Evidence of Youth-Appealing Cigarette Advertising Tactics from a Randomized, Controlled Experiment

Abstract

Introduction

Receptivity to tobacco advertising is an important component in the progression from exposure to advertising to use behavior, yet little is known about current tobacco advertising tactics that increase receptivity. This study tests the effect of three advertising features identified in earlier work as potentially appealing to adolescents and young adults: flora imagery, eco-friendly language, and sweepstakes.

Aims and Methods

We conducted an online survey in which 1,000 US adolescents (age 15–17) and 1,000 US young adults (age 18–24), equally stratified by smoking status, were exposed to three experimental modules manipulating presence/absence of each feature of interest on cigarette ads. After viewing each ad, participants reported their receptivity to the advertisement. Bivariable analysis and multivariable logistic regression analyses were used to assess the effect of each appeal’s presence on receptivity.

Results

Adolescents (aOR = 1.55, 95%CI: 1.12–2.14) and young adults (aOR = 1.43, 95%CI: 1.06–1.93) were more likely to be receptive to ads with flora imagery. The effect of sweepstakes on receptivity was modified by the specific ad for young adults. Ecofriendly language did not significantly impact receptivity among either group. Exploratory subgroup analysis found a significant ad by sweepstakes interaction among young adult noncurrent smokers.

Conclusion

These findings provide evidence that use of flora imagery increases cigarette advertising receptivity among adolescents and young adults, while the effect of sweepstakes on young adults might additionally depend on the ad shown. Restrictions on youth appealing advertising tactics, such as those identified in this study, as well as broader content-neutral advertising restriction policies, should be considered.

Implications

This study provides evidence for the appeal of flora imagery in cigarette advertising for both adolescents and young adults, as well as the complexity of how use of sweepstakes interacts with branding in specific ads to predict receptivity among young adults. Because receptivity to cigarette advertising is a documented step in the pathway between advertising exposure and product use, restrictions on use of these tactics should be implemented, and broader content-neutral advertising restriction policies should be considered.

Introduction

Cigarette advertising is causally associated with youth smoking behavior.^1,2 The tobacco industry targets young people to initiate smoking across developmental stages from childhood to young adulthood.³ Adolescents and young adults who are exposed to higher levels of cigarette advertising are more likely to start and maintain tobacco use.⁴

Receptivity is an important component in the progression from tobacco advertising exposure to use behavior. Receptivity to tobacco advertising is conceptualized as “a successive sequence of attachment to, or involvement in, tobacco-related advertisements.”⁵ This concept stems from the hierarchy-of-effects approach to advertising, in which ad exposure prompts a series of affective and cognitive processes including liking the ad, interest in the product, and intention to use the product, and ultimately can result in product purchase and use.^6,7 Youth with higher levels of tobacco advertising receptivity are more likely to engage in future tobacco use.⁸ Thus, it is crucial to understand what aspects of tobacco advertising increase youth receptivity—in other words, what characteristics of a tobacco advertisement increase its appeal to youth?

Themes relating to popularity and peer acceptance, fun, rebelliousness, and risk-taking have been highlighted as features in tobacco marketing that connect to unique adolescent needs and motivations.⁹ Tobacco industry document research also shows that cigarette companies historically targeted the youth market through advertisements that connected to concepts of independence and autonomy, nonconformity and expression, maturity, and social acceptance and admiration.^10–12 Youth were also targeted via promotional activities such as the offer of free, branded products.¹⁰

However, tobacco advertising features that appeal to youth may extend beyond content that seems intuitively youth-appealing. For example, our prior work assessed which features were common to cigarette ads that were appealing to youth and young adults. This study content analyzed the features of cigarette advertisements, and then assessed response to those ads among adolescents and young adults.¹³ This work found that cigarette ads that contained three features—(1) imagery of flora (eg, plants, flowers), (2) ecofriendly themes (eg, language about how the company protects the environment), and (3) sweepstakes—were more appealing than ads without these features. Sweepstakes are defined as “a marketing or consumer sales promotion which involves the offering of prizes to participants, where winners are selected by chance and no consideration is required.”¹⁴ The current study builds on this work and uses a randomized controlled experiment to identify whether the presence versus absence of each of these three tobacco marketing features affects receptivity to cigarette advertising among adolescents and young adults. We have three main hypotheses:

H1: The presence of flora imagery will increase receptivity to cigarette advertising among (a) adolescents and (b) young adults.

