Article Text

Download PDFPDF

Effects of plain package branding and graphic health warnings on adolescent smokers in the USA, Spain and France
  1. J Craig Andrews1,
  2. Richard G Netemeyer2,
  3. Scot Burton3,
  4. Jeremy Kees4
  1. 1Marquette University, Milwaukee, Wisconsin, USA
  2. 2University of Virginia, Charlottesville, Virginia, USA
  3. 3University of Arkansas, Fayetteville, Arkansas, USA
  4. 4Villanova University, Villanova, Pennsylvania, USA
  1. Correspondence to Dr J Craig Andrews, Marquette University, Milwaukee 53233, WI, USA; craig.andrews{at}marquette.edu

Abstract

Objective The purpose of this study is to provide an experimental test of the effects of plain pack branding and graphic health warnings (GHWs) in three different countries for an important and vulnerable population, that is, adolescents who are experimenting with smoking.

Methods The effects of plain pack branding (logo present, logo absent), and graphic visual warning level (absent, low, medium, high) are studied experimentally for their impact on adolescent cigarette craving, evoked fear, pack feelings and thoughts of quitting in the USA, Spain and France. A total of 1066 adolescents who were experimenting with smoking served as participants in the study. A quota sample produced 375 respondents in the USA, 337 in Spain and 354 in France.

Results Overall findings indicate that the GHWs were effective in impacting adolescent cigarette craving, evoked fear, pack feelings and thoughts of quitting. The plain pack effects were not as strong, yet reduced craving, increased fear, and decreased pack feelings for all three samples combined, and for US adolescent smokers individually, irrespective of the GHWs. For French adolescent smokers, plain pack effects for craving were limited to low/moderate GHW levels. For Spanish adolescent smokers, plain pack feeling effects were limited to the absence of the GHWs.

Conclusions The results show that plain packs can independently strengthen the more instantaneous, direct effects (short of quitting thoughts) found with the GHWs. Yet, the plain pack results were attenuated for Spanish and French adolescent smokers, who are currently exposed to GHWs.

  • Packaging and Labelling
  • Prevention
  • Priority/special populations
  • Public policy
  • Advertising and Promotion

Statistics from Altmetric.com

Request Permissions

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.

Introduction

Tobacco use is the single most preventable cause of death in the world today, accounting for almost 6 million deaths each year.1 In the USA, there are approximately 483 000 deaths annually due to smoking and an estimated 350 000 new, underage smokers.2 Sadly, 88% of all adult smokers begin in their teens or younger, with smoking being described by many as a ‘pediatric disease’.2–4 Currently, cigarette-smoking rates for high school students have declined to 9.2%, yet the use of any tobacco product remains high at 24.6%.5 Although research shows that graphic health warnings (GHWs) on tobacco packages have helped,6 perhaps the most important initiative in an extension of the GHWs to curb adolescent smoking is the use of plain packs that remove all branding colours, imagery, logos, and trademarks, except for a small brand name identifier and the GHW. The WHO has touted the 2011 Australian plain pack legislation as ‘setting a new global standard’ for control of a product that kills millions each year worldwide.7 Recently, the UK, Ireland, and France are in the process of implementing plain packing, with Spain (and others) retaining the right for future use.8 ,9 Although there are many studies of the impact of GHWs on adult smokers,6 ,10 ,11 and a few examining effects of plain packs and warnings on adult smokers, experimental plain pack research is needed on one of the most important and vulnerable populations, that is, adolescents experimenting with smoking.

GHWs and experimental research

Currently, 176 countries (including France in 2004, Spain in 2005, yet not the USA) have ratified or accepted the WHO's Framework Convention on Tobacco Control (FCTC) treaty requiring tobacco warning information in the form of text, visual (eg, GHW), or a combination of the two (WHO 2003).12 GHWs, offering a visual depiction of the health consequences associated with smoking, are now required in over 77 countries, beginning with Canada in 2001.13 In 2005, the EU adopted a library of 42 (GHW) visuals for use with text warnings, and France and Spain began requiring GHWs beginning in 2011.13 In the USA, although the Family Smoking Prevention & Tobacco Control Act (FSPTCA) (2009) mandated the use of the colour GHWs with text warnings, their use has been blocked by a US federal court decision.6

