Objectives: This study examined the interactive effects of cigarette package flavour descriptors and sensation seeking on adolescents’ brand perceptions.
Methods: High school students (n = 253) were randomly assigned to one of two experimental conditions and sequentially exposed to cigarette package illustrations for three different brands. In the flavour descriptor condition, the packages included a description of the cigarettes as “cherry”, while in the traditional descriptor condition the cigarette brands were described with common phrases found on tobacco packages such as “domestic blend.” Following exposure to each package participants’ hedonic beliefs, brand attitudes and trial intentions were assessed. Sensation seeking was also measured, and participants were categorised as lower or higher sensation seekers.
Results: Across hedonic belief, brand attitude and trial intention measures, there were interactions between package descriptor condition and sensation seeking. These interactions revealed that among high (but not low) sensation seekers, exposure to cigarette packages including sweet flavour descriptors led to more favourable brand impressions than did exposure to packages with traditional descriptors.
Conclusions: Among high sensation seeking youths, the appeal of cigarette brands is enhanced through the use of flavours and associated descriptions on product packaging.
Statistics from Altmetric.com
Numerous sweet flavours, such as lime, cherry, strawberry, honey, mango and apple, have been used by tobacco firms to describe and market their cigarettes to consumers. Government officials and consumer advocates in Australia, Canada, the United States and the European Union have charged the tobacco industry with introducing such sweet flavoured cigarettes as an attempt to appeal to new smokers.1 2 3 4 5 6 It is known that sweet flavours encourage trial of unfamiliar foods and are particularly palatable to youths.7 8 Accordingly, there is reason to be concerned that flavoured cigarettes may serve as a means to introduce youths to tobacco, and that these cigarettes may be especially appealing to high sensation-seeking youths who are predisposed towards stimuli that offer a novel and potentially heightened sensory experience.9 10 Consistent with this premise, a recent assessment of tobacco industry records revealed that menthol flavours were added to cigarettes to manipulate the sensory characteristics of the product and to enhance the likelihood of smoking initiation.11 These investigators also note that market research studies have found that menthol cigarettes are more popular among adolescents and young adults than older people.
Despite public health concerns, litigation, prohibitions and ongoing policy discussions regarding sweet flavoured cigarettes, little is known about the appeal of these products to youths. Accordingly, the current research provides an initial examination of the conditions under which adolescents’ impressions of cigarette brands are favourably impacted by the inclusion of sweet flavour descriptors (such as “cherry”) on cigarette packaging. In particular, we propose that, by enhancing the arousal potential associated with tobacco brands, sweet flavour descriptors boost the appeal of these products among high sensation seekers.
Cigarette makers have used a variety of product descriptors on cigarette packages. Some descriptors, such as “low tar”, “filtered” and “light”, may convey meaning to consumers regarding the relative dangers of the product. Physical properties of the product are described to consumers using terms, such as “slims”, “wides” and “100s.” In addition, an assortment of descriptors, such as “menthol”, “mild” and “full-bodied”, connote the taste of the product. The study reported in this article focuses on a subset of flavour descriptors, specifically sweet flavours, which have become common in the cigarette market during the past decade.
Across numerous countries, significant progress has been made in discouraging cigarette use in general and in prohibiting the use of misleading product descriptors. Particularly noteworthy among such efforts is the treaty negotiated by the World Health Organization’s Framework Convention on Tobacco Control. Article 11 of the treaty calls for packaging and labelling changes to tobacco products including the elimination of misleading descriptors such as “light” and “low tar”.12 Some jurisdictions have already prohibited these descriptors, but unfortunately, manufacturers have responded by substituting words (for example, “smooth”) that have been found to result in the same misleading effect.13 Recognising the importance of package-based communications, many countries have instituted mandatory on-package warnings regarding the health consequences of smoking (see www.tobaccolabels.org for examples).14 In addition, many countries have established restrictions on the promotion of tobacco products. For instance, the United States Master Settlement Agreement of 1998 aimed at limiting adolescents’ exposure to cigarette advertising by prohibiting tobacco companies from targeting youths in their advertising.
