Enhancing the effectiveness of tobacco package warning
labels: a social psychological perspective
E J Strahan, K White, G T Fong, L R Fabrigar, M P Zanna, R Cameron
See end of article for
Tobacco Control 2002;11:183–190
Objective: To outline social psychological principles that could influence the psychosocial and behav-
ioural effects of tobacco warning labels, and to inform the development of more effective tobacco
Data sources: PsycInfo and Medline literature searches and expert guided selection of principles and
theories in social psychology and of tobacco warning labels, including articles, books, and reports.
Conclusions: Tobacco warning labels represent a potentially effective method of influencing attitudes
and behaviours. This review describes social psychological principles that could be used to guide the
creation of more effective warning labels. The potential value of incorporating warning labels into a
broader public health education campaign is discussed, and directions for future research are
smoking. By 1991, 77 countries required health warnings on
their tobacco products,1although the nature of those health
warnings varies considerably across countries. The introduc-
tion of Canada’s new graphic warning labels in December
2000 has prompted other countries to review their require-
ments and may lead to new warning labels in many of those
countries. Hence, it seems timely to consider methods that
could be employed to enhance the effectiveness of tobacco
This paper will outline various social psychological princi-
ples that could be used to guide the creation of effective
tobacco warning labels. In addition, we will discuss the need
to integrate tobacco warning labels with other anti-smoking
efforts (for example,point-of-sale ads,television commercials,
etc).Finally,this paper will outline the value of future system-
atic research in the area of tobacco warning labels. We are
aware that policy makers are often under constraints as to
what types of messages they can implement (for example,
some countries may allow only health oriented messages on
their warning labels, as is currently the case in Canada under
the 1997 Tobacco Act). We conceptualise the psychological
principles outlined here as being suggestions that policy mak-
ers can selectively choose from depending on their particular
constraints (for example, whether the warning messages
must be health related), or their goals (for example, whether
they are targeting a specific demographic group).
We began this conceptual review by using our expertise to
identify the principles and research domains from social psy-
chology that we felt were applicable to understanding the
possible effects of tobacco warning labels. We reviewed the
social psychological literature on persuasion,fear appeals,and
various social psychological theories of behaviour change (for
example, theory of reasoned action/planned behaviour). We
then conducted an extensive review of the research literature
on warning labels, which included a review of all the sources
cited in a literature review by Mitchell,2as well as other
sources, to identify articles, chapters, reports and other mate-
rial that were directly or indirectly relevant to the social
psychological literature. We did not find any articles that cast
ver the past 35 years, warning labels have become a
popular method by which governments attempt to
inform their citizens of the health consequences of
their findings in terms of the social psychological principles
that we outline in this paper. Furthermore, a search of
Medline and PsycInfo databases for “social psychology” and
“warning labels” yielded no items. Thus, the articles and
reports cited in this conceptual review article came from an
expert guided search of the social psychological literature.
SOCIAL PSYCHOLOGICAL PRINCIPLES FOR
ENHANCING THE EFFECTIVENESS OF WARNING
Over the past 50 years, a vast body of social psychological
research regarding the processes and principles that affect
attitude and behaviour change has accumulated.3–5A number
of basic principles for enhancing the effectiveness of warning
labels can be derived from this research. These principles can
be divided into two categories. Content principles refer to fea-
tures of the message content itself that make the message
more effective. Process principles refer to design features (for
example, stylistic features) that make the message more
effective. Both categories of principles can be used to inform
the design of warning labels.
Currently, warning labels that appear on tobacco packages
typically consist of simple statements about the health risks of
smoking (for example, “Smoking can kill you”, “Smoking
causes lung cancer”). Although pointing out the health risks
of smoking may be beneficial, it is clear from the research on
attitudes and persuasion that the sole focus on negative health
risks may be too narrow.6The social psychological research on
attitudes and persuasion suggests several general content
principles that could guide the creation of new warning labels.
Promote attitudes and beliefs towards alternative
The core principle here is quite straightforward: messages are
more likely to be persuasive if they not only promote negative
attitudes toward an undesired behaviour (for example, smok-
ing), but also promote positive attitudes toward a mutually
exclusive desired behaviour (for example, quitting smoking).
Geoffrey T Fong,
Department of Psychology,
University of Waterloo,
200 University Avenue
West, Waterloo, Ontario,
N2L 3G1, Canada;
Received 17 July 2001
and revision requested
8 October 2001.
Accepted 1 March 2002
Research indicates that a person’s attitude toward smoking is
not the exact opposite of his/her attitude toward quitting
toward smoking and might believe it is unhealthy and danger-
ous to smoke, but if this person’s attitude and beliefs about
quitting smoking are even more negative, they will continue to
This research suggests that warning labels would be more
effective if they created a strong positive attitude toward quit-
ting, in addition to promoting a strong negative attitude
towards continued smoking. Incorporating positive, factual
messages about the benefits of quitting could represent a sim-
ple, but important, improvement. For example, some labels
could stress the immediate health benefits of quitting. One of
the inside messages introduced as part of the new Canadian
warning labels states that “Quitting smoking reduces your
chance of having a heart attack”. Other labels could stress the
financial benefits of quitting (for example, “If you smoke two
packs a day, quitting will save you over $3500 in the next
Utilise gain-framed messages
The importance of incorporating positive information into
smoking labels is further supported by research on gain-
framed versus loss-framed messages.8 9Health messages can
focus on the negative consequences of continuing to engage in
a health compromising behaviour (loss-framed messages) or
they can focus on the positive consequences of refraining from
a health compromising behaviour (gain-framed messages).