A wide variety of research studies have been conducted on GHWs involving many different outcomes.6 Still, relatively few controlled experiments have been conducted to test the effectiveness of GHWs,14–26 which are necessary to establish cause and effect relationships.27 ,28 For example, experimental studies have shown that the level of graphicness affects quit thoughts for adolescent smokers, with both high and moderate graphic levels being more effective than low or no graphic warnings.15 ,22

Plain pack research

Brand and logo information (‘trade dress’) (eg, name, colour, design, shape) are important elements in developing brands by differentiating one brand from another in the minds of consumers.29 Strong brand logos and identifying characteristics tend to aid brand recognition, convey consistent meaning to consumers, and evoke positive brand feelings.30 As applied to tobacco packaging, eye-tracking research has shown that plain cigarette packs (ie, packages with only a small, brand name identification and without colour brand logos) remove attention from brand information and shift it towards GHWs for non-smokers and weekly smokers, but not for daily smokers who consciously avoid the warnings.31 ,32 Also, plain tobacco packs displaying progressively fewer branding design elements are perceived to be increasingly unfavourable by adult smokers33 and result in an increase in strong cognitive, emotional, motivational, and avoidance reactions.34–37 In experimental auctions, bids from adult smokers are lowest for plain unbranded packs with prominent visual warnings.38 ,39 In a naturalistic field study, young adult UK smokers used plain packs for 2 weeks and regular packs for 2 weeks.40 Results showed that the plain packs increased (1) negative feelings about the packs, (2) pack avoidance behaviour, and (3) quitting thoughts to a greater extent than regular packs. Similar results were found in a field study of plain packs among young adult, roll-your-own, smokers in France.41 Yet, in another study, although standardised (plain) packs were shown to reduce cigarette craving over time, they did not affect adult smoker motivation to quit.42

Because prior plain pack studies have used adult smokers, an important question is “Will adolescent smokers react in the same way?” Focus groups with adolescents (not necessarily smokers) indicate that plain packs are likely to enhance the salience and impact of the graphic warning labels.43 Other survey work with adolescents (again not necessarily smokers) has shown shifts over time in pack perceptions, brand image, and beliefs of harmful consequences due to a combined presentation of the GWHs and plain packs.44 ,45 Our experimental study across three different countries focuses on two key outcomes from GHWs and plain packs—that is, short-term cigarette craving21 and longer-term thoughts of quitting.22 We also examine separate main and interactive effects for GHWs and plain packs in our experimental design, as previous research has traditionally studied only perceptions of their combined effects. Based on this discussion, it is predicted that:

H1: Stronger GHWs will lead to (a) a greater reduction in adolescent smoker cigarette craving and (b) a greater increase in thoughts of quitting than will weaker GHWs (ie, with text-only warnings).

H2: Plain packages will lead to (a) a greater reduction in adolescent smoker cigarette craving and (b) a greater increase in thoughts of quitting than will packages with brand logo information.

H3: There will be an interaction between the GHWs and plain pack branding such that the effect of stronger GHWs is greater for plain packs than for packs with brand logo information, leading to (a) a greater reduction in adolescent smoker cigarette craving and (b) a greater increase in thoughts of quitting. Such branding differences are expected to be stronger for non-graphic (text-only) health warnings.

We address these predictions across and within three countries, two of which (Spain and France) currently require GHWs and one which does not (the USA).

Methods

Sample and general procedure

Data collection

A professional marketing research firm, with expertise in research involving adolescent smokers and conducting international online surveys, was used to collect the data. A double consent procedure was employed, that is, parents were first contacted to obtain permission and then the adolescents were asked for permission. Participants were adolescents aged 13–18, split evenly across the USA, France and Spain, who indicated that they smoked at least once in the past 30 days.46 A set of academic linguistic experts and young adults fluent in the language for the sample were used to help translate the stimuli and questionnaire for both the pretest and main study from English (US) into French and Spanish. The stimuli and questionnaire were then back translated into English to ensure the intended meaning.

