Changing social attitudes to smoking

The downward trend in smoking in high-income countries has been associated with increasingly negative perceptions of smoking, both with respect to its health effects and its social acceptability.

Firstly, an increasing proportion of adults regard smoking as a risk to health. In the US, smoking prevalence declined from the 1950s across decades in which the proportion agreeing that smoking was a cause of lung cancer increased (from around 40 per cent in the early 1950s to over 90 per cent in the late 1990s) (Kim and Shanahan, Reference Kim and Shanahan2003; Link and Phelan, Reference Link and Phelan2009). In the 1970s, 58 per cent of US adults considered that passive smoking was harmful to health; a decade later the proportion had climbed to 88 per cent (MMWR, 1988). Similar patterns are evident in Britain. From the 1960s, public awareness of the link between smoking and lung cancer increased – and smoking prevalence declined (Marsh and Matheson, Reference Marsh and Matheson1983; Wald and Nicolaides-Bouman, Reference Wald and Nicolaides-Bouman1991). Today, surveys report high levels of public knowledge about the risks of smoking; less than 1 per cent of the population are not able to identify at least one health risk of smoking and nearly 90 per cent can identify one or more adverse effects of second-hand exposure to smoking (Wardle et al., Reference Wardle, Pickup, Lee, Hall, Pickering, Grieg, Moodie and MacKintosh2010).

Secondly and relatedly, public attitudes to smoking have changed. Summarising the changes in the US, Brandt noted that ‘the fragrant has become foul; an emblem of attraction has become repulsive; a mark of sociability has become deviant; a public behaviour now is virtually private’ (Brandt, Reference Brandt, Reynolds and Tansey1998: 176). A similar cultural shift is evident in the UK. In the 1940s, surveys described how ‘chocolates, cigarettes and really “classy” evenings out all put you one step higher on the ladder’ (Calder and Sheridan, Reference Calder and Sheridan1984: 134). By the 1960s, attitude change was well underway, with half the population agreeing that ‘smoking is a dirty habit’ (McKennell and Thomas, Reference McKennell and Thomas1967: 259). By the mid-1990s, over 50 per cent of non-smokers were reporting that they minded if a smoker smoked near them, a proportion that now exceeds 60 per cent (Lader, Reference Lader2009). Not wanting smokers nearby is not primarily about protecting one's health. While health concerns are cited, the unpleasant smell of cigarette smoking, including the smell it leaves on clothes, is the major reason given for disliking physical proximity to smokers; an increasing proportion of non-smokers report that being near a smoker makes them feel sick (Lader, Reference Lader2009).

Public health policies are seen to have played a central role in increasing public knowledge of, and changing public attitudes to, smoking (Troyer and Markle, Reference Troyer and Markle1983; Hammond et al., Reference Hammond, Fong, Zanna, Thrasher and Borland2006). In the US and the UK, health promotion campaigns were initially focused on the risks to the smoker (Berridge, Reference Berridge, Lock, Reynolds and Tansey1998; Brandt, Reference Brandt2007). From the mid-1970s, the risks to others, and specifically to children born to pregnant smokers, were incorporated into government campaigns (Berridge and Louglin, Reference Berridge and Loughlin2005). Through the 1980s and 1990s, smoking was increasingly recast as a threat to the wider public, a shift in public health discourse that was particularly marked in the US (Brandt, Reference Brandt1990; Kim and Shanahan, Reference Kim and Shanahan2003). Today, minimising non-smokers’ exposure to tobacco smoke has assumed a central place in tobacco control policy. The risks of exposure are communicated through information campaigns and warnings on cigarette packs as well as through environmental regulation: legislation governing smoking on public transport, in workplaces and in public places has progressively restricted the locations in which smoking is permitted.