Warning labels, as well as most anti-smoking campaigns,
focus nearly exclusively on loss-framed messages.10 11
Past research suggests that whether loss or gain-framed
messages are more effective depends on various factors, such
as whether respondents are motivated to engage in effortful
processing12and level of self-efficacy.13More recent evidence
indicates that loss-framed messages appear to be successful in
promoting behaviours related to early detection, such as clini-
examination,13and HIV testing.16In contrast, gain-framed
messages have been found to influence prevention related
behaviours such as sunscreen use,17preference for surgical
procedures,18–21and engaging in regular physical exercise.22
Because quitting smoking is a preventive health behaviour,
these studies suggest that messages designed to encourage
smokers to quit might be more effective if they were framed in
terms of gains rather than losses.
One recent study directly addresses the effects of gain-
framed versus loss-framed messages in smoking. Schneider
and colleagues11found that gain-framed health messages (for
example, “Quitting smoking reduces your chances of prema-
ture death and illness”) were more effective than the same
messages rewritten so that they were loss-framed (for exam-
ple, “Smoking increases your chances of premature death and
illness”) in changing smoking related beliefs, attitudes, and
It should be noted that this research on framing cannot be
used to conclude that gain-framed messages are effective and
loss-framed messages are not. Rather, it suggests that
gain-framed messages may be more effective in certain
settings than loss-framed messages.A second qualifying com-
ment is warranted here. The research on framing does not
conclude that gain-framed messages are superior to loss-
framed messages when the gain-framed message covers a dif-
ferent topic than a loss-framed message. For example, it can-
not predict that a gain-framed message such as “you’ll
experience greater self-esteem if you quit smoking” will be
more effective than a loss-framed message such as “if you
keep smoking, you will be many times more likely to get lung
cancer” because of differences in content. What the research
does suggest is that even health messages on warning labels,
which are currently exclusively framed in terms of losses,
might be more effective if some of the messages were
reframed in terms of gains. In short, it may be effective to use
multiple messages that focus on the both the benefits of quit-
ting and the costs of smoking.
There is one class of loss-framed messages that is of
particular relevance for warning labels. These are those that
involve fear appeals. For about 50 years, researchers and prac-
titioners alike have conducted research demonstrating the
effectiveness of fear appeals in influencing health relevant
attitudes and behaviour.6 23–26Out of the literature on fear
appeals has emerged some general conclusions about the con-
ditions under which fear appeals are most likely to be
effective.In a recent meta-analysis and literature review of the
research on fear appeals, Witte and Allen26drew several
conclusions, including two that are most relevant here: (1)
fear appeals can be effective in increasing healthy behaviour
and decreasing unhealthy behaviour; and (2) fear appeals are
effective to the extent that they are accompanied by efficacy
messages—that is, messages that provide information about
how to avoid the threat that is highlighted by the fear appeal.
Thus, loss-framed messages used on warning labels are
more likely to be effective if: (a) they are combined with gain-
framed messages (emphasising that quitting is possible and
beneficial); and (b) they advise the reader how to quit, or
where to get help. The new Canadian warning labels are con-
sistent with these recommendations: the new warning labels
consist of an outside message accompanied by a vivid photo-
graph that may invoke fear (fig 1), and 16 rotated inside mes-
sages (fig 2), many of which provide a strong efficacy
message—that is, information about actions one can take to
avoid the health threat depicted on the outside label. Further
research should clarify the ways in which combinations of
both loss-framed and gain-framed messages influence smok-
ing attitudes, beliefs, and behaviours.
Emphasise subjective norms
Subjective norms and social approval have a strong influence
on health behaviour. We are more likely to perform a
behaviour if we believe that the behaviour is valued or
expected within our reference group.We tend to accommodate
the expectations of important people in our lives.27This prin-
ciple implies that smokers who are contemplating quitting
may find the impulse to quit counteracted by perceptions that
smoking is the norm within their reference group.
There is evidence that subjective norms predict intentions
to perform health behaviours.28–32For example, Finlay and
colleagues30found that subjective norms predicted behav-
ioural intentions across a wide range of health behaviours,
such as avoiding risky sexual behaviour, paying attention to
health related advice, and taking medication as prescribed.
Most importantly, subjective norms have been found to
predict smoking intentions and behaviour.33 34
* Calculation based on cigarette packages costing $5.00.
Canadian package warning label that may invoke fear.
184Strahan, White, Fong, et al
The research on subjective norms suggests that warning
labels might be more effective if they conveyed credible mes-
sages indicating that significant referent groups (for example,
children, friends, physicians) were strongly in support of ces-
sation. Polling data could be used to provide support for sub-
jective norm based appeals. For instance, one could determine
the percentage of family members and friends that want their
loved ones to quit smoking and then incorporate this
information into a message (for example, “85% of family
members surveyed want their loved ones to quit smoking”).
In a related vein, labels might be more effective if they
highlighted the negative social consequences of smoking (for
example,“Smoking causes bad breath and yellow teeth”).van
der Plight and de Vries35found that both smokers and
non-smokers believed that smoking is bad for one’s health,
but only smokers believed that smoking helps people relax
and, most pertinent here, that smoking fosters social interac-
tion. Labels could present the negative social consequences of
smoking to counteract smokers’ perceptions that smoking
facilitates social interactions.
This point might be most relevant in considering the impact
of current warning labels on adolescents.The narrow focus on
the health risks associated with smoking may not be optimally
effective with adolescents. Adolescents are less likely than
adults to value their health, and as a result, the messages may
seem irrelevant to them. Furthermore, research has high-
lighted the role that social and self presentational reasons play
in adolescent smoking.36 37Therefore, warning labels that
focus on the negative social consequences of smoking may be
more effective among adolescents than those that focus on the
negative health consequences.
Focus on relevant attitudes of the target group
Different messages are relevant to different segments of the
population (for example, those who smoke v those at risk of
initiating smoking; adolescents v adults). Recent research
indicates a person’s three most important smoking related
beliefs are better predictors of smoking behaviour than a com-
posite of one’s general smoking beliefs.38Research that identi-
fies important beliefs and values relevant to smoking and
quitting held by target groups would, as social marketing
principles suggest, be a valuable first step in designing
messages of greater impact.