H2: The presence of eco-friendly language will increase receptivity to cigarette advertising among (a) adolescents and (b) young adults.

H3: The presence of sweepstakes will increase receptivity to cigarette advertising among (a) adolescents and (b) young adults.

Methods

Sample/Recruitment

Data were collected from 1,000 US 15–17-year olds (adolescents) and 1,000 US 18–24-year olds (young adults) who completed an online survey between December 2018 and January 2019. This sample size was powered to have ~.90 power to detect a difference between advertising tactics with a standard small effect size of .1¹⁵ using Cramers V. We contracted SSRS (ssrs.com), a survey research firm, to recruit participants from an online national panel. Quotas were used during sampling to ensure equal numbers of smokers and nonsmokers: for adolescent participants, 500 ever smokers and 500 never smokers, and for young adult participants, 500 current smokers who smoked at least 100 lifetime cigarettes and smoked in the past 30 days, and 500 noncurrent smoker who had not smoked at least 100 lifetime cigarettes and had not smoked in the past 30 days. Participants received an email invitation to participate in the study and were compensated with points that are deposited into an account and can then be converted into money.

Procedure

Participants completed an online survey in which they were exposed to three different experimental modules: one manipulating the presence or absence of flora imagery, one manipulating the presence or absence of ecofriendly language, and one manipulating the presence or absence of sweepstakes. Within each module, the pairs of ads were identical, other than the presence or absence of the stimulus of interest. Before completing the survey, participants were presented with an information sheet and indicated their agreement to participate in the study. This study was approved with a waiver of parental consent by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board (IRB00008763). Table 1 illustrates the study’s experimental design. Participants viewed three tobacco ads, one for each experimental module with the tactic either present or absent. Within each module, participants were randomly assigned to view an ad for one of five brands: American Spirit, Camel, Marlboro, Newport and 2000 (a Brazilian brand novel to US consumers).¹⁶ Ads were developed using existing imagery and language from existing tobacco ads. Order of module presentation was randomized, as was the selection of the ad brand and condition (ie, module tactic either present or absent). Manipulated ads used in the experiment are available upon request.

Table 1.

Study Design Structure and Sample Size by Advertising Claim Experimental Module, Tobacco Brand, and Age of Participants

Brand	Age group	Experimental module 1: Sweepstakes		Experimental module 2: Flora imagery		Experimental module 3: Eco-friendly language
		Absent	Present	Absent	Present	Absent	Present
American Spirit	Adolescents	98	109	109	101	96	96
	Young Adults	95	88	84	96	100	101
Camel	Adolescents	108	94	95	91	104	102
	Young Adults	89	98	103	103	92	89
Marlboro	Adolescents	102	86	99	105	101	104
	Young Adults	95	110	97	91	93	93
Newport	Adolescents	96	111	95	104	94	96
	Young Adults	98	82	101	88	101	103
2000a	Adolescent	98	98	103	98	103	104
	Young Adults	98	98	91	97	91	88
Total	Adolescents	502	498	501	499	498	502
	Young Adults	475	476	476	475	477	474

Adolescents include 15–17 year olds; young adults include 18–24 year olds.

2000 is a Brazilian brand of cigarettes that would be novel to an American study sample.

Outcome Measure

The primary outcome for this study was ad receptivity. To assess this, we employed a measure from earlier work.¹³ After viewing each ad, participants were asked how much they (a) liked the ad, (b) were curious about the product in the ad, and (c) were interested in using the advertised product. Liking of the ad was measured with an item adapted from the Population Assessment of Tobacco and Health (PATH) study. Participants were asked how much they liked the ad, with response options of dislike very much (0), dislike (1), neither dislike nor like (2), like (3), and like very much (4). Curiosity and interest were measured using items adapted from the Expanded Susceptibility to Smoke Index,¹⁷ in which participants were asked how much they disagreed or agreed with the statements, “This ad made me curious about the product” and “This ad made me want to use the product”. Response options were strongly disagree (0), disagree (1), somewhat disagree (2), neither agree nor disagree (3), somewhat agree (4), agree (5), and strongly agree (6). We used these items to create a new variable indicating any receptivity to the advertisement. A positive response to any of these three items indicated receptivity to the ad, while all neutral or negative responses indicated no receptivity.¹⁴ These items were combined to capture any level of interest or desire from the AIDA Hierarchy of Effects Model.¹⁸ Internal consistency of outcome items was measured via Cronbach’s Alpha and scale reliability coefficients ranged from 0.80 to 0.82 for the three ad features.

