The call to consider the social context of smoking is being increasingly heard in the tobacco control literature, not to replace (or deny) the importance of physiological dimensions of addiction, but to understand better two key phenomena in particular. First, social context is a factor in the growing concentration of smoking among socially and economically marginalised groups (the uneven social and geographic distribution of smoking). Second, social context may be key to understanding (and, ultimately, addressing) diverse sources of resistance to tobacco control.

This growing interest in “the social” parallels a broader shift within public health research from a biomedical model of illness and disease towards an understanding of the broader social determinants of health. Indeed, calls for greater attention to the social context of behaviour in the health promotion literature are not new. Such calls can be traced back to influential articles by McKinlay¹ and Labonte and Penfold,² among others.

Research has begun to examine the social context of smoking among specific population subgroups, including persons experiencing mental illness,³ homeless persons,⁴ older persons,⁵ pregnant and early postpartum women who smoke,^6,7 those living in areas of economic deprivation,^8,9,10 and adolescent smokers (especially young women).^11–13 Although the need to incorporate the social has now been recognised by many of the disciplines involved in tobacco control research, significant conceptual and operational (measurement) issues remain.^14,15 The lack of a shared understanding or agreement concerning what constitutes the social may stem in part from the range of epistemological and methodological approaches utilised in the primary disciplines involved in tobacco control research (that is, epidemiology, medicine, psychology). These varying disciplinary stances, coupled with an approach to tobacco use that is based on an addiction or lifestyle model, has resulted in an orientation that positions smoking primarily as an individual level health behaviour, and tobacco users as victims (of nicotine addiction; of unhealthy lifestyle choices; of the tobacco industry). As a result, the social meaning of smoking in the context of people’s everyday lives is underplayed.¹⁶

In this commentary we examine the nature and significance of the social as a domain of inquiry, and its relevance to smoking. We offer a brief overview and analysis of some of the strengths and persistent limitations in the literature and suggest avenues of theoretically informed research that we feel are particularly promising. In particular, drawing on a wide range of work in social theory from anthropology, sociology, cultural studies, and geography, we describe several key tenets or dimensions of the social that we believe are key to a fuller understanding of smoking as a collective social practice. Specifically, we wish to draw attention to the centrality of power relations in shaping the uneven social geography of smoking, as well as the following dimensions of the social: sociology of the body as it relates to smoking, collective patterns of consumption, the construction and maintenance of social identity, the ways in which desire and pleasure are implicated in these latter two dimensions in particular, and smoking as a social activity rooted in place. We are a group of social scientists from diverse disciplines (social epidemiology, geography, anthropology, sociology, human kinetics) from across Canada who have been funded to distil key insights from social theory for tobacco research and practice. Several of us (Poland, Frohlich, Sparks) have been involved, to varying degrees, in tobacco research.

WHAT IS SOCIAL CONTEXT?

It turns out that the social context that is so efficiently revealed through other media than the printed word, such as Max Harris’ photoraphic portrayal of smoking in “the biker”,¹⁷ or Travis Tritt’s new country song “Can’t tell me nothin’ ”,¹⁸ is rather difficult to unpack theoretically. Context can be defined as the circumstances or events that form the environment within which something exists or takes place and as that which therefore helps make phenomena intelligible and meaningful (interpreting something in context, versus out of context). The configuration of influences that support or hinder smoking at a variety of scales in terms of access to tobacco, advertising, prevalence and visibility of smoking, etc, are key, and it is these that have received the greatest attention in recent years. A recent (May 2003) special issue of the journal Addiction summarises and assesses research on the social context of adolescent smoking at a variety of scales: micro (family and peer social networks), meso (school, workplace, neighbourhood), and macro (policy, social acceptability, media). The bulk of the research reviewed focuses on explaining individual smoking behaviour by seeking to isolate predisposing and constraining factors at a variety of scales and settings, using statistical methods to assess robustness and generaliseability of association across populations and settings.