As public health researchers recognise, these polices have both material and cultural impacts. They serve to protect public health, and particularly the health of those unknowingly and unwillingly exposed to smoking – and they also signal the social unacceptability of smoking and, by extension, of smokers. Analyses of internal tobacco industry documents make clear that this is precisely what the industry fears. As early as 1984, it was acknowledged that ‘the single most important issue facing our industry is the erosion of the social acceptability of smoking’. The internal memo continued (Philip Morris International, 1984):

Shifting social norms is hypothesised to be an important mechanism through which tobacco control policies operate, by contributing to a public sentiment which is anti-smoking and supportive of further restrictions on smoking (Chapman and Freeman, Reference Chapman and Freeman2008; Hammond et al., Reference Hammond, Fong, Zanna, Thrasher and Borland2006). There is evidence to support this hypothesis: communities where smoking is less acceptable have lower smoking rates, lower cigarette consumption and, among smokers, greater willingness to quit (Alamar and Glantz, Reference Alamar and Glantz2006; Kim and Shanahan, Reference Kim and Shanahan2003). It is therefore concluded that ‘increasing the social unacceptability of smoking is a highly effective policy tool . . . Tobacco control programs should stress the dangers of environmental tobacco smoke and reinforce the non-smoking norm’ (Alamar and Glantz, Reference Alamar and Glantz2006: 1362).

Reinforcing the non-smoking norm is described as ‘social denormalisation’, a process that seeks ‘to push tobacco out of the charmed circle of normal desirable practice to being an abnormal practice’ (Hammond et al., Reference Hammond, Fong, Zanna, Thrasher and Borland2006: 225). It is widely regarded as essential if tobacco policies are to make headway in further reducing tobacco-related morbidity and mortality, particularly in the face of the continuing influence of the tobacco industry (Feldman and Bayer, Reference Feldman and Bayer2004). However, while the overall approach is not fundamentally questioned, public health researchers have expressed concerns that policies to denormalise smoking can, by design or default, denigrate smokers. As Stuber et al. (Reference Stuber, Galea and Link2008: 422) observe about policies restricting smoking in public spaces, ‘although smoke free air laws are imposed on the act of smoking and not on the smoker as an undesirable type of person, one need only look at the huddle of smokers commonly seen outside public buildings in inclement weather to witness the decreased social standing of smokers relative to non-smokers’. A smoker in a recent Scottish study put it more bluntly: ‘I think it's a load of crap, I don't like to be made to feel like a leper and feeling as if we're being persecuted because we smoke’ (quoted in Ritchie et al., Reference Ritchie, Amos and Martin2010: 625). As these comments suggest, stigma has emerged as an important issue for tobacco control policy.

The debate on smoking and stigma

Discussions about smoking and stigma are grounded in Irving Goffman's account of the experiences of the individual who ‘possesses a trait that can obtrude itself upon attention and turn those whom he meets away from him’ (Goffman, Reference Goffman1963: 15). While stigma is often treated as an individual attribute, Goffman emphasised that it is socially produced; his interest is therefore in ‘the structural preconditions of stigma’ (Goffman, Reference Goffman1963: 11). Social inequalities figure prominently among these preconditions (Link and Phelan, Reference Link and Phelan2001; Parker and Aggleton, Reference Parker and Aggleton2003). For identities to be successfully ‘spoiled’, those on the receiving end must be disadvantaged relative to the individuals, agencies and institutions engaged in denigrating them: ‘stigma is entirely contingent on access to social, economic, and political power’ (Link and Phelan, Reference Link and Phelan2001: 375).

Goffman's work has informed studies of smoking for over two decades. A Canadian study conducted in the 1980s drew on his insights to document how both smokers and non-smokers operated with negative stereotypes of smokers, seeing them as ‘dirty’, ‘inconsiderate’ and ‘weak-willed’ (Goldstein, Reference Goldstein1991). Through the 1990s and early 2000s, smoking research continued to reference Goffman (for example Brandt, Reference Brandt, Brandt and Rozin1997; Kim and Shanahan, Reference Kim and Shanahan2003). However, the last five years has seen a more sustained engagement with stigma, both theoretical and empirical (for example, Bayer and Stuber, Reference Bayer and Stuber2006; Burgess et al., Reference Burgess, Fu and van Ryn2009; Ritchie et al., Reference Ritchie, Amos and Martin2010; Thompson et al., Reference Thompson, Pearce and Barnett2007).

