To our knowledge, this is the first study to examine the impact of cigarette package design in Latin America, and one of the few studies in any country to focus on young women. The findings indicate that removing color and brand descriptors significantly reduced brand appeal. The greatest decreases in brand appeal were observed when both brand imagery and descriptors were removed from packages. These findings are consistent with previous studies suggesting that when cigarette packages are stripped of their color, imagery and fonts, youth find the packs less appealing [7, 16, 19, 23].
In the branded condition, single packs with female-oriented color schemes (e.g., pinks, light green) and designs (e.g., the black Virginia Slims pack) were generally rated more favorably than packs that appeared more gender neutral (e.g., the white Marlboro Original Gold pack). These findings are consistent with research on industry documents that suggest that marketers advertise ‘female’ brands of cigarettes as a means of satisfying their needs such as peer belonging, self-confidence, and female camaraderie . The “slim” and “lipstick” shaped packages were also rated relatively favorably, even in the plain conditions. In fact, these narrower packages were rated as "more appealing" by the largest proportion of people across all 10 packs in the plain-no descriptors condition. It is unclear whether this was due to the “slim” pack shape or the overtly feminine brand family names that remained on packs (e.g., Silk Cut, and Virginia Slims). Future research should consider experimentally manipulating pack shape and size. The present findings are consistent with past research and suggest that regulators may need to consider restricting pack size and shape if they wish to reduce brand appeal among young women .
The findings also highlight the potential appeal of “brand family names” such as Vogue. Indeed, the plain pack with only the name Vogue was rated as more appealing by a greater proportion of women than the Vogue pack that also featured the descriptor “bleue” (60% versus 39%, respectively). If plain packaging is implemented, one might expect the tobacco industry to create not only more appealing brand descriptors, but more alluring brand names, including those with pre-existing associations with glamour and fashion. Future studies could monitor the extent to which tobacco companies alter brand names and the use of brand descriptors following implementation of plain packaging regulations in Australia.
The removal of imagery and descriptors also influenced perceptions of taste. Across the brands, the greatest decreases in perceived taste were observed when both the imagery and the descriptors were removed. This was especially true for the five flavored cigarette packages. While removing the brand imagery significantly reduced taste ratings for two of the five flavored brands, removing both the imagery and the descriptors significantly reduced taste ratings for all five flavored brands. This finding indicates that plain packaging and greater restrictions on descriptors—or even prohibiting flavored cigarettes— could reduce positive perceptions of taste, and potentially undermine an individual’s desire to smoke . These results are particularly notable given that Brazil recently announced a ban on flavor additives, including menthol.
The findings on perceived smoothness were similar to those for brand appeal and taste. Across the brands, perceived smoothness was lowest for plain-no descriptors packs. Removing the brand imagery significantly reduced smoothness ratings for four packages –the four packages that were also rated most frequently as “female”. This may indicate that consumers associate feminine packs with being less harsh or smoother on the throat. Previous research indicates that perceived smoothness is an important predictor for smoking initiation among youth and is strongly correlated with perceptions of risk [21, 26, 27]. Brands perceived as smooth may promote experimentation or smoking initiation by reducing harshness and irritation .
It was initially hypothesized that plain packaging would reduce the misconception that particular brands were less harmful than others; however, no significant differences between conditions were observed. This difference may not have been observed because of a “floor effect” in the health risk ratings. For the majority of branded packages, less than 15% of the participants rated the packs as “less harmful”; therefore, there was little room for this proportion to significantly decrease across the other conditions. The presence of a floor effect may indicate that the tobacco control messages in Brazil, including large pictorial health warnings, may have been effective in communicating that all cigarette brands are equally harmful, and that there are no “safe” cigarettes [28, 29]. The effect of packaging on perceptions of risk may be different in other low-and middle-income countries with lower awareness about the health effects of smoking. Previous research suggests that plain packaging can increase attention to health warnings [21, 30]. Cigarette packages shown in the current study, however, did not display a pictorial health warning on the package front, as per existing regulations in Brazil. If health warnings were visible during the study, this may have led to differences in health ratings between the plain packages and branded packages.
Plain packaging reduced the likelihood that young women would associate particular packs with attributes such as being female, stylish, and/or sophisticated. These findings add to the evidence that cigarette packaging acts as a type of “lifestyle” advertising [16, 19].
The pack selection task provided additional evidence that branded packaging was more desirable than plain packaging among female youth. Participants were three times more likely to select a branded pack than a plain pack, consistent with previous research . The findings also demonstrate that, as one might expect, plain packages with more engaging brand names and descriptors were more likely to be selected. The two plain packages chosen most frequently had flavor descriptors (DJ Mix Strawberry Flavor and Peel Sweet Melon). This highlights the appeal of flavored cigarettes to youth, and provides further support for a ban on flavored cigarettes .
Strengths and limitations
Participants in this study were not recruited through random sampling and were limited to individuals with internet access. In 2011, Brazil had an internet penetration of 41%, or almost 76 million people . Individuals with internet access likely have a higher degree of education and literacy than the general population. In addition, the self-reported smoking prevalence in our sample (28.4%) was higher than national smoking prevalence estimates for young women. Therefore, the findings may not generalize to the broader population of female youth in Brazil.
Some degree of self-selection bias was also likely given that participants self-selected into the study by enrolling in GMI’s participant panel; however, participants were unaware of the topic of study when beginning the survey. In addition, cigarette branding and industry marketing is national in scope and varies very little across regions of Brazil ; therefore, one would expect little systematic differences in tobacco/branding exposure between the current sample and the general population. Furthermore, the between-subjects experimental design and randomization of participants to experimental conditions should ensure that any sampling biases were equally distributed across groups and therefore could not account for differences found among conditions.
The extent to which the current findings apply to other Latin American countries is unclear due to cultural differences between countries, for example, in terms of what packaging elements the participants find appealing, or stylish. Differences by country in the implementation of other tobacco control policies, such as advertising bans and the use of pictorial health warnings on packaging could also influence perceptions. However, pack design and other tobacco marketing practices are often similar across countries, and the findings from this study were similar to those obtained in studies conducted in many Western countries [7, 16].
In this study, participants viewed images of cigarette packages, rather than observing packs directly. It is inherently difficult for packaging studies to replicate “naturalistic” exposure to packaging as it occurs in real-life. Multiple factors such as package display, advertising, peer influences, and even the weight and feel of the packs may enhance the influence of package design. Consequently, the findings may underestimate the impact of package design.
The inclusion of the pack selection task allowed brand appeal to be assessed using not only a 5-point Likert scale, but also as a behavioral measure. The order in which the packs were selected in terms of frequency was generally consistent with the ratings of appeal in the individual pack rating section. This consistency provides some evidence of the validity of the pack rating questions as proxies for consumer behavior.