While this approach has generated important insights for the field, it suffers several limitations that suggest that further understanding will need to come from other approaches, both theoretical and methodological. First, the explanatory power of these models is, at best, modest, accounting typically for only a fraction of the variance in statistical terms. Second, even when it is successful in isolating factors statistically associated with variations in smoking, this approach tells us about what factors may be most relevant, but reveals relatively little about how or why they’re important. This is because decomposition of multiple (local) contexts into a series of factors or variables generaliseable across settings and populations (as is common in the field at the moment), with consequent emphasis on the allocation of variables to categories rather than preserving the integrity of context, means that factors are abstracted from context, denuded of context. Third, it is only relatively recently in tobacco control research and practice that there has been recognition that social context involves more than what is normally subsumed under the concept of “social influences” (“social norms”, peer pressure) or other aspects of the immediate interpersonal and physical environment (concepts that borrow heavily from social psychology, which has enjoyed a privileged disciplinary position within the field of tobacco control). In the investigation of young women’s smoking uptake, a new stream of “mixed method” studies suggests that there is much to be learnt by addressing young people’s social representations (that is, “images and identities” of smoking), mediated by age and sex based social group membership.^{11–13,19–21} These findings suggest that tobacco use might represent much more than adolescent experimentation or “risk behaviour”, and that smoking should be viewed as a practice that is very much linked to where, when, and with whom they smoke.²² One subfield where more sociological (and anthropological) analyses in tobacco research are to be found is research on women’s tobacco use. The influential work of feminist scholars such as Graham^23,24 and others^25,26 in the UK, and later Greaves²⁷ in Canada, has been critical to establishing the relationship between female smoking and social disadvantage, for example.

The work cited above has been enormously helpful in raising the profile and importance of the social context of smoking, and pointing to some of the dimensions that must be considered. Several of these studies have been informed by an explicit theoretical framework that guided the selection of empirical foci for examination. Several used novel methods of data collection (like peer network mapping) to address key aspects of the social context of smoking. Yet critical gaps remain. For example, with few exceptions,^{23,24,28–32} few have squarely addressed the issue of power relations and the central importance of place with respect to smoking. And none of this work provides a systematic and comprehensive basis for understanding the social embeddedness of collective social practices, for guiding how key dimensions of the social can be conceptualised with respect to smoking, and how they can be operationalised using qualitative and quantitative research methods. As previously noted, much research has been conducted on smoking in public health from an individual health behaviour perspective; in the social sciences, much work has gone into theorising the social world, but little work has brought these two spheres together. As a consequence, insights from social theory that may help us to understand the socially unequal distribution of smoking are still missing within mainstream discussions of smoking and have yet to gain prominence within tobacco control research.

As a beginning step in this direction, we have identified a number of dimensions of “the social” that we think are key to understanding the unequal distribution of smoking and also the role of tobacco control in shaping smoking practices (both historic reductions in prevalence and the persistence of smoking in certain groups). We express these in the form of working hypotheses or “tenets”, but caution against a deterministic reading of their possible import.

Before describing these dimensions of social context, we make a few key assumptions that need to be spelled out insofar as they may deviate from what is normally held in tobacco control. We use the prefix “social” (as in social context), not to downplay the importance of biology or physical environment (both of which are also shaped by the social, researchers are finding), but to underscore the importance of social relations and social structures. It is the local configuration of social relations (comprising social structures such as class, race, and gender; institutional practices, and collective and individual behaviour, and intersecting personal biographies) that constitute context, and that is key to our understanding. Further, we posit a dialectical relationship between agency and structure. That is to say, social structures (economic systems, culture, class and gender relations, to name just a few) influence and constrain human action, and what people think and do (their practices) serve to reproduce these same social structures (notwithstanding some resistance and modification, wholesale transformation of deeply structured social relations on a collective/regional scale is relatively rare). Also, we make a distinction between smoking as an individual health behaviour and smoking as collective social practice, favouring the latter. “Collective lifestyles” reflect a way of understanding behaviours as social practices, that is routinised and socialised behaviours common to groups (such as patterns of consumption).^33,34 These are generated at the intersection of social structure (norms, resources, policy, institutional practices that organise society), and agency (individual action, volition and sense of identity), and manifest concretely in specific places (for example, neighbourhoods). Following Bourdieu,^35,36 we regard lifestyle as an expression of “habitus” (practices and schemes of perception, preference and taste acquired over time as internalisation of objective life chances). Frohlich has used this concept to help understand the differential distribution of smoking among adolescents in Quebec.^29,30