The social psychological literature points out that it is
important not only to identify which attitudes and beliefs are
important for a given group, but also the underlying reasons
for those attitudes and beliefs.38Research in this area focuses
on the important question: what are the underlying functions
for holding particular attitudes? Attitudes fulfil various func-
tions,and understanding the function of the attitude provides
insight into the types of messages likely to have impact.39 40
The concept of attitude function has been studied in relation
to attitudes towards victims of AIDS,41attitudes towards
homosexuals,42attitudes towards automobiles,43 44and atti-
tudes towards advertising of consumer products.45–52
As applied to smoking, this suggests that it is not only use-
ful to know that the target group has positive attitudes
towards smoking, but also to be aware of why that group has
positive attitudes towards smoking. An adolescent may hold
with her peers, to keep her body weight low, or because she
enjoys the physiological effects of smoking. Knowledge of a
particular group’s attitude functions can be beneficial in the
creation of appropriate and effective smoking labels. For
example, it might be effective to create different labels for dif-
ferent brands, in much the same way as tobacco companies
create different brand images for different brands (for exam-
the package of a brand that is marketed to young males (fig
3)).This suggestion is similar in ways to the process of market
segmentation that is prevalent in any marketing endeavour.
This segmentation process may be profitably guided by apply-
ing the findings in the social psychological literature to warn-
ing labels (for example, in the application of fear appeals53).
Finally, it seems likely that rotating multiple warning label
messages that address these different attitude functions will
be an effective strategy.
Increase perceived self efficacy
People are more likely to attempt to change their behaviour
(for example, quit smoking) if they believe they can
succeed—that is, if they have a high level of self efficacy54or
perceived behavioural control.29 55Those with higher levels of
self efficacy are more likely to alter successfully a wide range
of health behaviours, including AIDS risk behaviours,56
sumption,60 61and dietary intake.62
Smokers who are high in self efficacy have a greater chance
of entering treatment to quit smoking and have a greater
chance of success than those who are low in self efficacy.63
Bandura64has described several strategies for enhancing self
efficacy. The effectiveness of warning labels may be consider-
ably enhanced by reinforcing people’s beliefs that they are
capable of quitting smoking. Labels could provide some
general statements about quitting efficacy (for example, “You
can quit smoking and reduce your risk of lung cancer”) or
specific information about the process of quitting that would
be encouraging (for example, “Smokers who quit tended to
try a number of times before they succeeded, so keep
trying!”). These kinds of messages would also serve to
enhance outcome efficacy (or behavioural beliefs, in the
theory of planned behaviour55).
Another promising avenue for enhancing self efficacy is to
provide information about quitting. The importance of
providing specific information about quitting is also high-
lighted by the classic literature on fear appeals in social
psychology. As mentioned earlier, this research suggests that
Canadian package warning label with strong efficacy
Canadian package warning label directed at young
Tobacco package warning labels 185
messages stressing the negative health consequences of
smoking create fear among smokers, and that such messages
may be more effective when presented in tandem with recom-
mendations for how the negative consequences can be
avoided.6 26Thus, labels that provide information about
quitting could enhance the effectiveness of not only the
current set of warning labels that focus exclusively on
negative health consequences, but also future warning labels
that would focus on enhancing people’s efficacy to quit. For
example, the new Canadian warning labels include quit tips,
efficacy messages, and a website for obtaining additional
information about methods of quitting.
Additional content principles
One other factor that appears to be important is whether or
not the anti-smoking label encourages interpersonal commu-
nication. For example, one recent study found that discussing
a smoking campaign with someone else was predictive of
positive behavioural outcomes among smokers.65Smoking
labels might profitably encourage people to talk to others
about smoking.A second additional factor that is important is
the ability to induce cognitive dissonance. Cognitive disso-
nance is an unpleasant state that is aroused when an
individual has an attitude that is discrepant to a behaviour or
to a new piece of information.66For example, an anti-smoking
label that makes salient to smokers that their beliefs (for
example,smoking is bad for your health) and their behaviours
(for example, smoking two packs of cigarettes a day) are con-
tradictory may be an effective means of reducing smoking.
Research has found that a hypocrisy manipulation which
makes people aware of their dissonant beliefs can be an effec-
tive means of changing health related behaviours.67This
research suggests that warning labels could be designed to
serve as a hypocrisy manipulation. For example, if smokers
believe that children should not smoke and they have told
their own children not to smoke,warning labels could remind
smokers that their own smoking behaviour is a major cause of
their own children’s smoking (fig 4). This should make
parents feel guilty about smoking and could ultimately lead to
a decrease in smoking behaviour. In addition, labels that ask
people to commit to quitting may create the need to be
consistent with that commitment and may encourage
Whereas content principles are concerned with how the con-
tent of the labels might be changed to improve their effective-
ness, process principles are concerned with the manner in
which a message is presented.In many countries,the warning
label is presented in a visually indistinct way.There are several
ways that the presentation could be enhanced to make warn-
ing labels more effective.
One set of relevant process principles can be derived from a
social psychological theory known as the elaboration likeli-
hood model.5A similar theoretical framework, the heuristic-
systematic model,has been formulated by Chaiken69 70and has
also been applied to understanding the cognitive and social
psychological principles underlying the effects of warning
According to the elaboration likelihood model, people are
sometimes persuaded as a result of thinking very carefully
about the content of a message and on other occasions by
considering factors that have little to do with the content of a
message. When a person is both motivated and able to think
carefully about a message, that person engages in extensive
elaboration of the message. This process involves carefully
scrutinising the merits of the arguments. Thus, a person may
assess the validity of the arguments in light of what they
already know,form inferences that go beyond the information
presented, and perhaps even seek out additional information.
When persuasion occurs as a result of extensive elaboration of
a message, they are taking the so called central route to
persuasion, and the persuasiveness of the message will
depend on the strength of the arguments.