Regulatory efforts are also under way in many countries to specifically address the production, marketing and sale of sweet flavoured cigarettes. Concerns about the potential appeal of flavoured cigarettes among youths have been corroborated by survey findings showing that young smokers (aged 17 to 19) were more likely to report having recently smoked flavoured cigarettes than smokers over the age of 25.15 While some regulators called for an EU ban on flavoured cigarettes, these products continue to be produced and sold within Europe.6 The South Australian Government is leading efforts in Australia to ban any display of fruit or confectionary cigarettes at the point of sale.16 Canada’s parliament is also considering laws that would prohibit the sale of fruit-flavoured cigarettes and small cigars.3 In the US, state-level bans against the sale of flavoured cigarettes have been introduced in New York, Minnesota, West Virginia, Connecticut, Illinois, North Carolina and Texas.17 These regulatory efforts had some initial impact, as RJ Reynolds agreed to stop marketing Camel, Kool and Salem flavoured cigarette in the US.18 However, only seven months after the agreement, the company introduced Camel Signature Blends flavoured cigarettes.19 Adding impact to state regulatory efforts, a recently passed US federal law, titled the Family Smoking Prevention and Tobacco Control Act, bans the use of all artificial or natural flavours in cigarettes (with the exception of menthol) (HR1256).
Sensation seeking and flavour descriptors
We propose that the influence of flavour descriptors lies in their ability to alter the arousal potential of a cigarette brand’s marketing communications (such as its packaging). Arousal potential refers to the degree to which a stimulus is capable of gaining attention and exciting the nervous system.20 The arousal potential of a stimulus has been theorised to depend on a variety of factors including its novelty, affective connotations, incongruity, colour, size and sensory modality.21 In the context of cigarette products, a flavour descriptor included on product packaging is likely to be incongruent with past product experiences and the stimulating colours and graphics may create the perception of a sensory rich product experience. Accordingly, the inclusion of flavour descriptors is expected to enhance the arousal potential of cigarette packaging.
Research has shown that the greater the arousal potential of marketing stimuli, the greater the arousal induced in consumers.22 23 Importantly, however, a long history of research in psychology also indicates that individuals vary in terms of their stimulus arousal preferences.24 25 From this premise, Zuckerman and colleagues theorised that sensation seeking is a measureable individual difference in optimal levels of stimulation and arousal.26 More specifically, sensation seeking has been defined as a general personality trait characterised by the search for unique, complex, ambiguous and emotionally intense stimuli and the willingness to take risks to experience such stimulation.9 10 High sensation seekers gravitate towards novel stimuli and dislike monotonous situations, whereas low sensation seekers prefer the familiar relative to novel things and situations.10 It is also noteworthy that past research has found that high sensation-seeking youths are more likely to smoke cigarettes than their low sensation-seeking peers.27
Mismatches between the arousal potential of marketing stimuli and individuals’ optimal stimulation level can result in unfavourable product judgments and evaluations. In particular, studies have shown that marketing stimuli with high arousal potential have greater appeal among high sensation seekers than low sensation seekers.28 29 Accordingly, we propose the effects of sweet flavour descriptors on adolescents’ brand impressions will be moderated by sensation seeking such that incorporating a flavour descriptor into the cigarette package design will have a positive effect on high (but not low) sensation seekers’ brand related beliefs, attitudes and intentions.
The proposed interaction between flavour descriptors and sensation seeking was examined by conducting an experiment in which high school youths viewed cigarette packages with traditional or flavour descriptors, and answered questions regarding brand related beliefs, attitudes and intentions. We use the term “traditional” to refer to cigarette packaging that does not include sweet flavour descriptors, but may contain general statements regarding taste (for example, full-bodied) or tobacco blend (for example, domestic). We focused on cigarette packaging since, owing to numerous advertising restrictions, packaging has played an increasingly important part as a communication medium for tobacco manufacturers.30 Further, tobacco industry documents indicate that cigarette packs have robust effects on product perceptions and are instrumental in brand image development.31 32 33
In total, 253 high school students, at a school in either the central or south-east United States, completed the study. The sample was about evenly split between the two schools. The participants in the south-eastern high school included a higher percentage of African Americans (29% vs 0%), a smaller percentage of Mexican Americans (0% vs 11%) and a greater percentage of adolescents who reported smoking during the past month (26% vs 17%). Overall, the average age of the participants was 15.7 years and 60% were females. Fifty-seven per cent of the sample reported an awareness of flavoured cigarettes and 19% indicated that they smoke flavoured cigarettes at least once in a while.
