Article Text
PDF
Abstract
Objective Certain tobacco companies use health-oriented descriptors (eg, 100% organic) on product packaging and advertising of combustible cigarettes or electronic cigarettes (e-cigarettes) that create a ‘health halo’ around smoking and vaping. Previous observational research suggests that such language may be associated with more favourable attitudes and reduced risk perceptions toward these brands compared with others. This study aimed to determine the effects of health-oriented descriptors on smokers’ attitude toward the brand, perception of packaging information, comparative harm versus other brands and intention to purchase either combustible cigarettes or e-cigarettes.
Method US adult smokers were randomly assigned to view either a health-oriented language package (‘100% organic,’ ‘all natural’ or ‘no additives’), traditional marketing language package (‘fine quality,’ ‘premium blend’ or ‘100% original’) or a no-language package of a combustible cigarette brand (Study 1, n=405) or an e-cigarette brand (Study 2, n=396) in an experimental design.
Results Study 1: Participants in the health-oriented condition reported more favourable perceptions toward the package information, lower comparative harm and higher intention to purchase combustible cigarettes versus the no language control. In addition, participants in the health-oriented condition reported more positive attitude toward the brand and lower comparative harm versus the traditional marketing condition. Study 2: Compared with the traditional marketing condition, participants in the health-oriented condition reported greater intention to purchase Absolute e-cigarettes. There were no significant differences in attitude toward the brand, perception of packaging information and comparative harm versus other brands across conditions.
Conclusions The effect of health-oriented language was significant for combustible cigarettesand e-cigarette packages. Policies to restrict health-oriented language on cigarette and e-cigarette packaging are recommended.
- packaging and labelling
- electronic nicotine delivery devices
- advertising and promotion
Statistics from Altmetric.com
Introduction
In 2015, the US Food and Drug Administration (FDA) issued cease and desist letters to three tobacco companies regarding the use of language and imagery that suggest their products were less harmful than other products.1 The FDA prohibited such language as ‘natural’ and/or ‘additive free.’ A recent content analysis by Moran and colleagues found that Natural American Spirit (one of the brands implicated in these FDA letters) continued to use this type of language in their marketing and product packaging.1 There is emerging evidence that cigarette packaging labelled as ‘natural’ is perceived to be more appealing and less harmful among smokers.2 In a nationally representative sample (Population Assessment of Tobacco and Health study (PATH)), the majority of smokers of Natural American Spirit believed that their brand was less harmful.3 Another study among US adults reported that about half of the participants regarded American Spirit cigarette packs that displayed the descriptors ‘Made with Organic Tobacco’ and ‘100% Additive-Free’ (48% and 59%, respectively) as less harmful than packs that were modified to exclude these descriptors.4 Misperceptions about health risks arising from these descriptors may influence tobacco use behaviours. For instance, one survey among US adults found that higher misperceptions that American Spirit cigarettes had lower health risks were associated with increased intention to purchase American Spirit cigarettes.5
Electronic cigarette (e-cigarette) advertising and packaging frequently include health-oriented claims compared with combustible cigarettes.6–9 For example, Grana and colleagues found that e-cigarette retail websites contained statements suggesting that e-cigarettes only contained water vapour and nicotine.8Research documenting the prevalence or effects of specific descriptors including ‘natural’ or ‘organic’ on e-cigarette packaging on consumer perceptions and behaviours is lacking. Anecdotally, images of e-cigarette liquids and packaging can be found on certain e-cigarette websites offering natural or organic e-liquids (eg, E-Swisher, Vape Organics).10 We are therefore interested in the effect of these types of claims on e-cigarette packaging in comparison with combustible cigarette packaging.
In this study, we compared the effects of health-oriented packaging language which potentially conveyed a health halo around smoking and vaping,11 in this case, ‘100% organic,’ ‘all natural’ and ‘no additives’ versus controls for both combustible cigarettes and e-cigarettes. The study outcomes were attitude toward the brand, credibility of the information on the package, comparative harm of the brand versus other brands and intention to purchase the product.