On other occasions, people may lack the motivation (for
example, they may not see the topic as personally relevant)
and/or ability (for example, they may lack sufficient knowl-
edge about the topic or may have distractions in their environ-
ment) to elaborate carefully on a message. In such cases, they
will not consider the strength of the arguments, but instead
look for simple characteristics of the message or the context in
which it is encountered to provide them with a basis for
determining whether they should accept the message. For
example, they may use the attractiveness or credibility of the
persuader as a basis for deciding to accept the message. In so
doing, they are taking the peripheral route to persuasion, and
the persuasiveness of the message will depend not on message
content, but on other features of the message, or the message
context—and such non-central features can be just as persua-
People are exposed to warning labels under many different
circumstances,some of which favour elaboration,and some of
which do not. Hence, it would be wise to build features into
warning labels that influence people under a wide variety of
different levels of elaboration.
Colour is one way to capitalise on process principles. For
instance, consider the impact of presenting warning labels in
bright orange print. Orange is associated with warning and
danger signs: under conditions of low elaboration, orange in
labels may convey a “warning” message, independent of con-
tent, because of this sort of association. Under conditions in
which people are able to examine the content of the label, but
perhaps are only moderately motivated to do so, the bright
colour may attract attention,and prompt them to consider the
message content. Finally, with those motivated to read the
message carefully, the association of orange with danger
might evoke feelings congruent with the message content and
thus increase message acceptability and impact.
Iconic symbolic images may also be useful for conveying
persuasive messages. If a widely recognised athlete, who sym-
bolises health and vigour, appeared on tobacco package warn-
ing labels, endorsing the anti-smoking message, this might
increase the impact of the message in several ways. Under
conditions of low elaboration,the image would be a peripheral
cue signalling the health message independent of content.
People would understand and accept the message because
they like and trust the iconic figure.Under conditions of mod-
erate elaboration, people might be more willing to read and
reflect upon what this popular spokesperson has to say.Under
conditions of high elaboration, positive feelings toward the
spokesperson could enhance the likelihood that people would
agree with the message.
Overexposure, or wear-out, is a major problem for any mes-
sage that is presented many times over a period of time.When
people receive a message multiple times, the message’s effec-
tiveness tends to increase over the first few exposures, but
then begins to diminish over time.72A survey conducted in the
their own smoking behaviour.
Canadian package warning label warning parents about
186 Strahan, White, Fong, et al
summer of 1999 revealed that among over 2000 Canadian
adults and 746 Canadian youth aged 12–18 years,65% of adult
smokers and 74% of youth smokers agreed that the warning
labels that were introduced in 1994 were “worn-out and had
lost their effectiveness”.73Other studies found that individuals
exposed to newly designed warnings were significantly more
likely to remember the concept of the warnings than
individuals exposed to old mandated warnings.74 75Our review
of the research on gain and loss-framed messages and on the
functions of attitudes suggests that a variety of anti-smoking
messages would be effective, and such a strategy would also
address this overexposure problem.
Advertising campaigns have developed strategies to coun-
teract the overexposure problem. These include changing ads
and commercials by employing different variations of the
same theme, promoting the products via different spokesper-
sons, or even following people as they progress through a
series of decisions or judgments involving the product. Policy
makers have suggested that rotating smoking label messages
may counteract the overexposure problem.76Designing warn-
ing labels in various appropriate colours and broadening the
content of the messages may help to counteract overexposure.
Any method that reduces overexposure increases the likeli-
hood that people will read the warning labels, and attracting
people’s attention is the first step in the influence process.
INCORPORATING WARNING LABELS INTO A
BROADER ANTI-SMOKING CAMPAIGN
There are limits to what warning labels can be expected to
accomplish in isolation. Warning labels are very brief
messages, considerably shorter than the typical persuasive
communication, and thus there are limitations to the amount
of information that they can convey. Furthermore, policy con-
straints limit the type of information that can appear on
smoking warning labels.
Warning labels might be considerably more effective if ref-
erences were made to the labels as part of a broader,
coordinated anti-smoking campaign. This more extensive
anti-smoking campaign would not be subject to the policy
constraints that the warning labels are subject to, and would
thus be better positioned to employ the social psychological
principles outlined above.
A multimedia campaign using many avenues for communi-
cation has the potential to persuade people using both elabo-
rated and unelaborated channels of communication. For
instance, catchy, colourful posters and labels could be used to
persuade via the peripheral route, whereas more informative
print ads and television commercials could be used to
persuade people via the central route.Thus,people at all levels
of elaboration could be reached if such an anti-smoking cam-
paign was employed.
Furthermore, warning labels on cigarette packages could
serve as retrieval cues to remind people of, or reinforce, more
elaborate anti-smoking messages that they would be receiving
from other sources.77Warning labels would be more effective if
they were specifically designed to remind people of anti-
example, television commercials, magazine ads, billboards,
etc), point-of-sale displays, and school based programmes. It
has been suggested that health campaigns work better when
they use multiple media (for example, print, radio, and televi-
sion) and when a common message is repeated.78If warning
labels were coordinated with broader, more extensive cam-
paigns, they would likely have a greater impact on the public.
Such broader anti-smoking campaigns, because of the greater
space or time available, could also take fuller advantage of the
principles of influencing alternative behaviours to smoking,
utilising gain-framed messages, emphasising
norms, focusing on the relevant attitudes and attitude
functions of the target group, and enhancing self efficacy.
As new social psychological principles are identified, these
might also be incorporated into broader anti-smoking
campaigns. For example, Cialdini has recently formulated a
persuasive technique to encourage people to resist an
argument that is deceptive or duplicitous.79In his research,
Cialdini finds that a persuasive counter-argument that uses
aspects of the original advertisement and then counter-argues
against it, with an element of ridicule, is most effective. Cial-
dini has dubbed this technique as the “poison parasite”
because it contains two elements, one poisonous (the counter
argument) and one parasitic (the mnemonic link to the origi-
nal argument). An excellent example of this are the “Joe
Chemo” ads which directly ridicule and counter-argue the
“Joe Camel” ads.† Although such techniques would admit-
tedly be very difficult to incorporate directly onto tobacco
packaging, they could be used as part of a ongoing multime-
dia anti-smoking campaign—cues of which could be con-
tained within a warning label.