Parental consent and student assent were obtained before data collection. To recruit participants, students were provided with a handout that briefly described the study and requested their participation. The handout explained that participants would (1) be asked to look at pictures of different cigarette packages, (2) answer questions about their opinions of the packages and products, (3) be asked questions about their beliefs and behaviour related to cigarettes, advertising, and taking risks, and (4) receive $5 (£3; €3.5) for participating in the study. The two schools were compensated for their administrative help with the recruitment process.
Participants were randomly assigned to the conditions in a 2 (cigarette descriptor) ×3 (brand) ×6 (brand order) mixed factorial design. Cigarette descriptor was a between-subjects factor involving exposure to flavour or traditional (that is, control condition) descriptors. Brand was a within-subjects factor with all participants exposed to packages for three different brands. The order of exposure to these brands was counterbalanced.
The experimental task was completed in a classroom setting with 10–20 youths per session. Teachers, from the schools where the data were collected, attended a study facilitator training session and assisted with administering the study. This approach allowed concurrent data collection across several classrooms. Experimental packets for the conditions were randomly distributed to the participants. The packets included instructions, stimuli and measures. Following a cover page, participants encountered instructions stating that they would be evaluating three cigarette package images and that they should examine the packages as they would if they were viewing them in a store.
The cigarette package stimuli were one dimensional, approximately 2¼×3½ inch (5.7×8.9 cm), full colour representations of the front of the package. Each package image represented a different brand of cigarette and was presented separately. After viewing the first package image, participants turned to the next page and responded to a series of measures (that is, belief, attitude and intention) regarding the brand. Following these measures, participants were exposed to the second brand/package along with its associated measures, and finally participants viewed the third brand/package and responded to the final set of brand-related measures. The final portion of the experimental packet included measures of past and present smoking behaviours, sensation seeking and demographic characteristics.
For debriefing, we provided each participant with a fact sheet about the dangers of tobacco in general and the specific risks/concerns associated with flavoured tobacco, which they were required to read and sign. Participants also received an anti-tobacco bracelet, and anti-smoking posters were left with classroom teachers for future posting and/or distribution to participants. To reduce the opportunity for participants to talk about the study with peers who had not yet participated, most experimental sessions were run concurrently.
Two existing brands (Camel and American Spirit) and one fictitious brand (Onyx) provided the basis for the cigarette package stimuli. Camel has been one of the top three cigarette brands among adolescents and American Spirit has been reported to be the fastest growing cigarette brand in the US.34 35 36 37 The fictitious brand was included in the study to permit a relatively “noise-free” assessment and thereby maximise experimental control. For the fictitious brand, we created a brand name (Onyx) and professionally designed package image (see fig 1). A pretest (n = 36) revealed that, relative to other options, “Onyx” was perceived by high school students to be both an appealing and realistic cigarette brand name.
The traditional descriptor condition served as a baseline for comparison. The Camel package shown in this condition was largely consistent with this brand’s standard packaging. To create a traditional package for the American Spirit brand, we removed the terms “Natural” and “100% Additive-Free Natural Tobacco” from the package and added “regular full-bodied taste.” The traditional package for Onyx was professionally designed and contained traditional descriptors similar to the other brands.
Stimuli for the flavour descriptor condition were constructed via systematic changes to the traditional package design stimuli. To select the flavour that would be used in this condition, we pretested the appeal of 12 flavours via seven-point scales gauging general appeal, appeal as a lip balm flavour and appeal as a cigarette flavour. Relative to other flavours, cherry was favourably scored across all three measures. In addition, cherry was selected because it appealed to both genders and because it is a common tobacco flavour in the marketplace. Thus, for the flavoured condition, we added the flavour name (“Cherry”) and subtle cherry-related background graphics to the packages (see fig 1). Other elements such as the brand names, logo related graphics and package sizes were held constant.
After exposure to each of the three cigarette packages, participants completed a series of seven-point scaled belief, attitude and intention measures regarding the specific cigarette brand. Using scales anchored by very unlikely (1) and very likely (7), “hedonic belief” items assessed the likelihood that the brand is enjoyable, relaxing and good tasting. For example, the enjoyable belief measure asked: “How likely is it that [brand name] cigarettes are enjoyable.” The three hedonic belief items associated with each brand were averaged (α = 0.89 to 0.91).