Pretest

A pretest of 150 adolescent smokers (50 US, 50 Spain and 50 France) was used to help select three graphic visuals (low, medium and high graphicness) from nine pictures depicting mouth cancer from around the world. (Mouth cancer has been noted as one of the most recognised and effective package warnings and is particularly relevant to the sample).47 Three age quotas (ie, 13–14, 15–16 and 17–18), based on the University of Michigan Monitoring the Future data on adolescent smokers, were used to help ensure representative samples in all adolescent age groups. A 50/50 split was made on gender to ensure equal representation in the samples, and smoking status in the past 30 days46 was screened as well. Each adolescent was exposed to a randomly-ordered set of nine pictures depicting mouth cancer and then asked a set of questions regarding perceived graphicness, evoked fear and fit with a text warning on mouth cancer under each picture. For the key measure of graphicness, each respondent evaluated the given picture on four, seven-point scales: graphic—not graphic; vivid—not vivid; powerful—weak; and intense—not intense; α=0.97).15 ,22 Finally, one additional single item, seven-point measure assessed the degree to which respondents felt the picture fit a text warning of ‘mouth cancer’, and a second, seven-point measure assessed the perceived believability of the picture. The level of evoked fear from each picture was measured with four, seven-point scales: fearful—not fearful; afraid—not afraid; anxious—not anxious; and nervous—not nervous; α=0.97).15 ,22 Standardised mean scores were calculated and CIs within each country showed items that were significantly different from one another (p≤0.05). The graphicness results for the moderate and low graphic visuals were consistent across all three countries aiding the selection of the picture for the moderate and low categories. The high graphic picture evaluation varied somewhat across three countries, with the selected picture providing the greatest consistency across all measures.

Main study procedure and design

The main study again used a double consent procedure (ie, parents gave their permission, then adolescents) before being screened on age, gender, and smoking status (in the past 30 days).46 The study sample consisted of 1066 adolescents experimenting with smoking—375 in USA, 337 in Spain and 354 in France, with quotas producing a 50/50 split on gender and age categories of 15% (13–14 years.), 33% (15–16 years.), and 52% (17–18 years.) based on the University of Michigan Monitoring the Future data. Following screening questions, respondents were told they would be viewing a cigarette package and then asked some questions about it without any reference to warnings. After successful screening for consent and quotas, respondents were randomly assigned to one of the eight experimental versions of cigarette packs (see figure 1 for the English version). The top one-half of the package facing included the text warning and graphic visual health warning (if included), with the GHW taking up one-third of the total package facing. The bottom half of the pack facing included either the traditional brand logo and colours or the plain pack version (brand name minimised in white on a drab olive background, similar to Australia's plain packs). Thus, the main study used a four (GHW level: absent/control, low, medium, high)×2 (plain pack branding: plain pack with minimised brand name and no brand logo or colour, traditional brand logo with current brand name and logo and colour) between-participants design. The leading brand available across all three countries (Marlboro) was used for pack stimuli.48

Figure 1

Examples of the plain pack branding and graphic health warnings manipulations.

Main study measures

Key dependent variables of interest include cigarette craving (single-item, 6-point scale); evoked fear (4-item, 7-point scale, α=0.94); pack feelings (3-item, 7-point scale, α=0.87); and quitting thoughts (4-item, 7-point scale, α=0.81). Cigarette craving was measured by asking respondents, “After having viewed the cigarette package, how much do you want to smoke a cigarette right now?” with end points of ‘not at all’—‘very much’.21 The evoked fear measure asked respondents how the package made them feel with end points of ‘not fearful’—‘very fearful’; ‘not afraid’—‘very afraid’; ‘not nervous’—‘very nervous’; and ‘not anxious’—‘very anxious’.15 ,22 For pack feelings, respondents were asked, ‘Personally, when I looked at the cigarette pack, I felt…’ ‘embarrassed’—‘not embarrassed’, ‘ashamed’—‘not ashamed’, and ‘accepted’—‘not accepted’.40 Thoughts of quitting were measured on seven-point scales by the following: (1) ‘The information shown on the cigarette package would help me quit smoking’ (‘strongly disagree—strongly agree’), (2) ‘The information on the cigarette package motivates me to quit smoking’ (‘strongly disagree—strongly agree’), (3) ‘How important is it for you to quit smoking?’ (‘not at all important—very important’), and (4) ‘How often do you think about quitting smoking?’ (‘not often—very often’).22