RESEARCH TO GUIDE LABEL DEVELOPMENT
In this paper we have outlined some basic principles from
social psychology that could enhance the effectiveness of
tobacco warning labels (see fig 5 for a summary of these prin-
ciples).In some cases,we have offered ways that these specific
principles could be operationalised. However, it is clear to us
that regardless of the apparent applicability of these social
psychological principles, there must be a plan for testing the
actual effectiveness of these principles, and for that matter,
any other principles that may be relevant in creating a more
effective tobacco warning label campaign.We agree with other
reviewers of the current research on tobacco warning labels
that to be effective,warnings must be continuously developed,
tested, targeted, monitored, and revised over time.80
To date,research on tobacco warning labels has been some-
what limited (see review by Mitchell2). There have been stud-
ies assessing awareness or recall of existing warning labels,81
believability of messages,82and subjective impressions of the
effectiveness of existing, new, or proposed warning labels.83
But the correlational design of many of the existing studies
leads to problems of interpretation.
In one of the best designed study of its kind to date,
Borland84 85conducted a longitudinal study in which smokers
were surveyed by phone before and six months after new,
larger, and enhanced warning labels were introduced in Aus-
tralia in 1995. In both cross sectional and longitudinal
samples, smokers contacted after the new enhanced warning
labels had been introduced provided survey responses that
were consistent with the notion that the new warning labels
had some beneficial effects, including greater likelihood of
noticing the health warnings and refraining from smoking on
at least one occasion.
There is a need for more powerful research methods to sup-
plement the existing correlational and focus group research
and to guide the development of the next generation of labels.
We envision a research programme that would employ multi-
ple research methods and contexts, including a combination
of laboratory experiments and field studies.
Experimental research would provide a more rigorous test
of the possible effectiveness of smoking labels in changing
smoking attitudes, beliefs, and behaviours. For example,
although the success of mass media anti-smoking campaigns
has been well documented,86–88experimental research would
allow researchers and tobacco control policy experts alike to
isolate the key factors that make such campaigns successful.
In addition, although research on message framing has been
promising, further research could determine whether it is a
†For the Joe Chemo ads, see http://www.joechemo.org
Tobacco package warning labels187
particular type of gain frame that is most persuasive, whether
a combination of messages might be most effective, or
whether it is the novelty of the message that is important. To
get a more realistic picture of the effects of tobacco warning
labels, such controlled experiments should involve longer
exposure to the labels.We recognise that pragmatic considera-
tions may preclude a full test marketing of potential warning
labels per se, but we can envision research that tests the effec-
tiveness of messages that would eventually appear on warning
themselves—for example, as part of a health education
programme. Although not perfect, such research would
provide much needed preliminary data that might help policy
makers formulate warning labels (and health education
programmes more generally) with greater effectiveness.
As an example of this, before the introduction of the new
Canadian warning labels,Health Canada conducted an exten-
sive set of research studies, including experimental studies of
the effects of changing certain features of the warning labels
(for example, size, presence of photographs). This research
provided a powerful empirical base that supported the
effectiveness of the new labels and facilitated their implemen-
Because experimental methods are limited in ecological or
external validity, field research could examine the efficacy of
tobacco warning labels in the real world. Although experi-
mental techniques are not possible when large scale label
campaigns are implemented, natural experiments that utilise
a quasi-experimental designs could be employed. One
example is the research by Fong and his colleagues
(unpublished data) in their quasi-experimental, longitudinal
survey of 12 000 high school students in Canada and the USA,
with survey waves both before and after the introduction of
the new Canadian warning labels. Such quasi-experimental
than thewarning labels
research could then be further complemented by more
experimental research.That is,if Canada’s new warning labels
do appear effective in increasing smoking cessation, further
experimental research might be utilised to determine more
precisely what aspects of the labels are most important (for
example, their graphic nature, their novelty, their ability to
elicit negative reactions, etc).
Furthermore, social psychological research indicates the
benefits of assessing the relevant psychosocial mediators of
smoking behaviour—for example, attitudes and intentions
toward smoking. When measuring smoking attitudes and
intentions, researchers could employ more subtle techniques,
because they would reduce the problem of demand character-
istics. For example, measures of attitudes and intentions that
are embedded within other measures, measures that tap
affective responses, and measures that assess consequential
behavioural reactions to smokers might be employed.89
Furthermore,implicit measures of attitudes (that is,the auto-
matic evaluation associations people have to attitude objects)
such as adaptations of the Implicit Association Test90 91could
be a promising avenue for indirectly tapping into smoking
attitudes. Finally, it is important to use theory in guiding
research. In this article, we have outlined various principles
based on social psychological theory that may aid in the crea-
tion of effective smoking labels.
The value of more rigorous research that would evaluate
possible public health policies in tobacco control is highlighted
by the fact that tobacco companies have spent decades and
many millions of dollars creating advertising campaigns that
have been found to be extraordinarily effective in creating,
maintaining,and expanding their market.They have intimate
knowledge about ways to make a product that is inherently
unattractive, unappealing, and dangerous to the health of
oneself and others into a symbol that is attractive, sexy,
appealing, sophisticated, and even associated with healthy
activities (for example, Virginia Slims tennis). Cigarette
advertising companies developed these techniques through
careful research, much of which has utilised the findings and
theoretical perspectives of social psychology. Tobacco market-
ing and advertising campaigns that the public eventually sees
labels, these labelling strategies may be most effective when supported by a broader anti-smoking campaign that reinforces, elaborates upon,
and reminds consumers of these images.