Brand attitude and trial intentions were measured using indirect measures. Such measures involve responding from the perspective of another person (such as an acquaintance or friend) and thereby result in participants feeling less pressure to respond in a normative fashion. Fisher provides evidence that indirect measures result in participants projecting their own beliefs and evaluations and that this measurement format reduces social desirability bias.38 Such bias was considered likely in the current study given the negative stigma associated with cigarettes. Thus, to assess brand attitude, participants responded to the statement, “My friends’ opinions of [brand name] cigarettes would probably be”, on two seven-point scales anchored by unfavourable/favourable and negative/positive (r = 0.83 to 0.89). Employing a scale anchored by “very unlikely” (1) and “very likely” (7), the single-item trial intention measure asked participants “In the future, how likely is it that your friends will try [brand name] cigarettes?”
The final section of the survey included demographic questions as well as questions about past smoking behaviours. Participants were asked “have you ever smoked” (response scale: yes, no), “in the last month have you smoked cigarettes” (response scale: yes, no) and “do you smoke cigarettes” (five-point scale varying from “not at all” to “a pack or more a day”). This section of the survey also included a two-item attitude towards cigarettes measure (r = 0.66), and a short-form measure of sensation seeking (α = 0.84).39 Example sensation-seeking scale items are: “I would like to explore strange places” and “I like friends who are exciting and unpredictable”. A median split was used to divide participants into lower and higher sensation-seeking groups (Mlower = 3.12, Mhigher = 5.88, p<0.001).
We tested for descriptor effects using a repeated measures analysis of covariance model, with cigarette descriptor and sensation seeking as between-participants factors and brand as a within-participants factor. Additionally, we included school location as a two-level blocking variable. General attitude towards cigarettes, smoking frequency, awareness of flavoured cigarettes and flavoured cigarette smoking frequency were related to outcomes and were thus included as covariates. Brand order, gender, ethnicity and age were unrelated to outcomes and were therefore not included in further analysis. Thus, the final model used in analysis was a 2 (descriptor: flavoured vs traditional) ×2 (sensation seeking: high vs low) ×2 (school location) ×3 (brand) repeated measures ANCOVA model, with brand as the only within-participants factor.
Table 1 presents descriptive statistics for all outcomes by descriptor condition, brand and sensation seeking. With respect to the hedonic beliefs measure, the analysis revealed a significant main effect of package descriptor (F (1, 215) = 18.36, p<0.001). Specifically, the flavour descriptors (M = 3.50) led to more positive beliefs about the hedonic qualities of the brands than the traditional descriptors (M = 2.64). However, this main effect was qualified by a significant interaction between package descriptor and sensation seeking (F (1, 215) = 10.17, p = 0.002); see fig 2A for a graphic representation of the interaction and table 1 for brand level descriptive statistics. All other main effects and interactions were not significant (ps >0.1). Contrasts revealed that, among lower sensation seekers, hedonic brand beliefs did not differ between the two package descriptor conditions (p = 0.32; MTraditional = 2.81, MFlavour = 3.03). However, within the higher sensation-seeking group, the flavour descriptors (M = 3.98) led to more favourable hedonic brand beliefs than the traditional descriptors (M = 2.47; p<0.001).
With regard to the brand attitude measure, a main effect was present for school location (F (1, 211) = 11.42, p<0.001) such that brand attitudes were more favourable at the school located in the south-eastern US (M = 3.15) than the central US location (M = 2.44). The only other significant effect was an interaction between package descriptor and sensation seeking (F (1, 211) = 10.47, p<.001; see fig 2B). A contrast revealed a marginally significant effect (p = 0.10) of the cigarette descriptor within the lower sensation-seekers’ condition such that attitudes were more favourable among those exposed to the traditional descriptors (M = 2.91) than the flavour descriptors (M = 2.53). A second contrast revealed a significant effect (p = 0.003) of the descriptor manipulation among higher sensation seekers with brand attitudes being more favourable among those exposed to the flavour than the traditional descriptors (MTraditional = 2.40, MFlavour = 3.44).
The trial intentions measure also revealed a significant main effect of school location (MSouthwest = 3.25, MCentral = 2.67; F(1, 215) = 7.71, p = 0.006). Once again, the only other significant effect was an interaction between package descriptor and sensation seeking (F(1, 215) = 8.92, p = 0.003; see fig 2C). In assessing the interaction, a contrast revealed that among lower sensation seekers trial intentions were marginally greater in the traditional descriptor condition (M = 3.19) than the flavour descriptor condition (M = 2.77; p = 0.07). Among the higher sensation seekers, the flavoured descriptors (M = 3.36) led to higher trial intentions than the traditional descriptors (M = 2.53; p = 0.01).
