. As such, the concept of agency has clear relevance to psychological models of suicidal behaviour (such as O’Connor’s IMV), which engage directly with the role of individual choice and the ability to reflect, and the psychological factors which might mediate this. These insights indicate a potentially productive focus for sociological and psychological collaboration.

However, a key divergence between psychological and sociological approaches with regard to the role of life histories and their impact on current reasoning, reflection and agency lies in the different lenses applied. While the quantitative studies which inform psychological thinking maintain a necessary focus on an individual, with fairly linear reasoning; sociological theory encourages a broader, more complex view of the relationships between individual agency, social structures, and cultural contexts, filling in gaps which are featured in models like O’Connor’s IMV (e.g. in the form of background or environmental factors) but not fully fleshed out. This is reflected in Redley’s (2003) analysis, where he argues that the material realities of life in an area of multiple deprivation shape the availability of different ways of interpreting and giving meaning to one’s life and future options. Similarly, Fincham and colleagues (Fincham et al., 2011; Shiner et al., 2009) have highlighted the importance of life-stage, and the social expectations regarding this, in understanding the ways in which individuals reflected upon their lives in ways which might make suicide more possible. Finally, papers by Fullagar (2003) and Bryant and Garnham (2015) point to the importance of political and ideological contexts in shaping individual responses to hardship, in particular, the effect of neoliberalism on emphasizing personal responsibility for ‘failure’ to achieve or persevere in the face of economic challenges.

Building on intersections in future suicide research

Contrasting attempts by health psychologists Kirtley and O’Connor, and sociologist Chandler, to advance understandings of the relationship between socioeconomic inequalities and suicide highlights important intersections between health psychology and sociology, which may offer productive arenas for future interdisciplinary collaborations. In short, future suicide research would benefit immeasurably from careful, critical studies which draw on pluralistic methodologies (Hjelmeland and Knizek, 2011) which incorporate insights from psychology and sociology, as well as from other disciplines. Such collaboration is necessary if researchers are to advance efforts to understand and intervene in suicide on an individual and population level. The intersections noted above point to fertile starting points for such collaborative efforts.

The health psychology literature reviewed by Kirtley and O’Connor was shown to work well with O’Connor’s IMV model for understanding the psychological drivers for suicide, and how these may be adapted to understand the role of socioeconomic contexts. A strength of the IMV is its acknowledgement of the relevance of social contexts – via ‘environment’. However, as discussed above, reliance on quantitative, psychologically informed studies limits the ability of researchers to creatively think between macro and micro contexts. This is a key point to which sociological perspectives might productively contribute. Current sociological work addressing socioeconomic inequalities and suicide is also constrained, though, by a similar reliance on quantitative approaches. This reliance means that reflections on the impact of socioeconomic inequalities on individual cases of suicide or self-harm are, in most cases, little more than guesswork and imaginative projection.

A further criticism which has been levelled at psychological models of suicide, which draw largely on quantitative evidence, is their linearity, and what Hjelmeland and Knizek (2016) refer to as ‘cause and effect thinking’. Interestingly, the literature discussed in this paper evidenced such thinking across psychological, sociological and qualitative literatures. This is an important, though perhaps less promising, intersection. A useful illustration of the challenges of such thinking can be seen in the way in which emotions or feelings are enrolled across the studies reviewed. In O’Connor’s IMV model, feelings are clearly situated prior to a suicidal act/thought, and while positioned in a social context of sorts, are nonetheless framed as being within an individual. A person experiences defeat or humiliation, they feel entrapped, and for some individuals (with particular vulnerabilities, or exposures, e.g. to others’ suicidal practices), this may lead them to suicide or self-harm.

In contrast, other work on emotions and suicide, drawing on both sociology and on a post-colonial-inspired critical theory, emphasizes the way in which particular political and ideological programmes, enacted through policy, media reporting, and public understandings, may promulgate particular ‘psychic lives’ (Bryant and Garnham, 2015; Fullagar, 2003; Mills, 2018). In these analyses, drawing on qualitative data, emotions are understood to be generated socially through government policies (e.g. austerity), economic ideologies (e.g. neoliberalism), but internalized and reinterpreted as individual concerns and responsibilities. Such approaches question the status of accounts which unproblematically link feelings of, for instance, shame or humiliation to an individual response to difficult circumstances, emphasizing the cultural and social origins of shame, while retaining awareness of the individual and sometimes fatal consequences of such feelings.

Research such as Mills’ (2018) psychopolitical autopsy, Byng et al.’s (2015) enrolment of agency and mastery, and McDermott and Roen’s (2016) detailed analysis of accounts of queer youth who self-harm, all point towards a rich arena where sociological, psychological and theoretical expertise from elsewhere might work together to ‘revitalize’ suicidology (Fitzgerald et al., 2016). However, if this is to occur, then the methodological and theoretical disconnects highlighted in this article need to be addressed. The emergent field of ‘critical suicide studies’ is one such arena where this may occur, providing a space where more ‘dangerous liaisons’ between psychology, sociology and others might occur. However, if such work is to have greater impact, then these efforts must reach ‘mainstream’ suicidology as well. ⁶ Steps towards this would include the development of collaborative, multi-disciplinary and mixed methods research teams; and a change in the publication policies of mainstream suicide research journals (Hjelmeland and Knizek, 2011). Further, sociologists should be encouraged to undertake suicide research in greater numbers. Insights from sociology are much needed, and yet they remain curiously side-lined within suicide research.