DIMENSIONS OF SOCIAL CONTEXT

We identify six dimensions of social context from the social theory literature in geography, sociology, and anthropology that have the potential to generate important new insights, dimensions that are currently lacking in the mainstream tobacco research literature. These dimensions help us to flesh out the dialectic relationship between agency and structure and collective lifestyles. What we describe here can hardly be anything more than suggestive, given the space available here, and given the early stage of our work.

First, we recognise the central role of power relations in shaping the uneven social geography of smoking. Power is the capacity to act (power to), especially so that one’s interests prevail over those of others (power over).³⁷ Grabb identifies three fundamental bases of power: control of material resources (means of production, wealth), control of human resources (labour power), and control of ideas (ideology, hegemony, and cultural dominance).³⁸ Each has been the subject of intensive study and numerous publications. Suffice it to say, for the moment, that a focus on power relations draws attention to the ways in which the social and geographic patterning of smoking parallels the effects of other processes of marginalisation and disadvantage. It invites us to reflect on why this might be the case, whether that be the intentional racial and class profiling of tobacco advertising,³⁹ the role of smoking as a coping mechanism for dealing with the psychosocial sequelae of social disadvantage,²⁴ or other possible explanations.

Second, we underscore the central importance of physicality, sociality, and the body in smoking. Following the historic work of Mauss⁴⁰ and Bourdieu,^35,36 and a more recent and extensive literature on the sociology of the body⁴¹ we note that smoking is not only a social practice but also a bodily one, in that how the cigarette is held and smoked can demonstrate bodily competence (being “cool”) that must be acquired through practice,⁴² as new adolescent smokers who’ve been mocked by their peers for “not doing it right” can attest. Further, Bourdieu³⁶ has demonstrated the close intersection of social class and bodily capital such as physique (for example, stocky versus wiry frame) and gait, which would suggest that not only what one smokes (brand) but also how (for example, blowing rings), where, and with whom are closely linked to social position and cultural capital.

Two additional dimensions are drawn from the extensive literature on consumption and cultural studies. We would maintain that lifestyle practices are embedded in collective patterns of consumption selected from among what is economically and socially feasible/appropriate, in order to construct and maintain a social identity that both establishes and expresses difference among and between social groups. An interesting example of work in this area is Ioannou’s analysis of youth smoking in Cypress in the everyday contexts of consumption as part of what she terms the “active stylization of life” (linked to image management among peer groups).⁴³ The marketing practices of the tobacco industry are implicated in this broad examination of the close linkages between conspicuous consumption and identity formation/maintenance in late modern consumer societies. A closely linked fifth dimension is that of desire/pleasure, an area that has been receiving greater attention in the critical addictions^44,45 and health promotion literatures of late.^34,46 We understand this to include not only the pleasure associated with bodily competence and (to quote a participant in an earlier study) the “promeathean pleasure of holding fire”,⁴⁷ but also the ambivalent relationship that health promotion has historically had towards pleasure and desire (for example, popular resistance to “health police” who are seen to be hell bent on removing every last pleasure from the working class).

Last but not least, we see smoking as a social activity rooted in place. This is understood to include more than simply where smoking takes place (mapping utilisation). To quote John Agnew, “in order to explain human behaviour one must deal with the micro-episodes of everyday life and their embeddedness in concrete milieux or contexts.”⁴⁸ Space is both a condition (container) and a consequence of (shaped by) human activity.⁴⁹ Elsewhere, Poland et al,⁵⁰ building upon earlier work on settings for health promotion,⁴⁷ have described how distinctive cultures emerge in specific places that govern how people behave and the meanings that are derived from experience, and how technologically mediated power relations have particular consequences for how place matters for intervention design and implementation. It is understood that issues of race, class, and gender cut across and impact upon each of these dimensions. What remains is to investigate these empirically through in depth interviews with smokers (in progress).