Results

Manipulation checks were used to ensure participants recognised the brand information on the package and 90% recognised that they were exposed to the Marlboro brand across all treatments. Also, in the absence of graphic warnings, those exposed to the colour, brand logo condition (M=5.16) felt the package's brand was significantly brighter and more colourful (on a seven-point agreement scale to ‘The package's brand was bright and colourful’) than those exposed to plain packs (M=2.84; t=10.25; p≤0.01).

A multivariate analysis of covariance was first conducted to examine main effects for the GHWs, then for plain pack branding, and interaction effects between the GHWs and plain pack branding. Respondents' age, gender and familiarity with the Marlboro brand served as covariates. Also, the category breakdown for respondent 30-day smoking frequency46 is as follows: 29.6% reported having smoked 1–2 days; 17.4% smoked 3–5 days; 11.8% smoked 6–9 days; 14.7% smoked 10–19 days; 8.0% smoked 20–29 days; and 18.5% reported having smoked all 30 days. Smoking rates for all 30 days in France (28%) were significantly higher than in Spain (15.7%) and the USA (12%; χ2=25.74, p≤0.05). Table 1 shows the overall multivariate and univariate findings for all countries combined as well as the effects within each country. Tables 24 then display the marginal means and SDs for each independent variable.

Table 1

Multivariate and univariate results for the effects of graphic health warning levels and plain packaging on adolescent smokers from the USA, Spain and France

Table 2

Means (and SDs) for the effects of graphic level (GL) conditions on adolescent smokers from the USA, Spain and France

Table 3

Means (and SDs) for the effects of plain pack (PP) conditions on adolescent smokers from the USA, Spain and France

Table 4

Means (and SDs) for the effects of country (C) conditions on adolescent smokers from the USA, Spain and France

Effects of GHWs

Hypothesis 1 predicted that stronger graphic visual health warnings should result in a greater reduction in cigarette craving and a greater increase in thoughts of quitting than will weaker (text-only) warnings. Supporting H1a and H1b, our findings in table 1 indicate that the graphic visual level had a significant and negative effect on cigarette craving (F(3,1065)=21.55, p≤0.01) and had a significant and positive effect on thoughts of quitting (F(3,1065)=10.83, p≤0.01). As shown in table 2, Bonferroni contrasts indicate that both high (M=1.83) and moderate (M=2.07) graphic visual warning levels significantly reduced cigarette craving for adolescent smokers compared to both control (no graphic) (M=2.67) and low graphic visual warning (M=2.53) conditions (all p≤0.05). Also in table 2, Bonferroni contrasts reveal that both high (M=4.76) and moderate (M=4.60) graphic visual warnings significantly increased adolescent smoker thoughts of quitting compared to both control (no graphic) (M=4.08) and low graphic visual warning (M=4.24) conditions (all p≤0.05). In addition, table 2 also shows similar effects of the graphic visual warning levels on evoked fear and pack feelings.

Effects of plain pack branding

Hypothesis 2 predicted that plain cigarette packages should lead to a greater reduction in cigarette craving and a greater increase in thoughts of quitting than will packages with the current brand logo information. The results in table 1 indicate that the plain pack treatment had a significant and negative effect on cigarette craving (F(1,1065)=8.58, p≤0.01) supporting H2a, yet did not support H2b predicting a positive effect on thoughts of quitting (F(1,1065)=0.21, p=0.65). Bonferroni contrasts in table 3 show that the plan packs (M=2.15) significantly reduced cigarette craving for adolescent smokers compared to the brand logo condition (M=2.40) (p≤0.05). Yet, there were no significant differences between the plain pack condition (M=4.40) and brand logo condition (M=4.44) (p>0.05) on adolescent smokers' thought of quitting. However, table 3 shows that the plain pack condition significantly increased evoked fear and reduced pack feelings. Thus, although independent effects of the plain packs are shown on more immediate measures of craving, evoked fear and pack feelings, these effects did not extend to thoughts of quitting.