Additional Measures

Participants provided the following sociodemographic information: gender, race/ethnicity, age, parental education, and sexual orientation (for young adults only). We also assessed smoking status based on ever use, past 30-day use and susceptibility to smoking, to create a four-category indicator of smoking status: nonsusceptible never smoker, susceptible never smoker, ever smoker—not in the past 30 days, and ever smoker—smoked in the past 30 days.

Analysis

Forty-nine young adults entered conflicting information regarding their smoking status and were excluded from the analytic sample, leaving a total analytic sample size of 1,951 (500 adolescent never smokers, 500 adolescent ever smokers, 500 young adult past 30-day smokers, and 451 young adult not past 30-day smokers). We assessed randomization of conditions across key participant characteristics (eg, gender, race/ethnicity, age, highest level of parental education, smoking status, sexual orientation) and experimental characteristics (eg, ad position (1st, 2nd, 3rd) and brand of ad shown) (see Supplementary Tables 1A and 1B). Chi square tests were used in bivariable analyses to assess differences in proportion of participants who were receptive to the ads when the different advertising tactics were present or absent. We used multiple logistic regression to calculate adjusted odds ratios of receptivity for each condition (flora imagery, eco-friendly language, and sweepstakes) separately for adolescents and young adults. Models were adjusted for characteristics that were revealed by the randomization check to significantly differ across experimental conditions, as well as prognostic covariates identified a priori.¹⁹ Model 1 included brand of ad viewed, condition order, smoking status, gender, race/ethnicity, age, parental education, and sexual orientation (for young adults only) to control for their association with advertisement receptivity. Model 2 included an advertisement × condition interaction to test if the effect of advertising feature varied by brand of the advertisement. Significant interactions were followed up with marginal estimates of the predicted probability of receptivity for each brand by presence and absence on the tactic. Graphs of these marginal estimates are presented with 95% confidence intervals. Supplemental subgroup analyses were also conducted by age and smoking status. All tests of association were two-sided (p < .05). All analyses were conducted in STATA version 15.²⁰

Results

Sample Demographics

Within the adolescent sample, the average age was 16.3 years, 55.7% identified as female, 44.8% reported identifying as non-Hispanic White, and 40.9% had a parent with at least a four-year college degree. The average age of the young adult sample was 20.9 years; 54.6% identified as female, 49.0% reported identifying as non-Hispanic White, 40.8% had a parent with at least a four-year college degree, and 19.0% identified as lesbian, gay, bisexual or another sexual identity. Table 2 presents participant characteristics.

Table 2.

Demographic Characteristics of Adolescent (15–17 years old) and Young Adult (18–24 Years Old) Participants.

	Adolescents (N = 1000)	Young adults (N = 951)
	n (%)	n (%)
Age, mean (SE)	16.31 (.02)	21 (.07)
Identified as a woman	545 (55.7%)	509 (54.0%)
Identified as a man	434 (49.1%)	434 (46.0%)
Race/Ethnicity
Non-Hispanic White	448 (44.8%)	464 (48.8%)
Non-Hispanic Black	179 (17.9%)	177 (18.6%)
Hispanic	215 (21.5%)	206 (21.7%)
Other	158 (15.8%)	104 (10.9%)
Highest level of parental education
Less than high school degree	57 (6.3%)	39 (4.3%)
High School graduate or equivalent	235 (26.0%)	249 (27.6%)
Some college	242 (26.8%)	248 (27.5%)
A four-year college degree or more	370 (40.9%)	367 (40.6%)
Sexual orientation
Lesbian, Gay, Bisexual, or Another Sexual Identity	--	176 (18.8%)
Heterosexual/Straight	--	758 (81.2%)
Smoking status and intensity
Past 30-day smoker	121 (12.1%)	500 (52.6%)
Not past 30-day smoker	379 (37.9%)	159 (16.7%)
Susceptible never smoker	241 (24.1%)	125 (13.1%)
Not Susceptible never smoker	259 (25.9%)	167 (17.6%)