While a brief summary cannot capture the differences of view and nuances of argument that characterises this work, it can identify the central foci of debate. Describing the negative ways in which smoking is depicted is one major focus (Goldstein, Reference Goldstein1991; Stuber et al., Reference Stuber, Galea and Link2008). For example, an Australian study of cultural representations of smokers in various media, including news reports, online polls and dating websites, found them characterised as malodorous, selfish, unattractive and excessive users of public services (Chapman and Freeman, Reference Chapman and Freeman2008). Qualitative studies have confirmed these findings. Non-smokers describe smoking as ‘a disgusting habit’ and smokers as ‘outcasts’ and ‘lepers’ marked by their smell (‘reek’, ‘stink’, ‘stale’) and appearance (‘dirty’, ‘brown teeth’, ‘grey, dry, wrinkly skin’) (Farrimond and Joffe, Reference Farrimond and Joffe2006). Smokers are acutely aware of these negative depictions and report a pervasive social disapproval which they feel reflects ‘an unfair stigmatizing judgement of them as a person’ (Louka et al., Reference Louka, Maguire, Evans and Worrell2006: 445; see also Platt et al., Reference Platt, Amos, Godfrey, Martin, Ritchie and White2009; Ritchie et al., Reference Ritchie, Amos and Martin2010). However, not all researchers are convinced that being a smoker qualifies as a stigma. For example, the discrediting trait that smokers possess is seen to have a transitory quality, operating when they smoke but absent at other times (Ritchie et al., Reference Ritchie, Amos and Martin2010). Burris has more fundamental concerns. While accepting that smoking is widely regarded as undesirable, he rejects ‘an effete sensitivity in which even the least whiff of social disapproval of a behavior is seen as coercive or stigmatizing’ (Burris, Reference Burris2008: 475).

A second area in the debate about the smoking and stigma concerns ‘policy-induced stigma’ (Bayer, Reference Bayer2008: 470), a term that signals how policies can mobilise and sanction hostility towards those deemed to be a threat. The potential for such effects is widely appreciated within the public health community: the history of public health is scarred by policies which, pursued in the name of health protection and promotion, have served to intensify public vilification and state-sanctioned discrimination against already disadvantaged groups. Studies have documented how, across the nineteenth and twentieth centuries, poorer communities, including migrant and indigenous groups, were cast as ‘the contaminating other’ whose habitual behaviours were seen to threaten ways of life that were, in contrast, presented as normal, healthy and desirable (Finch, 1996; Petersen and Lupton, Reference Petersen and Lupton1996). In the more recent past, public health responses to HIV/AIDs are similarly seen to have intensified pre-existing inequalities constructed around sexuality, gender and social class (Bayer and Stuber, Reference Bayer and Stuber2006; Parker and Aggleton, Reference Parker and Aggleton2003).

Given this ignoble history, it might be expected that stigma-inducing policies would be abhorred. Indeed, concerns are raised (see, for example, Bayer and Stuber, Reference Bayer and Stuber2006; Burgess et al., Reference Burgess, Fu and van Ryn2009). But they typically fall short of unequivocal condemnation. Instead, the broad conclusion is that tobacco control policies can legitimately involve a degree of stigma if they achieve their objective of protecting people's health. The moral claims of non-smokers figures prominently in this assessment. Thus, discussing the restrictions on smoking in public places, Ritchie et al. (Reference Ritchie, Amos and Martin2010: 68) argue that ‘it is inevitable that such a policy will cause some marginalisation (or stigma) of some smokers . . . but that . . . is justified by the protection of non-smokers from physical harm’. The potential health benefits of stigmatising policies are seen to extend to smokers, for example by encouraging them to quit: ‘if they work, they may represent a significant contribution to the wellbeing of the very people they burden’ (Bayer, Reference Bayer2008: 470). The continuing need to blunt corporate efforts to promote their product lends further weight to the case that stigma can be an acceptable instrument of public health policy (Chapman and Freeman, Reference Chapman and Freeman2008; Ritchie et al., Reference Ritchie, Amos and Martin2010). While this principle is broadly accepted, caution is urged around its practice: for example, sensitivity to how stigma is deployed by the anti-tobacco movement and an appreciation of its psychological impact on smokers (Bayer and Stuber, Reference Bayer and Stuber2006; Burgess et al., Reference Burgess, Fu and van Ryn2009).