This study fills a need for research to inform the policy debate regarding the sale of flavoured cigarettes and answers the call for research on moderators of adolescents’ receptivity to promotions for high risk, addictive products.40 Across three brands and two school settings, findings indicate that high sensation-seeking youths develop stronger hedonic brand beliefs, more favourable attitudes and greater intentions of trial for brands with flavour versus traditional descriptors. Across attitude and trial intention measures, our results also provide some evidence that low sensation seekers may prefer cigarettes with traditional descriptors relative to those with flavour descriptors. The latter finding is consistent with low sensation seekers’ preferences for familiar stimuli; however, the marginally significant results across the attitude and intentions measures do not permit firm conclusions.
We have theorised that the use of sweet flavour descriptors enhances the appeal of cigarette brands among high sensation seekers by heightening the arousal potential of the marketing communications used in promoting such products. While other research has shown that manipulating marketing communications can impact the appeal of public service announcements directed towards high sensation seekers, to the best of our knowledge, the current study is the first to illustrate such effects on brand preferences. Further, the current study illustrates that the robust effects of simply adding “cherry” and associated graphics to cigarette packaging, and the generalisability of the effects across three brands and two, fairly diverse student populations.
It is noteworthy that beyond the use of flavour descriptors, cigarette manufacturers may heighten the arousal potential of their products via other package-based communications, thereby enhancing product appeal among high sensation-seeking youths. For instance, intense or unique graphics and colours, unusual brand names and original package shapes or materials may make these products more attractive to young sensation seekers. From this perspective, the ongoing product and packaging innovations within the tobacco industry may present a significant threat to this segment of youths who are predisposed towards risky behaviours. In the context of the WHO’s Framework Convention for Tobacco Control (Article 11 regarding the packaging and labelling of tobacco products), sweet flavoured cigarettes appear to fall into the category of products that are likely to create a false or erroneous impression that the product is less harmful than other tobacco products. Yet packaging and labelling restrictions, independent of product restrictions, may not adequately protect youths given the ability and swiftness of the tobacco companies to introduce substitute words on their packaging that may be equally appealing.
The current research should also raise concerns about the use of sweet flavour descriptors in other dangerous product markets that may appeal to high sensation-seeking adolescents. Sweet flavour descriptors are widely incorporated into the marketing communications associated with cigars, non-filtered bidis (imported from India and other south-east Asian countries), cloves, cigarette rolling paper and smokeless tobacco.41 42 43 Distillers have also used sweet flavour descriptors in conjunction with vodka, wine coolers and flavoured malt beverages. While the sweet flavour descriptors used within these markets resemble those used in the current study, further research is needed to verify that the current findings generalise to these product contexts.
Our study is not without its limitations. Owing to participant anonymity requirements and other data collection restrictions, assessing sensation seeking at a separate occasion from the measurement of the outcome variables was not feasible. While we believe that the likelihood of demand artefacts (that is, hypothesis guessing and compliant responses) is low, we are unable to rule out such explanations. In addition, aspects of our experimental stimuli and procedures are inconsistent with the complex marketplace settings wherein cigarettes are typically evaluated and selected. Therefore, our findings may not fully reflect brand evaluations formed under normal circumstances. Finally, while we theorise that the attitudinal effects found in our study are mediated by cigarette flavour descriptors enhancing the arousal potential of the cigarette package communications, we did not include measures that allow us to directly assess this process. It is possible that in responding favourably to cigarettes with flavour descriptors high sensation seekers were expressing a preference for novel and stimulating tastes; such a finding would be consistent with research which has shown that high sensation seekers prefer novel tastes.10 Future research could incorporate measures to clarify the mediating process resulting in the appeal of flavoured cigarettes among high sensation seekers.
Despite the legislative progress that has been made to date, the use of flavour descriptors by cigarette companies and by manufacturers of other tobacco and alcohol products will probably remain a public health concern and in the policy spotlight for years to come. Current agreements and rulings are limited to certain companies and product categories, do not cover global markets, are awaiting government action and, in some instances, may be delayed by appeals. Accordingly, further efforts aimed at informing the policy debate and identifying effective intervention strategies by expanding our understanding of the appeal of these products among adolescents should remain a research priority.
The authors thank the editor, reviewers and Michael Slater for their assistance and valuable insights.
Funding This research was funded by a grant from the Robert Wood Johnson Foundation (No 55499).
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.