Tobacco Advertisement Receptivity

Table 3 presents the proportion of participants who reported receptivity to ads by presence versus absence of the three different advertising tactics. Among adolescents, the proportion receptive to the ad with flora imagery (28.66%) was significantly larger than the proportion receptive to the ad without flora imagery (23.00%, p = .041). There were no significant differences in receptivity to ads with and without sweepstakes and ads with and without eco-friendly language for adolescents. Overall, rates of receptivity were higher overall among young adults. A significantly larger proportion of young adults were receptive to ads with sweepstakes (50.11%) compared to ads without sweepstakes (43.13%, p = .038). There were no significant differences in receptivity to ads with and without flora imagery and ads with and without eco-friendly language among young adults.

Table 3.

Proportion of Adolescents (15–17 Years Old) and Young Adults (18–24 Years Old) Receptive to Each Ad by Absence/Presence of Experimental Module Advertising Tactic.

	Adolescents	Young adults
	N (%) Receptive	N (%) Receptive
Experimental module 1: Sweepstakes
Absent	122 (24.40%)	204 (43.13%)
Present	139 (28.02%)	238 (50.11%)
Experimental module 2: Flora imagery
Absent	115 (23.00%)	196 (41.26%)
Present	143 (28.66%)	220 (46.41%)
Experimental module 3: Eco-friendly language
Absent	138 (27.71%)	212 (44.54%)
Present	140 (28.00%)	217 (45.97%)

Bolding indicates that there was a statistically significant difference between proportion who were receptivity when a tactic was present compared to when it was absent at p < .05.

Table 4 presents results of adjusted logistic regression models for each ad tactic. Findings indicate that adolescents had 55% higher odds of receptivity to ads with flora imagery compared to ads without this feature (p = .008); 95% confidence intervals indicate a range of odds of increased receptivity between 12% and 114%. Young adults had 43% higher odds of receptivity to ads with flora imagery compared to ads without this feature (p = .02); 95% confidence intervals indicate a range of odds of increased receptivity between 6% and 93%. Results similarly suggest that adolescents and young adults similarly had higher odds of receptivity to ads containing sweepstakes. Compared to those who viewed an ad without sweepstakes, adolescents who viewed an ad containing sweepstakes were 37% more likely to be receptive (p = .057). However, 95% confidence intervals indicate that the effect of sweepstakes on likelihood of receptivity ranges from a 1% decreased odds of receptivity to an 88% increased odds of receptivity. Young adults who viewed an ad containing sweepstakes, compared to those who viewed an ad without sweepstakes, were 29% more likely to be receptive (p = .091). However, 95% confidence intervals indicate that the effect of sweepstakes on likelihood of receptivity ranges from a 4% decreased odds of receptivity to a 74% increased odds of receptivity. Presence of eco-friendly language is unlikely to affect receptivity. Adolescents who viewed an ad containing eco-friendly language were only 97% as likely as those who viewed an ad without eco-friendly language to be receptive (p = .873); 95% confidence intervals indicate the range of this effect was between 71% and 133%, thus failing to offer support for the hypothesis that eco-friendly language increases receptivity. Young adults who viewed an ad containing eco-friendly language were 17% more likely to be receptive to the ad, compared to those who viewed an ad without eco-friendly language (p = .313); 95% confidence intervals indicated the range of this effect was between 86% and 158%, similarly failing to offer support for the hypothesis that eco-friendly language increases receptivity.

Table 4.

Adjusted Logistic Regression Models^a of Ad Receptivity Among Adolescents (15–17 Years Old) and Young Adults (18–24 Years Old) Stratified by Experimental Module Advertising Tactic.