Steps in creating effective smoking labels: a checklist. Although this checklist contains strategies for the creation of effective smoking
‡These reports are available on the Health Canada website at
188Strahan, White, Fong, et al
are the culmination of a long process of experimental
research, pilot studies, and extensive field studies. Although
tobacco control campaigns seldom have the kind of resources
available to them that cigarette advertising campaigns have,it
is important to understand that certain aspects of the process
of creating an effective advertising campaign of any kind are
essential. And the most central aspect of that process is to
conduct high quality research to test the effectiveness of pos-
sible methods. Social psychological principles—particularly
those that are embedded in models of communication—have
been shown to be effective in guiding the creation of effective
mass media campaigns in tobacco control,92 93and have the
potential to guide the creation of more effective warning
In closing, there are a number of challenges in designing a
more effective set of tobacco warning labels. This article is a
first step toward this end.Over the next few years,as we enter
a promising time of new opportunities for developing public
health campaigns that intend to educate and inform people
about the dangers of tobacco use, future campaigns, such as
those that will result in the creation of new tobacco warning
labels,could be profitably shaped by social psychological prin-
ciples and informed by research.
The preparation of this article was supported by the Centre for
Behavioural Research and Program Evaluation of the Cana-
dian Cancer Society/National Cancer Institute of Canada, the
Canadian Tobacco Control Research Initiative/Social Science
and Humanities Research Council of Canada, the Lyle S Hall-
man Institute Fund, and the National Cancer Institute of the
United States. The authors gratefully acknowledge the contri-
butions of Elizabeth Leal,Tara Elton,Jessica MacKay,Karolina
Pajda, and Melissa Kuntz. We would also like to thank two
anonymous reviewers for their insightful comments on an
earlier draft of this manuscript.
E J Strahan, G T Fong, M P Zanna, Department of Psychology,
University of Waterloo, Waterloo, Ontario, Canada
K White, Department of Psychology, University of British Columbia,
Vancouver, BC, Canada
L R Fabrigar, Department of Psychology, Queen’s University, Kingston,
*R Cameron, Department of Health Studies and Gerontology, University
*Also the Centre for Behavioural Research and Program Evaluation,
Canadian Cancer Society/National Cancer Institute of Canada
1 The World Bank. Curbing the epidemic: governments and the
economics of tobacco control. Washington DC: The World Bank, 1999.
2 Mitchell BM. Review of documents on tobacco package warnings and
toxic constituent labeling. Report prepared for the Center for Behavioural
Research and Program Evaluation, National Cancer Institute of Canada.
University of Waterloo, Waterloo, Ontario, Canada: Brenda Mitchell
and Associates, 1999.
3 Cialdini RB. Influence: science and practice. 3rd ed. New York: Harper
4 Eagly AH, Chaiken S. The psychology of attitudes. Fort Worth, Texas:
Harcourt Brace Jovanovich, 1993.
5 Petty RE, Cacioppo JT. The elaboration likelihood model of persuasion.
In: Berkowitz L, ed. Advances in experimental social psychology.
Hillsdale, New Jersey: Erlbaum, 1986:123–205.
6 Leventhal H. Findings and theory in the study of fear communications.
Adv Exp Soc Psychol 1970;5:119–86.
7 Fishbein M. A theory of reasoned action: some applications and
implications. In Howe HE, Page MM, eds. Nebraska symposium on
motivation. Lincoln, Nebraska: University of Nebraska Press,
8 Kahneman D, Tversky A. Prospect theory: an analysis of decisions
under risk. Econometrica 1979;47:263–91.
9 Tversky A, Kahneman D. The framing of decisions and the psychology
of choice. Science 1981;211:453–8.
10 Rothman AJ, Martino SC, Bedall BT, et al. The systematic influence of
gain- and loss-framed messages on interest in and use of different types
of health behavior. Pers Soc Psychol Bull 1999;25:1355–69.
11 Schneider TR, Salovey P, Pallonen U, et al. Visual and auditory message
framing effects on tobacco smoking. J Appl Soc Psychol (in press).
12 Maheswaran D, Meyers-Levy J. The influence of message framing and
issue involvement. J Marketing Res 1990;27:367.
13 Meyerowitz BE, Chaiken S. The effect of message framing on breast
self-examination attitudes, intentions, and behavior. J Pers Soc Psychol
14 Rothman AJ, Salovey P, Antone C, et al. The influence of message
framing on health behaviour. J Exp Soc Psychol 1993;29:408–33.
15 Banks SM, Salovey P, Greener S, et al. The effects of message framing
on mammography utilization. Health Psychol 1995;14:178–84.
16 Kalichman SC, Coley B. Context framing to enhance
HIV-antibody-testing messages targeted to African American women.
Health Psychol 1995;14:247–54.
17 Detweiler JB, Bedell BT, Salovey P, et al. Message framing and
sunscreen use: gain-framed messages motivate beach-goers. Health
18 Levin IP, Schnittjer SK, Thee SL. Information framing effects in social and
personal decisions. J Exp Soc Psychol 1998;24:520–9.
19 Marteau TM. Framing information: its influence upon decisions of
doctors and patients. Br J Soc Psychol 1989;28:89–94.
20 McNeil BJ, Pauker, SG, Sox HC, et al. On the elicitation of preferences
for alternative therapies. N Engl J Med 1982;306:1259–62.
21 Wilson DK, Kaplan RM, Schneiderman L. Framing of selections of
alternatives in health care. Soc Behav 1987;2:51–9.
22 Robberson MR, Rogers RW. Beyond fear appeals: negative and positive
persuasive appeals to health and self-esteem. J Appl Soc Psychol
23 Janis IL. Effects of fear arousal on attitude change: recent developments
in theory and experimental research. In: Berkowitz L, ed. Advances in
experimental social psychology, vol 3. New York: Academic Press,
24 Mongeau P. Another look at fear arousing messages. In: Allen M, Preiss
R, eds. Persuasion: advances through meta-analysis. Cresskill, New
Jersey: Hampton Press, 1998:53–68.