Methods
We conducted two randomised controlled experiments in November 2016 among US adult daily smokers. In Study 1, participants were told that they would view a combustible cigarette package (n=405), while in Study 2, participants were told that they would view an e-cigarette package (n=396). Participants were recruited through Qualtrics Survey panels. They were told that the survey would take approximately 15 min. If a participant did not respond for a request for participation or had an invalid email address, the next panel member was selected for recruitment. Participants were paid in Qualtrics points, which were equivalent to between $0.75 and $0.85 dollars, depending on how they used their points. Other studies have used similar Qualtrics panels.12–14 To be eligible, participants have to have smoked at least 100 cigarettes in their lifetime and smoked every day in the last 30 days. Qualtrics was able to differentially recruit participants who reported smoking on a baseline questionnaire. We oversampled for African American and Hispanic participants. Of the 16 530 invitations sent to panel members, 1678 participants viewed the survey and 801 completed it. Of the participants who did not complete the survey, 57 quit part of the way through the study and 851 did not meet the inclusion criteria with the majority (626) being non-daily smokers.
Procedure
Participants were sent a link, through Qualtrics, to participate in the study. They were then consented and answered questions about their smoking behaviour, e-cigarette use, general attitudes toward healthy products and attitudes toward tobacco packaging. In Study 1, participants were randomly assigned to view one of seven cigarette packs for a non-US brand (Absolute) (figure 1). In the health-oriented language condition, the cigarette packs displayed the brand name and one of the following phrases: (1) ‘100% organic,’ (2) ‘no additives,’ or (3) ‘all natural’. The traditional language control packs displayed the brand name and one of the following phrases: (1) ‘fine quality,’ (2) ‘premium blend’ or (3) ‘100% original.’ The no language control pack only displayed the brand name and no labels. In Study 2, participants were told the product was an e-cigarette and the stimuli images and labels were identical to those used in Study 1. Although these terms seem less intuitive in referring to the e-cigarette device (ie, the device itself could not be organic or natural), we hypothesised that the labels could be construed by consumers to refer to the contents of e-cigarettes including the liquid nicotine and flavourings, and therefore influence their perceptions of the product as a whole. In addition, such labels have been used in actual e-cigarette packaging and also on the device itself (eg, Ecig Organic brand15). We selected an international brand intentionally to minimise the likelihood that participants have been exposed to the brand previously. In addition, we chose the package design because it could plausibly contain either regular cigarettes or e-cigarettes based on comparison with existing cigarette and e-cigarette packaging.15 After viewing the cigarette (or e-cigarette) pack, participants answered questions about the outcome variables and their demographic information.
Seven cigarette packs for a non-US brand (Absolute).
Outcome variables
Attitude toward the Absolute brand of cigarettes (or e-cigarettes) was measured using six items. Participants indicated on a seven-point Likert type scale (eg, 1=‘strongly agree’, 7=‘strongly disagree’) if they agreed or disagreed whether Absolute cigarettes would be described as environmentally friendly, healthy, safe, better for you, better tasting and authentic. These items were reverse-coded so that 1= ‘strongly disagree’ and 7=‘strongly agree’ and averaged into a scale (for combustible cigarettes, Cronbach’s alpha=0.93, mean=4.15, SD=1.49; for e-cigarettes, Cronbach’s alpha=0.91, mean=4.96, SD=1.22). These items were adapted from existing research on tobacco industry and e-cigarette industry marketing claims8 16
Perception of information credibility, veracity and effectiveness on package was measured using a five-item scale version that was originally developed to evaluate antismoking messages.16 It has been applied in similar contexts.17 Items were averaged to create a single scale. The stem item was, ‘The information in the message I just saw’ and the adjectives to describe that information were, ‘credible,’ ‘believable,’ ‘new,’ ‘important’ and ‘made me think.’ These items were rated on a scale from 1=‘strongly agree’ to 5=‘strongly disagree’. Each item was reverse-coded so that a higher number on the scale response reflected stronger agreement and the five items were averaged into a scale (for combustible cigarettes, Cronbach’s alpha=0.87, mean=2.44, SD=0.91; for e-cigarettes, Cronbach’s alpha=0.88, mean=2.44, SD=0.83).
Comparative harm was measured using the following item for combustible cigarettes and e-cigarettes, ‘Do you think that Absolute (cigarettes/e-cigarettes) might be 1=‘less harmful’, 2=‘no different’, or 3=‘more harmful’, compared with other (cigarettes/e-cigarettes) brands?’ Participants were also given the option of responding, ‘don’t know.’ Participants who responded ‘don’t know’ were treated as missing (13% of participants in Study 1 and 8% in Study 2).