	Adolescents		Young adults
	Model 1: sweepstakes aOR (95% CI)	Model 2: sweepstakes aOR (95% CI)	Model 1: sweepstakes aOR (95% CI)	Model 2: sweepstakes aOR (95% CI)
Experimental condition (ref: absent)
Condition present	1.37 (0.99–1.88)	0.86 (0.39–1.90)	1.29 (0.96–1.74)	1.92 (0.98–3.77)
Brand of ad viewed (ref: 2000 b)
American Spirit	1.81 (1.07–3.08)	1.49 (0.71–3.12)	1.81 (1.13–2.89)	2.81 (1.43–5.51)
Camel	1.90 (1.13–3.22)	1.07 (0.50–2.25)	1.90 (1.18–3.04)	2.13 (1.06–4.27)
Newport	1.04 (0.60–1.80)	0.71 (0.32–1.61)	1.18 (0.74–1.89)	1.00 (0.51–1.91)
Marlboro	3.14 (1.87–5.27)	3.08 (1.53–6.20)	2.55 (1.61–4.05)	4.73 (2.37–9.45)
Condition × Ad interactionc
American Spirit & condition present	–	1.52 (0.54–4.33)	–	0.42 (0.16–1.07)
Camel & condition present	–	3.17 (1.10–9.14)	–	0.80 (0.31–2.07)
Newport & condition present	–	2.06 (0.68–6.26)	–	1.45 (0.56–3.76)
Marlboro & condition present	–	1.00 (0.36–2.80)	–	0.32 (0.12–0.81)
Wald test for interaction	–	7.28 (p = .1217)	–	13.42 (p = .0094)

	Model 1: Flora aOR (95% CI)	Model 2: Flora aOR (95% CI)	Model 1: Flora aOR (95% CI)	Model 2: Flora aOR (95% CI)
Experimental condition (ref: absent)
Condition present	1.55 (1.12–2.14)	1.44 (0.60–3.46)	1.43 (1.06–1.93)	1.64 (0.77–3.50)
Brand of ad viewed (ref: 2000 b)
American Spirit	3.82 (2.22–6.55)	3.36 (1.56–7.25)	4.37 (2.62–7.27)	4.60 (2.13–9.92)
Camel	2.14 (1.20–3.79)	1.70 (0.74–3.91)	4.04 (2.48–6.57)	5.06 (2.47–10.34)
Newport	1.96 (1.12–3.44)	1.66 (0.72–3.84)	3.99 (2.41–6.61)	3.53 (1.69–7.37)
Marlboro	2.34 (1.34–4.10)	3.27 (1.48–7.23)	5.21 (3.15–8.61)	6.23 (3.00–12.95)
Condition x Ad interactionc
American Spirit & condition present	–	1.31 (0.44–3.84)	–	0.91 (0.33–2.55)
Camel & condition present	–	1.54 (0.49–8.86)	–	0.651 (0.24–1.71)
Newport & condition present	–	1.34 (0.43–4.19)	–	1.31 (0.48–3.62)
Marlboro & condition present	–	0.53 (0.17–1.62)	–	0.71 (0.26–1.92)
Wald test for interaction	–	5.54 (p = .2364)	–	2.89 (p = .5767)

	Model 1: Eco-friendly aOR (95% CI)	Model 2: Eco-friendly aOR (95% CI)	Model 1: Eco-friendly aOR (95% CI)	Model 2: Eco-friendly aOR (95% CI)
Experimental condition (ref: absent)
Condition present	0.97 (0.71–1.33)	1.15 (0.55–2.41)	1.17 (0.86–1.58)	2.39 (1.15–4.95)
Brand of ad viewed (ref: 2000 b)
American Spirit	1.93 (1.17–3.17)	2.39 (1.17–4.87)	2.60 (1.60–4.22)	4.56 (2.26–9.20)
Camel	2.09 (1.29–3.40)	2.70 (1.36–5.35)	2.45 (1.49–4.03)	2.91 (1.44–5.87)
Newport	1.01 (0.59–1.73)	0.96 (0.44–2.09)	1.15 (0.71–1.87)	1.83 (0.90–3.73)
Marlboro	1.2 (0.74–2.03)	1.09 (0.53–2.26)	2.81 (1.71–4.62)	4.81 (2.35–9.86)
Condition × Ad interactionc
American Spirit & condition present	–	0.65 (0.24–1.77)	–	0.32 (0.12–0.86)
Camel & condition present	–	0.59 (0.22–1.56)	–	0.75 (0.27–2.06)
Newport & condition present	–	1.10 (0.38–3.20)	–	0.40 (0.15–1.07)
Marlboro & condition present	–	1.26 (0.46–3.49)	–	0.34 (0.13–0.93)
Wald test for interaction	–	3.76 (p = .4397)	–	8.02 (p = .0907)

Bolding indicates significance at p < .05; aOR = adjusted Odds Ratio; 95%CI = 95% Confidence Interval.