25 Sutton SR. Fear-arousing communication: a critical examination of theory
and research. In: Eiser JR, ed. Social psychology and behavioral
medicine. New York: Wiley, 1982:303–37.
26 Witte K, Allen M. A meta-analysis of fear appeals: implications for
effective public health campaigns. Health Educ Behav
27 Fishbein M, Azjen I. Belief, attitude, intention and behavior: an
introduction to theory and research. Reading, Massachusetts:
28 Ajzen I, Fishbein M. Understanding attitudes and predicting social
behavior. Englewood Cliffs, New Jersey: Prentice-Hall, 1980.
29 Ajzen I, Madden JT. Prediction of goal-directed behavior: Attitudes,
intentions, and perceived behavioral control. J Exp Psychol
30 Finlay KA, Trafimow D, Jones D. Predicting health behaviors from
attitudes and subjective norms: between-subjects and within-subjects
analyses. J Appl Soc Psychol 1997;27:2015–31.
31 Gottlieb NH, Gingiss PL, Weinstein RP. Attitudes, subjective norms and
models of use for smokeless tobacco among college athletes: implications
for prevention and cessation programming. Health Educ Res
32 Marcoux BC, Shope JT. Application of the theory of planned behavior to
adolescent use and misuse of alcohol. Health Educ Res
33 O’Callaghan FV, Callan VJ, Baglioni A. Cigarette use by adolescents:
attitude-behavior relationships. Subst Use Misuse 1999;34:455–68.
34 de Vries H, Dijkstra M, Kuhlman P. Self-efficacy: the third factor besides
attitude and subjective norm as a predictor of behavioral intentions.
Health Educ Res 1988;3:273–82.
35 van der Plight J, de Vries N. Belief importance in expectancy-value
models of attitudes. J Appl Soc Psychol 1998;28:1139-1354.
36 Biglan A, McConnell S, Severson, HH, et al. A situational analysis of
adolescent smoking. J Behav Med 1984;7:109–14.
What this paper adds
Tobacco warning labels represent a potentially effective
method of communicating the health hazards of smoking;
moreover, they represent an opportunity for motivating
reductions in smoking behaviour. Theory and research in
social psychology offers a set of principles that could be
used to create more effective warning labels, but to date
there has been no review of those relevant social psycho-
logical principles as they pertain to tobacco warning
This article represents the first such review. We provide
a general description of some important social psychologi-
cal principles that are relevant not only for the design of
warning labels, but also for the design of anti-smoking
campaigns more generally.
Tobacco package warning labels 189
37 Presti DE, Ary DV, Lichtenstein E. The context of smoking initiation and Download full-text
maintenance: findings from interviews with youths. J Subst Abuse
38 Maio GR, Olson JM. Emergent themes and potential approaches to
attitude function: the function-structure model of attitudes. In: Maio GR,
Olson JM, eds. Why we evaluate: functions of attitudes. New Jersey:
Lawrence Erlbaum Associates, 2000:417–42.
39 Katz D. The functional approach to the study of attitudes. Public Opinion
40 Smith MB, Bruner JS, White RW. Opinions and personality. New York:
41 Pryor JB, Reeder GD, Vinacco, JR, et al. The instrumental and symbolic
functions of attitudes towards persons with AIDS. J Appl Soc Psychol
42 Herek GM. Beyond “homophobia”: a social psychological perspective
on attitudes towards lesbians and gay men. J Homosex 1984;10:1–21.
43 Ennis R, Zanna MP. Attitudes, advertising, and automobiles: a functional
approach. Adv Consumer Res 1993;20:662–6.
44 Ennis R, Zanna MP. Attitude function and the automobile. In: Maio, GR,
Olson JM, eds. Why we evaluate: functions of attitudes. Hillsdale, New
Jersey: Erlbaum, 2000:395–415.
45 Belk RW. Materialism: trait aspects of living in the material world.
J Consumer Res 1985;12:265–80.
46 Belk RW. Possessions and the extended self. J Consumer Res
47 Belk RW. Extended self and extending paradigmatic perspective. J
Consumer Res 1989;16:129–32.
48 DeBono KG. Investigating the social-expressive and value-expressive
functions of attitudes: implications for persuasion processes. J Pers Soc
49 DeBono KG. On the processing of functionally-relevant consumer
information: another look at source factors. Adv Consumer Res
50 DeBono KG, Packer M. The effects of advertising appeal on perceptions
of product quality. Pers Soc Psychol Bull 1990;17:197–200.
51 Shavitt S. The role of attitude objects in attitude functions. J Exp Soc
52 Snyder M, DeBono KG. Understanding the functions of attitudes: lessons
from personality and social behavior. In: Pratkanis AR, Breckler SJ,
Greenwald AG, eds. Attitude structure and function. Hillsdale, New
Jersey: Lawrence Erlbaum Associates, 1989:339–60.
53 Burnett JJ, Oliver RL. Fear appeal effects in the field: a segmentation
approach. J Marketing Res 1979;16:181–90.
54 Bandura A. Exercise of personal agency through the self-efficacy
mechanism. In: Schwarzer R, ed. Self-efficacy: thought control of action.
Washington DC: Hemisphere Publishing Corp, 1992:3–38.
55 Ajzen I. From intentions to actions: a theory of planned behavior. In Kuhl
J, Beckmann J, eds. Action control: from cognition to behavior. Berlin:
56 Bandura A. Social cognitive theory and exercise of control over HIV
infection. In: DiClemente RJ, Peterson JL, eds. Preventing AIDS: theories
and methods of behavioral interventions. New York: Plenum Press,
57 Oman RF, Duncan TE. Woman and exercise: an investigation of the
roles of social support, self-efficacy, and hardiness. Med Exerc Nutr
58 McAuley E, Courneya KS. Adherence to exercise and physical activity
as health promoting behaviors: attitudinal and self-efficacy influences.
