The ‘moral career’ of cigarette smokers: A French survey
PATRICK PERETTI-WATEL1,2,3, SANDRINE HALFEN4, &
ISABELLE GRE´MY4
1Health and Medical Research National Institute (INSERM), Research Unit 379, Social Sciences
Applied to Medical Innovation, Institut Paoli Calmettes, Marseilles, France, 2Southeastern Health
Regional Observatory (ORS PACA), Marseilles, France, 3French National Cancer Institute (INCA),
Paris, France, and 4Health Monitoring Centre of Paris Ile-de-France Region (ORSIF), Paris, France
Abstract
This paper aims to illustrate the relevance of Howard S. Becker’s sociological model of deviance for a
better understanding of contemporary adult smoking. From this perspective, one crucial aspect of
smoking is smokers’ ability to develop and entertain convincing rationalizations that help them to deny
smoking hazards and challenge anti-tobacco messages. Several hypotheses are derived from this model
and most of them are successfully tested with quantitative data from a cross-sectional survey conducted
in the Paris Ile-de-France Region. As expected, most smokers agreed that smoking damages health and
considered being well-informed about smoking hazards, even those who denied these hazards for
themselves. Moreover, smokers’ rationalizations were closely correlated to cigarette consumption and
duration since smoking initiation. Paradoxically, risk denial was also stronger among smokers who
have some characteristics usually considered as protective factors against smoking (especially future-
orientation and importance attached to one’s health). More generally, our sociological perspective
leads to consider smokers’ risk denial as the result of acquired cognitive skills instead of the
consequences of lack of information or psychological bias. We believe it provides a promising avenue
for further research.
Keywords: Cigarette smoking, risk denial, moral career, France
Introduction
Do smokers underestimate risks?
According to the World Health Organization (WHO), tobacco is the second major cause of
death in the world, with about 5 million deaths each year. Most developed countries have
joined the global war on smoking. Nevertheless, despite extensive anti-tobacco policies,
smoking prevalence is still high in these countries. For example, in most European countries
between a quarter and a third of adults are still smoking (WHO 2003). Socio-behavioural
scientists frequently explain the persistence of such high smoking prevalence rates as a result
of enduring underestimation of health-related consequences of smoking. However, studies
conducted to assess smokers’ perceptions of risk have produced confusing or inconsistent
results (Sutton 1999, Weinstein et al. 2005). Such studies usually ask smokers to
Correspondence: Patrick Peretti-Watel, ORS PACA Inserm U379, 23 rue Stanislas Torrents, 13006 Marseille, France. Tel: 33 4 96
10 28 61. Fax: 33 4 96 10 28 99. E-mail: peretti@marseille.inserm.fr
Health, Risk & Society,
September 2007; 9(3): 259 – 273
ISSN 1369-8575 print/ISSN 1469-8331 online  2007 Taylor & Francis
DOI: 10.1080/13698570701486070

numerically estimate their own risk of experiencing tobacco related illness, or to compare it
with that of someone else, either a non-smoker or an average smoker.
Studies using the numerical approach frequently found substantial over-estimation of
risks (Marsh and Matheson 1983, Viscusi 1990, Boney McCoy et al. 1992), but the results
could be very different depending on whether the risk was assessed using proportions or
percentages (Borland 1997), and whether respondents have to assess one risk (getting a lung
cancer) or concurrent risks (lung cancer, car accident, suicide, homicide, etc.) (Slovic
2000). Moreover, smokers also overestimate their life expectancy (Schoenbaum 1997).
Despite some inconsistencies related to data collection methods, studies using the
comparative approach are more reliable in showing that smokers are prone to believe that
they have a lower risk of developing a smoking-related disease than the average smoker
(Hansen and Malotte 1986, McKenna 1993, Weinstein et al. 2005). This finding is usually
interpreted as an ‘optimism bias’: people generally believe that their personal risk is less than
the risk faced by others in the same situation (Weinstein 1989).
Thus it is often claimed that smoking cannot be interpreted as a choice made in the
presence of adequate information about the potential harm, as smokers’ awareness and
understanding about such harm is quite superficial and not systematically related to a
perceived personal risk (Fox 2005, Kozlowski and Edwards 2005, Weinstein et al. 2005).
According to Chapman and Liberman (2005), there are four levels of information (having
heard that smoking increases health risks; being aware that specific diseases are caused by
smoking; accurately appreciating the meaning, severity, and probabilities of developing
smoking-related diseases; personally accepting that the risks inherent in other levels apply to
one’s own risk of contracting such diseases), and a smoker must have reached the fourth
level to be considered fully informed. From this point of view, the ‘optimism bias’
demonstrates a lack of information and most smokers, if not all, are not fully informed.
From misperception to risk denial: The ‘moral career’ of smokers
Beyond the (mis)perception of risk issue, some studies also highlighted the propensity of
smokers to endorse ‘smoking myths’ and self-exempting beliefs, for example ‘exercise
undoes most smoking effects,’ ‘cancer mostly strikes people with negative attitudes,’ ‘lots of
doctors and nurses smoke, so it cannot be all that harmful,’ etc. (Chapman et al. 1993,
Oakes et al. 2004, Weinstein et al. 2005). Some of these studies endorsed a social psychology
perspective based on the theory of ‘cognitive dissonance’ coined by Festinger (1957). This
theory is based on two assumptions: people need consistency between their behaviours and
their beliefs because inconsistency is disturbing and emotionally costly, and this consistency
is attainable by adjusting beliefs rather than changing behaviours. Within this theoretical
framework, observable beliefs are the result of a rationalization process intended to reduce
the dissonance. For example, smokers may entertain specific beliefs to dissipate the
cognitive dissonance created by widespread information about health-consequences of
smoking. From this perspective, smokers do not simply lack information about smoking
hazards, they rather actively entertain alternative beliefs to convince themselves that their
habit is not so risky.
The present paper proposes a similar perspective within a sociological model of deviance
initially developed by Howard Becker. According to Becker (1963), people who engage in a
behaviour labelled as deviant by the mainstream society have to neutralize this label with
convincing rationalizations, and these rationalizations will evolve all along the deviant’s
career (e.g., for a cannabis user: from initiation to regular use). Such neutralization process
shapes the deviant’s ‘moral career’ and they should be considered as a crucial dimension for
260 P. Peretti-Watel et al.