In both areas of debate, individuals are typically represented by their smoking status. Thus, the question of whether smoking and smokers are stigmatised, like the stigma-inducing potential of policy, is generally discussed as if smokers were similar to non-smokers in other respects. Yet, at the same time, it is acknowledged that smokers and non-smokers have other identities and are different and unequal in multiple ways. For example, many studies note that smokers are disproportionately drawn from disadvantaged groups. In consequence, ‘all the restrictive and burdensome public health measures designed to further limit the prevalence of tobacco consumption are borne by those at the bottom of the social ladder’ (Bayer, Reference Bayer2008: 470). With disadvantaged smokers vulnerable to the ‘dual stigmatisation’ of poverty and smoking (Thompson et al., Reference Thompson, Pearce and Barnett2007: 508), tobacco control interventions need to help disadvantaged smokers ‘translate feelings of stigma into productive rather than deleterious responses’ (Burgess et al., Reference Burgess, Fu and van Ryn2009: S156). Studies have also compared the experience of smoking-related stigma across socioeconomic groups. They conclude that disadvantaged smokers report greater (Farrimond and Joffe, Reference Farrimond and Joffe2006), less (Stuber et al., Reference Stuber, Galea and Link2008) and similar levels of smoking-related stigma (Ritchie et al., Reference Ritchie, Amos and Martin2010).

In these important ways, socioeconomic inequalities in smoking permeate the debate about smoking and stigma. However, there is little direct engagement with social class. Thus, while socioeconomic differentials in smoking are noted, there is little discussion of whether and how the dynamics of social class are implicated in the social denormalisation of smoking. In consequence, socioeconomic inequalities, both in people's circumstances and in their smoking behaviour, are not interrogated as ‘structural preconditions of stigma’ (Goffman, Reference Goffman1963: 11).

Social class, smoking and stigma

Like other foundational concepts in the social sciences, there is no one agreed definition of social class (Savage, Reference Savage2000). However, in broad terms, social class refers to enduring inequalities in people's lives, which are structurally and discursively produced and resisted (Breen and Rottman, Reference Breen and Rottman1994; Weis, Reference Weis2008). The centrality of such enduring inequalities to the production of smoking as a stigma is illustrated in this section in two ways: through the changing social profile of smokers and the increasing importance of the body and behaviour in public discourses about social class and moral worth.

Changing patterns of smoking

Evidence suggests that tobacco companies initially targeted their new product at dominant social groups. In the US for example, cigarettes conveyed the social distinction of affluent, white, heterosexual men (Brandt, Reference Brandt2007). The association with social prestige was also central to opening up a female market for the new tobacco produce: the popular media set women's smoking within the heterosexual courtship of the social elite (Tinkler, Reference Tinkler2001). The limited evidence indicates that these marketing strategies were successful (Brandt, Reference Brandt2007; Hilton, Reference Hilton2000). However, while the habit was taken up first by young adults in socially advantaged groups (Graham, Reference Graham2009a; Lopez et al., Reference Lopez, Collishaw and Piha1994), prevalence quickly rose among lower-income groups. For example by the 1940s in Britain, smoking rates were uniformly high across socioeconomic groups, both for men (over 60 per cent) and women (over 40 per cent) (Wald and Nicolaides-Bouman, Reference Wald and Nicolaides-Bouman1991).