Appl Prev Psychol 1994;2:65–77.
59 Sullum J, Clark MM, King TK. Predictors of exercise relapse in a college
population. J Am Coll Health 2000;48:175–80.
60 Dorsey, AM, Miller KI, Scherer, CW. Communication, risk behavior and
perceptions of threat and efficacy: a test of a reciprocal model. J Appl
Commun Res 1999;27:377–95.
61 Lee NK, Oei TPS, Greely JD. The interaction of alcohol expectancies and
drinking refusal self-efficacy in high and low risk drinkers. Addiction Res
62 Armitage CJ, Conner M. Distinguishing perceptions of control from
self-efficacy: Predicting consumption of a low-fat diet using the theory of
planned behavior. J Appl Soc Psychol 1999;29:72–90.
63 Tunstall CD, Ginsberg D, Hall SM. Quitting smoking. Int J Addict
64 Bandura A. Social foundations of thought and actions: a social cognitive
theory. Englewood Cliffs, New Jersey: Prentice-Hall, 1986.
65 Hafstad A, Aaro LE, Langmark F. Evaluation of an anti-smoking mass
media campaign targeting adolescents: the role of affective responses
and interpersonal communication. Health Educ Res 1996;11:29–38.
66 Festinger L. A theory of cognitive dissonance. Stanford, California:
Stanford University Press, 1957.
67 Stone J, Aronson E, Crain AL, et al. Inducing hypocrisy as a means of
encouraging young adults to use condoms. Pers Soc Psychol Bull
68 Aronson E, Fried C, Stone J. Overcoming denial and increasing the
intention to use condoms through the induction of hypocrisy. Am J Public
69 Chaiken S. Heuristic versus systematic information processing and the
use of source versus message cues in persuasion. J Pers Soc Psychol
70 Chaiken S. The heuristic model of persuasion. In: Zanna MP, Olson JM,
Herman CP, eds. Social influence: the Ontario symposium, vol 5.
Hillsdale, New Jersey: Erlbaum, 1987:3–39.
71 Zuckerman A, Chaiken S. A heuristic-systematic processing analysis of
the effectiveness of product warning labels. Psychol Marketing
72 Bornstein, RF. Exposure and affect: overview and meta-analysis of
research. Psychol Bull 1989;106:265–89.
73 Environics Research Group. Canadian adult and youth opinions on the
sizing of health warning messages. Report to Health Canada, 1999.
74 Fischer PM, Krugman DM, Fletcher JE, et al. An evaluation of health
warnings in cigarette advertisements using standard marketing research
methods: what does it mean to warn? Tobacco Control 1993;2:279–85.
75 Krugman DM, Fox RJ, Fletcher JE, et al. Do adolescents attend to
warnings in cigarette advertising? An eye tracking approach.
J Advertising Res 1994;39-52.
76 Canadian Cancer Society. Controlling the epidemic – round two and
round three; a submission to the Chief, Legislative and Regulatory
Processes, Environmental Health Directorate, Health Protection Branch,
Ottawa, on amendments to the Tobacco Products Control Regulations as
published in the Canada Gazette Part I, 20 March 1993, Ottawa. 16
77 Keller KL. Memory factors in advertising: the effect of advertising
retrieval cues on brand evaluations. J Consumer Res 1987;14:316–33.
78 Backer TE, Rogers EM, Sopory P. Designing health communication
campaigns: what works? Newbury Park, California: Sage, 1992.
79 Cialdini RB. The poison parasite defense: a new way to sap a
competitor’s persuasive strength. American Psychological Society,
William James Distinguished Lecture at the Annual meeting of the
Midwestern Psychological Association (MPA), 2001.
80 Krugman DM, Fox R. Journal of Health Communication
81 Richards JW, Fischer P, Connor FG. The warning labels on cigarette
packages are ineffective. JAMA 1989;261:45.
82 Cecil H, Evans RI, Stanley MA. Perceived believability among
adolescents of health warning labels on cigarette packs. J Appl Soc
83 Linthwaite P. Health warnings. Health Educ J 1985;44:218–19.
84 Borland R. Tobacco health warnings and smoking-related cognitions and
behaviours. Addiction 1997;92:1427–35.
85 Borland R, Hill D. Initial impact of the new Australian tobacco health
warnings on knowledge and beliefs. Tobacco Control 1997;6:317–25.
86 Hafstad A, Aaro LE, Langmark F. Evaluation of an anti-smoking mass
media campaign targeting adolescents: the role of affective responses
and interpersonal communication. Health Educ Res 1996;11:29–38.
87 Hu T, Sung HY, Keeler TE. Reducing cigarette consumption in California:
tabocco taxes vs. an anti-smoking media campaign. Am J Public Health
88 Jenkins CN, McPhee SJ, Le A, et al. The effectiveness of a media-led
intervention to reduce smoking among Vietnamese-American men. Am J
Public Health 1997;87:1031–4.
89 Towson SMJ, Pepperall G, Zanna MP. The impact of a smoking
prevention program on the evaluation of smokers and non-smokers.
Presented at the annual meeting of the Canadian Psychological
Association, Ottawa, 1984. Abstract published in Canadian Psychology
90 Swanson JE, Rudman LA, Greenwald, AG. Using the Implicit
Association Test to investigate attitude-behaviour consistency for
stigmatized behaviour. Cognition Emotion 2001;15:207–30.
91 Greenwald AG, McGhee DE, Schwartz JLK. Measuring individual
differences in implicit cognition: the Implicit Association Test. J Person Soc
92 Atkin, CK, Wallack, L. Mass communication and public health:
complexities and conflicts. London: Sage, 1990.
93 Rice, RE, Atkin, CK. Public communication campaigns. London: Sage,
190Strahan, White, Fong, et al