GHWs and plain pack branding

Hypothesis 3 predicted an interaction between GHWs and plain pack branding. It was predicted that stronger graphic visual health warnings and the absence of brand logo information will lead to (a) a greater reduction in cigarette craving and (b) a greater increase in thoughts of quitting than with weaker graphic visual health warnings and the presence of brand logo information. The branding differences also should be stronger for non-graphic (text-only) health warnings. Yet, the results in table 1 did not support H3a for a graphic warning×plain pack effect on cigarette craving (F(3,1065)=0.79, p=0.50), nor H3b for a graphic warning×plain pack effect on thoughts of quitting (F(3,1065)=0.68, p=0.56). Thus, separate (independent) main effects occurred for the graphic warnings and plain packs on cigarette craving and for graphic warnings on quitting thoughts. Though there were no significant 2-way graphic warning×plain pack effects overall, a significant 3-way multivariate effect was found across the dependent variables in table 1 (Wilks' λ=0.97; F=1.51, p≤0.05). As discussed below, significant 2-way effects for graphic warning×plain pack occurred for cigarette craving in France and for pack feelings in Spain.

Country effects

As noted in table 1, there were main effects for country (ie, USA, Spain or France) on cigarette craving (F(2,1065)=9.24, p≤0.01) and evoked fear (F(2,1065)=8.99, p<0.01). Bonferroni contrasts in table 4 reveal that cigarette craving was significantly stronger in the France (M=2.54) than in either Spain (M=2.19) or the USA (M=2.09; both p≤0.01). Similarly, evoked fear was significantly less in France (M=3.60) than in either Spain (M=3.93) or in the USA (M=4.14; both p≤0.05).

We also offer an examination of separate intracountry effects (with smaller cell sizes) for the graphic warnings and plain packs in table 1. The results for the US are identical to the overall effects. In Spain, a significant graphic warning×plain pack effect was found for pack feelings (see table 1). Specifically, plain packs (M=3.58) versus branded logos (M=4.54) significantly reduced pack feelings in the absence of the GHWs (F(1,85)=8.63, p≤0.01). (There were no plain pack effects for other GHW conditions.) In France, a significant graphic warning×plain pack effect occurred for cigarette craving (see table 1). In particular, plain packs (M=2.55) versus branded logos (M=3.29) significantly reduced craving under low GHWs (F(1,89)=4.72, p≤0.05) and moderate GHWs (plain pack M=2.17; branded logos M=2.91; F(1,85)=5.57, p≤0.05). (There were no plain pack effects for the other GHW conditions.) Again, caution is in order due to the smaller cell sizes and reduced statistical power in these intracountry analyses.

Discussion

Most experimental research on the GHWs has occurred in a limited set of countries such as the USA. Yet, how will different levels of graphicness affect adolescent smokers across other countries than the USA, some of which have had the warnings in place for some time (eg, France, Spain)? More importantly, can plain packs (ie, without brand colours and logos) independently add to these effects for adolescent smokers? Will plain pack effects vary by country? Our research with over 1000 adolescents experimenting with smoking in the USA, France and Spain sought to answer these questions.

Our results across three countries indicate that graphicness impacts quit thoughts, with high and moderate GHWs significantly increasing adolescent smokers' thoughts of quitting compared to both control (no graphic) and low GHW conditions. In turn, both plain packs and graphicness are found to separately impact more immediate-term measures of reduced cigarette craving, increased evoked fear and decreased pack feelings. In general, effects of GHWs in Spain and France are greater than plain pack effects, and this appears reasonable given outcomes, such as evoked fear and craving, that may be more likely for a graphic pictorial than for removal of brand logo information that may affect brand identification and consumer–brand relationships. Also, note that results are tempered by a significant, overall 3-way effect (graphic warning level×plain pack×country) on the dependent measures. For Spanish adolescent smokers, the plain packs reduced pack feelings without exposure to the GHWs (control condition). In France, the plain packs significantly reduced cigarette craving with exposure to the low and moderate GHWs. Some habituation and/or reactance may be present for French adolescent smokers with their higher smoking frequency and reduced fear when exposed to plain packs with high GHW levels.