Intention to purchase was measured using the Juster Scale.17 Participants were asked how likely they were to purchase Absolute cigarettes (or e-cigarettes) in the next 3 months on an 11-point scale. Responses ranged from 10=Certain, practically certain (99 in 100) to 0=No chance, almost no chance (1 in 100) (for cigarettes, mean=3.95, SD=3.69, and for e-cigarettes, mean=5.14, SD=3.53).
Control variables
Level of education was recoded into high school or less, some college, and college or more. We also included gender, whether participants were Hispanic, race (White or non-White), age, typical number of days smoked in the past 7 days and e-cigarette use (daily, some days, rarely and never) in the analyses.
Analysis
Our analyses indicated that the random assignment to conditions was successful. There was no significant difference in participant characteristics across conditions in both studies. We conducted linear regression analyses for Study 1 (combustible cigarette packaging) and Study 2 (e-cigarette packaging) separately to predict attitude toward the brand, perception of the information on the packaging and intention to purchase. We completed an ordinal logistic regression for the comparative harm items (since the scale was ordinal in nature). For each analysis, we compared the health-oriented language condition versus the no language condition (only the Absolute brand was displayed with no labels on the packaging). We repeated the analyses to compare the health-oriented language condition versus the traditional language condition (displayed the Absolute brand and labels including ‘fine quality,’ ‘premium blend’ or ‘100% original’); these results can be found in online supplementary tables 2–5. We collapsed participants who viewed the three different labels in the health-oriented language condition as one group. Similarly, we collapsed participants who viewed the three labels in the traditional language conditions instead of conducting the analyses across individual labels. The rationale for this approach is to avoid case-category confounding, which occurs when an individual message in a given class (for example health-oriented versus traditional marketing descriptors) produces an effect that is unique to that message as opposed to a class of messages.18 Therefore, it is important to have multiple stimuli within a single experimental category. Additionally, in a series of sensitivity analyses, we found no significant differences between labels within the health-oriented and the traditional language conditions (see online supplementary table 1). We fitted unadjusted and adjusted models controlling for age, gender, race, ethnicity and level of education. All analyses were conducted using Stata 14. Due to non-normality in the residuals, we performed multiple regression analyses and obtained bootstrapped standard errors (2000 replications) and bias-corrected 95% CIs. This is a method that has been used previously to address this concern.19–21 The Michigan State University committee for the protection of human subjects approved all aspects of this study.
Supplementary file 1
Results
Participants
Study 1: combustible cigarette packaging
Participants were 50% male, 9.6% Hispanic, 76.3% White and 34.0% were college-educated. Mean age of participants was 40.6 (SD=12.7). Median number of cigarettes smoked per day was 15 (IQR=10). About two in five participants smoked their first cigarette within 5 min of waking (43%) or between 6 and 30 min of waking (44%). In Study 1, 10% of participants used e-cigarettes every day, 31% on some days, 30% rarely and 29% reported never using e-cigarettes.
Study 2: e-cigarette packaging
Participant characteristics were similar to Study 1. About half the participants (47.7%) were male, 9.1% Hispanic, 82.3% White, and 38.4% were college-educated. Mean age of participants was 39.4 (SD=11.5). Median number of cigarettes smoked per day was 15 (IQR=10), 41.7% of participants smoked their first cigarette within 5 min of waking and 44.9% between 6 and 30 min of waking). In Study 2, 17% of participants used e-cigarettes daily, 45.5% used e-cigarettes on some days, 28% used e-cigarettes rarely and 9.6% never used e-cigarettes.
Attitude toward the Absolute brand of cigarettes (or e-cigarettes)
Study 1: combustible cigarette packaging
Participants in the health-oriented condition had more positive attitudes compared with those in the traditional language control condition (B=0.36, p=0.012) (online supplementary table 2). Compared with college-educated participants, those with high school or lower education and some college education had more negative attitudes toward the brand. E-cigarette use (daily or on some days) was associated with more positive attitudes toward the Absolute combustible cigarette brand (table 1, online supplementary table 2).
Attitude toward combustible and electronic cigarettes
Study 2: e-cigarette packaging
There were no significant effects of condition on attitude toward brand of e-cigarette. E-cigarette use (daily or on some days) was associated with more positive attitudes toward the Absolute e-cigarette brand.
Perception of information on package
Study 1: combustible cigarette packaging
Compared with the no language control condition, participants in the health-oriented language condition found the labels to be more credible, believable, important, etc. (B=0.28, p=0.032) (table 2). Participants with some college perceived the packaging information less favourably compared with college-educated participants. E-cigarette users had more positive perceptions of the Absolute cigarette packaging information than people who had not used e-cigarettes in the past 30 days or never users in the analysis comparing the no language control and the experimental conditions (table 2, online supplementary table 3).