SOCIAL CONTEXT: WHY BOTHER?

Why is the social context of smoking such a pressing issue? Because not fully understanding and accounting for socio-spatial disparities in smoking is, we suggest, hampering tobacco control efforts. Despite recent evidence of success among populations with lower income and education levels, further reductions in tobacco consumption will need to address the clustering of smoking among the socially disadvantaged and marginalised groups such as persons who have serious mental illness or who are homeless.^3,4 This concentration of smoking among particular subgroups of the population is not a naturally occurring or “random” event, but is tied to how societies are organised, and thus to the practices of institutions and persons therein. Underlying these inequities are complex social processes which are fundamental to understanding the continued prevalence and unique social distribution of smoking. Also, without wishing to exaggerate its prevalence or importance, resistance to tobacco control measures from a variety of sources must be acknowledged and appropriately responded to (not just dismissed). True, much of this resistance can be accurately described as the machinations of those with an economic interest in smoking, such as the tobacco industry, some advertisers and retailers, and some restaurant and bar owners. However, outside of a few cessation programmes, tobacco control seems to have, as a field, ongoing difficulty in engaging with smokers and in understanding what smoking means to them (for example, the voices and participation of smokers is notably absent from most tobacco initiatives, including research conferences). Yet we cannot afford to be out of step with what smokers think, feel, and need. Nor can we afford to reinforce unwittingly growing social class, educational, and other divisions between the target population (smokers) and those who design health promotion interventions (this too can become a source of irritation and resistance among low income smokers branded as “stupid” or uncouth by relatively privileged health professionals in ways that may compound class and race cleavages).

IMPLICATIONS FOR POLICY AND PRACTICE IN TOBACCO CONTROL

If there is a danger implicit in the exercise we have undertaken, it is the over-simplification of the “social” in ways that miss its essential contribution to collective lifestyles (for example, “class” relations imply more than is encompassed by variables such as socioeconomic status, measured at the individual level). What is required is a thoughtful operationalisation that preserves the complexity of the social context, and that while perhaps grounded in experience, does not overemphasise the individual to the exclusion of broader social processes. We therefore maintain that a key additional element of any investigation of the social context of smoking informed by social theory must be reflexivity with respect to the social (and historical/material) location of the researcher, the tobacco control practitioner, and of health promotion as a field. By reflexivity we mean maintaining a self critical attitude and questioning taken-for-granted assumptions regarding the (political) nature of our work and its (intended and unintended) effects, as well as the social distribution of these effects.^51,52 Reflexivity represents something of a post-modern break with the assumptions of relative predictability and certainty characteristic of a modernist emphasis on evidence based best practice.⁵³

Although it is still unclear what a more reflexive tobacco control practice will comprise (we are seeking funding to examine this in greater detail), an initial review of the literature suggests that it entails: (1) attention to the tacit knowledge and perspectives that practitioners bring to their work⁵⁴; (2) an openness to being transformed by the experience of engaging with smokers from very different social backgrounds who may question our (tobacco control) practices; (3) a questioning of “received knowledge” (what we hold to be self evident and true); (4) a curiosity about—and openness towards—other perspectives and ways of seeing; (5) mindfulness and presence; and (6) an awareness of power relations and one’s own social location and positionality (how we fit in to class and sex relations and how this affects the work we do, individually, and as a field).^53–56 We believe that a greater understanding of the social context of smoking will enable a greater sensitivity to the ways in which conventional tobacco control may unintentionally compound other forms of social exclusion among marginalised persons. In particular, we believe that current practice may be generating resistance to tobacco control among “hard to reach” groups, and may account for why such groups are experienced as “hard to reach”. This may be due in part to a mismatch in fundamental assumptions, “life-world” and lived experience between middle class professionals and their increasingly socially excluded “clientele”.³¹ Although it is still early, we foresee the potential for a better understanding of the social context of smoking to significantly enhance tobacco control research and practice.