Limitations and future research issues

Our experimental study manipulated GHWs and plain pack branding for adolescent smokers across three countries; offering a high degree of internal validity. However, although the online exposure context, sample focus, stimuli tested, measures used, and parental and adolescent consents (to meet IRB requirements) are consistent with prior tobacco warning reseach,15 ,22 they could serve to limit the generalisability of findings. There certainly are other warning themes to study beyond mouth cancer, as well as for other countries, smoking frequencies/situations46 (eg, progression from experimenters to any 30-day use to daily smokers), plain pack variations,33 brand variant descriptors (eg, colour, flavour, filter),49 and other brands and tobacco products (eg, e-cigarettes, hookah, cigars, cigarillos, smokeless tobacco).50 Also, mediation analysis may reveal interesting pathways to quitting from the plain packs and GHWs. Future research might consider using longitudinal data to examine any possible wear-out patterns in countries currently using the plain packs (eg, Australia, Ireland, and the UK, as well as France in 2016). Efforts in this regard will help contribute to a better understanding of how adolescents at risk of becoming lifetime smokers will react to countermeasures, like plain packs and GHWs, aimed at reducing dependence on tobacco.

Conclusion

As found in previous GHW research,15 ,22 the level of graphicness remains important in affecting quit thoughts for adolescent smokers, with both high and moderate graphic levels being more effective than low or no graphic warnings. Our overall results also show that plain pack branding can add to GHW effectiveness—yet only for the more immediate measures of cigarette craving, evoked fear and pack feelings. No doubt, the impact of plain pack branding may take some time and longitudinal assessments may be in order. Country effects for the plain packs indicate that they potentially can aid French adolescent smokers in reducing cigarette cravings—yet only at either low or moderate GHW levels. In Spain, the plain packs reduced pack feelings without exposure to the GHWs (control condition). Yet, the main effects of plain pack branding for Spanish and French adolescent smokers were not significant, perhaps implying that the removal of coloured brand logos for this global and US-based brand is not as important to Spanish and French adolescent smokers as for US smokers for these specific variables. For the US, without either GHWs or plain packs currently appearing on cigarette packs, our plain pack treatment reduced cigarette craving, increased evoked fear and lessened pack feelings without the need of GHWs. Overall, plain packs are found to add to GHWs for more immediate effects (eg, reducing cigarette craving, increasing evoked fear, and decreasing pack feelings) depending on the adolescent smoker's prior exposure in different countries.

What this paper adds

  • Plain pack branding on cigarette packaging is an important and emerging international tobacco control policy. Plain packs often appear in conjunction with graphic health warnings (GHWs) on cigarette packs. This paper builds on the existing literature by examining what the separate and interactive effects might be for plain packs and GHWs. Moreover, we extend previous research on GHWs and plain packs to an important segment—adolescent smokers—and do so across three different countries.

  • Our between-participants experiment with over 1000 adolescent smokers across the USA, Spain and France studies the effects of plain pack branding (logo present, logo absent) and GHW level (absent, low, medium, high) on adolescent cigarette craving, evoked fear, pack feelings and thoughts of quitting.

  • Overall results indicate graphicness impacts quit thoughts; whereas plain packs and graphicness affect more immediate measures of craving, fear/emotion, and pack feelings. The manipulations were less effective for quit thoughts for adolescent smokers in Spain and France, who are currently exposed to GHWs.

  • Plain packs can independently strengthen the more instantaneous, direct effects found with the GHWs.

Acknowledgments

The authors gratefully acknowledge the following individuals for help with the package stimuli, translation and back translation work: Brendan Andrews, Colleen Andrews, Marie Delage, Ana Escudero, Jenna Fanduzzi, Vinay Matta, Kelsey Otero, Megan Otero, and David Wyatt.

References

Footnotes

  • Funding This research received no specific grant from any outside funding agency in the public, commercial or not-for-profit sectors. Internal support was received from Marquette University Miles, Kellstadt, and Way-Klingler funds.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Marquette University Institutional Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement We have access to all data reported in the manuscript and will provide this data on request to the editors or their assignees.