Perception of package information for combustible and electronic cigarettes
Study 2: e-cigarette packaging
There were no significant effects of experimental condition on perception of information on e-cigarette packaging. Participants with high school or less education perceived the packaging information more favourable compared with college-educated participants. Compared with people who had not used e-cigarettes in the past 30 day or never users, participants who used e-cigarettes every day or some days reported more negative perceptions of Absolute e-cigarette packaging information.
Comparative harm
Study 1: combustible cigarette packaging
Compared with the no language control (OR=0.35, p=0.001) (table 3) or the traditional advertising control (OR=0.41, p=0.001) (online supplementary table 4), those in the health-oriented condition perceived the cigarette brand to be less harmful than other cigarette brands. Participants who were younger perceived the cigarette brand to be less harmful than other brands.
Ordinal logistics regression analysis of comparative harm of combustible and electronic cigarettes brands with other brands
Study 2: e-cigarette packaging
There were no significant predictors in this model.
Intention to purchase
Study 1: combustible cigarette packaging
Compared with the no language control, the health-oriented condition was associated with greater intention to purchase the Absolute brand of cigarettes (B=1.09, p=0.018) (table 4). Older participants, women (versus men), those with high school or less and some college education (vs college education) had lower intentions to purchase Absolute cigarettes (table 4, online supplementary table 5). Participants who used e-cigarettes every day or some days (vs no past 30-day use or never use) had higher intentions to purchase Absolute cigarettes.
Intention to purchase combustible and electronic cigarettes
Study 2: e-cigarette packaging
Compared with the traditional marketing condition, participants in the health-oriented condition reported greater intention to purchase Absolute e-cigarettes (B=0.85, p=0.017) (online supplementary table 5). Participants who were non-White (versus White) and those who used e-cigarette rarely, some days or every day in the past 30 days (vs never use or no past 30-day use) had higher intentions to purchase e-cigarettes. Older age, women (versus men) and those with higher school or less (versus college education) had lower intention to purchase Absolute e-cigarettes.
Discussion
We found significant effects of a single exposure to health-oriented descriptors on combustible cigarette packaging on multiple outcomes compared with traditional marketing language and/or no-language conditions using a randomised experiment. In addition, we found that health-oriented language on an e-cigarette pack increases intention to purchase compared with traditional marketing language. These findings are consistent with earlier studies examining misperceptions of risk in association with health-oriented descriptors on cigarette packaging. In a study of tobacco cigarette packaging labels among school-aged youth, Czoli and Hammond found that packages with ‘natural’ descriptors were perceived to be more appealing and less harmful.2 Their items were somewhat different than in the present study (eg, they were single item scales geared toward a younger population and did not include intention to purchase or perception of packaging information). However, their experimental results suggest a similar trend that tobacco cigarette marketing with health-oriented descriptors may make the products appear less risky. Findings in this present study are also consistent with Pearson and colleagues’ survey research, which compared consumers’ risk perceptions of Natural American Spirit packs with and without health-oriented descriptors (‘Made with 100% Organic Tobacco’ and ‘100% Additive-Free’) and found that packs with these descriptors were associated with reduced perceived harms compared with those that were modified to omit the descriptors.22
These subtle but significant effects from a brief exposure to cigarette and e-cigarette packaging descriptors are worrisome for public health, tobacco regulations and tobacco control research and practice. Previous research has been conducted using content analysis, qualitative research or cross-sectional surveys. This study adds to the body of evidence on the short-term effects of health-oriented descriptors on tobacco product packaging on consumer perceptions and behavioural intentions by using randomised controlled experiments. Based on the accumulating body of evidence, our results support banning the use of these descriptors on cigarette and e-cigarette packaging and promotional materials and stricter enforcement to ensure that these labels do not appear on advertising or packaging to prevent consumer misperceptions about the harm of cigarette smoking and e-cigarette use. Similar prohibitions have been found to be effective in the case of Australia.23 24
Further, we found that even with identical stimuli, results were substantially different when participants were told they were viewing a combustible cigarette package versus an e-cigarette package. Nevertheless, we observed a significant increase in intention to purchase the e-cigarette brand among smokers who viewed the health-oriented labels compared with the traditional marketing labels, suggesting that this type of health-oriented language associated with e-cigarette packaging may have some effect. One explanation for the small number of significant effects of health-oriented language in the context of e-cigarettes in this study may be because much of the e-cigarette advertising already contains language that imply lower harm, for example, ‘contains water vapour.’8 The ubiquity of this language may have created an overall perception that e-cigarettes pose little harm. In addition, all the participants were smokers and the majority have ever used e-cigarettes so they may have a general perception that e-cigarettes are safer and therefore perceived little difference between the language conditions. Whatever the reason, the results suggest that effects of health-oriented descriptors used for combustible cigarette packaging may not necessarily generalise to packaging of a different tobacco product.