The social cachet of cigarettes inevitably declined as the habit spread: social distinction cannot be claimed for a commodity bought and consumed on a mass scale. By the time that trend data were first available – in the 1940s in the UK – male prevalence was already declining in the higher socioeconomic groups (Wald and Nicolaides-Bouman, Reference Wald and Nicolaides-Bouman1991). Interestingly, this downward trend predated the scientific papers (Doll and Bradford Hill, Reference Doll and Bradford Hill1950) which warned of the health risks of smoking and the reports (e.g. RCP, 1962) which moved the issue into the public domain.

Socioeconomic differentials in smoking continued to widen in Britain through the 1970s, 1980s and 1990s (Figure 1), decades in which social attitudes shifted against smoking. In the US, too, the increasing social unacceptability of smoking coincided with its increasing concentration in disadvantaged groups (Figure 2). Similarly in Australia, ‘smoking is increasingly a badge of unemployment, low socioeconomic status and low educational achievement’ (Chapman and Freeman, Reference Chapman and Freeman2008: 27). As this suggests, smoking identities are inseparable from class identities; the negative labels which studies suggest are increasingly attached to smoking (‘a disgusting habit’) and to smokers (‘inconsiderate’, ‘weak-willed’, ‘dirty’) are inevitably class labels as well.

Figure 1.

Prevalence of cigarette smoking among men and women in the highest (professional) and lowest (unskilled manual) socio-economic groups, Britain, 1958–2000

Sources: Wald and Nicolaides-Bouman (Reference Wald and Nicolaides-Bouman1991, table 5.2), Office for National Statistics (2001, table 8.8).

Figure 2.

Prevalence of cigarette smoking among men and women in the highest and lowest educational groups, US, 1975–2005

Source: CDC (2010, table 62).

Drilling deeper into the class patterning of smoking, quantitative analyses suggest that smoking careers are shaped by persisting disadvantage. UK studies have documented how disadvantaged pathways, from poor childhood circumstances through early school leaving and into poor adult circumstances, are associated with an increased risk of smoking. For women, domestic trajectories also matter. Thus, early motherhood and single motherhood increase smoking risk over and above the effects of childhood disadvantage, educational disadvantage and poor adult circumstances (Graham et al., Reference Graham, Inskip, Francis and Harman2006; Graham et al., Reference Graham, Hawkins and Law2010). Among mothers on disadvantaged pathways – whose childhood circumstances were poor, who left school at the minimum permitted age, became a teenage mother and now lived on a low income – smoking rates approach 70 per cent (Graham, Reference Graham and Graham2009b).

These quantitative analyses suggest that being a smoker is a social marker of class-related disadvantage. Qualitative studies of the everyday lives of disadvantaged groups − of young people from poorer backgrounds, teenage mothers and welfare claimants, for example – make clear that they are framed both by material hardship and by a public discourse which undermines claims to moral worth (Charlesworth, Reference Charlesworth2000; Graham and McDermott, Reference Graham and McDermott2006; Seccombe et al., Reference Seccombe, James and Walters1998). As one example among many, a young mother summarised her experience of stigma by noting, ‘they look at you like you are a slag or something. No one smiles at you or anything, they just give you dirty looks’ (Letherby et al., Reference Letherby, Wilson, Bailey and Brown2001: 20).

Stereotypes of smokers draw on the repertoire of negative representations that young mothers, like other disadvantaged groups, experience and resist on a daily basis. As both quantitative and qualitative studies make clear, non-smokers align smokers with other stigmatised groups. In a UK study, non-smokers referred to ‘teenage smokers’, ‘smoking mothers’, ‘lower social class benefits kind of image’, ‘poor council house tenants’, ‘working class’ and ‘unemployed’ (Farrimond and Joffe, Reference Farrimond and Joffe2006: 487). An Australian study of the ‘spoiled identity’ of smokers similarly noted that smokers were seen as ‘undereducated and a social underclass’ (Chapman and Freeman, Reference Chapman and Freeman2008: 27). Such negative representations are actively challenged, with smokers asserting a moral identity in the face of representations which cast them as selfish and undeserving (Coxhead and Rhodes, Reference Coxhead and Rhodes2006; Thompson et al., Reference Thompson, Pearce and Barnett2007).