Finally, it is important to consider the effect of these findings more generally for regulation of other aspects of tobacco product packaging. This study contributes to the increasing body of evidence to suggest health-oriented language and images on product packaging may make cigarettes appear to be less risky.3 25 This may be particularly important to support legislation surrounding plain packaging. As the amount or type of information that is allowed to be displayed on tobacco products is being considered across countries, it is important to note that a few simple words may have a meaningful effect on intention to purchase, attitudes and perceived harm of cigarettes and intention to purchase e-cigarettes.
Limitations and strengths
During the period of data collection for this study, the FDA had not ruled on the use of health-oriented language by tobacco companies. In January 2017, Santa Fe Tobacco Company reached an agreement with the FDA to remove the terms ‘additive free’ and ‘natural’ from all Natural American Spirit product labels, advertising and promotional materials within 7 months of the agreement.26 However, Santa Fe may retain the use of the term ‘natural’ in the brand name and trademarks. It is unclear whether this language has continued to be used at present. Participants were from a convenience sample and may not be representative of adult smokers and did not include non-daily or former smokers. Further, only 10% of the participants who were invited to complete the survey viewed the survey and half of these respondents who viewed the survey completed it (the majority were not eligible for the study as they were non-daily smokers). Self-selection bias may be an issue. Therefore, we need to replicate this study in a nationally representative sample of smokers. This study describes results from a single brief exposure to one pack only and may not generalise to smokers receiving repeated exposures to health-oriented language present in other forms, for example, print advertising, coupons, internet and social media, over longer durations of time. Future studies should include longitudinal designs and test the effects of exposure to health-oriented language across multiple media. We did not include a baseline measure of e-cigarette and traditional cigarette risk in order to avoid priming participants to think about risk particularly before viewing the stimuli. This meant that we cannot control for baseline differences in perceived product risk. Future research should therefore include these variables as covariates. In addition, the pack that we selected may not be representative of different brands of cigarettes and e-cigarettes. This study does not describe the underlying mechanisms to explain how exposure to health-oriented language influences each outcome. In future studies, we will examine mediating pathways using additional techniques, for example, psychophysiological measurements, qualitative interviews to explore specific theoretical pathways.
This study has several strengths. We used randomised experiments to maximise internal validity and ensure that conditions only varied in terms of the presence or absence of health-oriented or traditional marketing descriptors. We attempted to minimise case-category confounding by including three health-oriented descriptors and three traditional marketing descriptors. We also included a no-language control condition as well as a traditional language condition, allowing us to understand whether the effects came from having any advertising language at all versus a pack lacking such language. We used a non-US brand to avoid potential bias due to familiarity with brand names in the USA and the study was adequately powered to detect differences between conditions.
Conclusion
In sum, the study findings contribute to the literature by suggesting that restriction of health-oriented language for combustible cigarette and e-cigarette packaging and promotional materials may be important.
What this paper adds
This study suggests that health-oriented language on combustible cigarette and e-cigarette packs increases intention to purchase.
Health-oriented language on combustible cigarette packs also increased attitudes toward the cigarette brand, increased favourable perceptions of the packaging information and reduced comparative perceived harm of the brand. These effects were not observed for e-cigarette packs.
The effects of this language on another tobacco product (e-cigarettes) are different than the effects on traditional combustible cigarettes, suggesting differential regulation and future research.
The findings suggest that stricter policies and enforcement to limit the use of health-oriented language on combustible cigarette and e-cigarette packages would be recommended.
References
Footnotes
Contributors All authors contributed to the design of this study. AS-J completed all analyses. AS-J and ASLT wrote most of the first draft of the manuscript. KY also wrote one section. All authors participated repeatedly in revising, formatting and conceptualizing future drafts.
Competing interests None declared.
Ethics approval Michigan State University’s Institutional Review Board approved this study.
Provenance and peer review Not commissioned; externally peer reviewed.