Results are adjusted for brand of ad viewed, condition order, smoking status, gender, race/ethnicity, age, parental education, and sexual orientation (for young adults only). Full results are available in Supplementary Table 2. Model 1 does not include Condition by Brand interaction.

2000 is a Brazilian brand of cigarettes that would be novel to an American study sample.

Interaction estimates are ratios of odds ratios that represent the magnitude and significance of effect measure modification.

Additionally, there was a significant interaction between presence of sweepstakes and brand of ad seen among young adults (See Figure 1). The presence of sweepstakes had a diminished effect when used in the Marlboro ad as when compared to the 2000 ad, such that the Marlboro ad with sweepstakes had a lower probability of receptivity than Marlboro ad without sweepstakes, whereas sweepstakes presence increased the appeal of the 2000 ad. Additionally, the Newport ad that contained sweepstakes had higher probability of receptivity than the Newport ad without sweepstakes.

Figure 1.

Marginal estimates of the interaction effect between brand of ad and presence of sweepstakes among young adults.

Exploratory subgroup analyses by age and smoking status (see Supplementary Table 3A and 3B for full results) were generally aligned with the main findings. For example, adolescent never smokers had significantly higher odds (aOR = 1.92) of receptivity to ads with flora imagery compared to ads without flora imagery, and there was a significant brand by sweepstakes interaction among young adult ever smokers who had not smoked in the past 30 days.

Discussion

There are several relevant findings from this study that can inform tobacco industry advertising restrictions. Most notably, the presence of flora imagery in advertisements increased ad receptivity among adolescents (supporting H1a) and young adults (H1b). These findings strengthen previous evidence of the appeal of flora imagery in cigarette advertising among adolescents.¹³ There was also a marginally significant relationship between sweepstakes and increased receptivity among adolescents and young adults, offering some support for H3a and H3b. The effect of sweepstakes on receptivity for young adults may also be dependent upon the specific ad viewed. Use of sweepstakes and flora imagery are common in tobacco advertising²¹ and are not necessarily features that would be intuitively identified as appealing to young people.

The relationship between flora imagery and increased ad receptivity among both adolescents and young adults is particularly concerning. Imagery has been a major feature of tobacco advertising for decades. As early as the 1950s, researchers and trade analysists advised the tobacco industry to rely on imagery and ad copy to imply their products were healthier, rather than explicit health claims, as more subtle cues could reduce the likelihood that consumers would be reminded of the health allegations facing tobacco.²² As a component of the complex advertising system that imbues products with symbolic meanings, flora imagery can be used to signify meaning beyond its surface presention,²³ for example, by potentially indicating a product is “natural” or less harmful. Positive feelings about flora imagery can then diffuse to the product itself, affecting product perceptions and user behaviors.²⁴ Additionally, the recent increase in eco-conscious or “green” consumers and their increased susceptibility to the persuasive effects of natural imagery in advertising for “nongreen” produts,²⁵ may help to explain why this advertising tactic could be effective in tobacco ads.

The appeal of sweepstakes to adolescents is also of concern, as prior work found this tactic was used in nearly 30% of cigarette advertisements.²¹ In addition, exposure to tobacco advertisements with sweepstakes is disproportionately higher among youth of lower socioeconomic status compared to youth of higher socioeconomic status.²⁶ Additionally, sweepstakes serve an instrumental function beyond their ability to increase advertising receptivity, as they are commonly used to obtain information about consumers or potential consumers which can facilitate future targeted marketing.^27,28 Although tobacco companies may restrict participation in sweepstakes to individuals aged 21 and over, this study’s findings indicate that even without participation, sweepstakes may nonetheless attract youth.

We did not observe a significant effect of eco-friendly language on receptivity for adolescents or young adults. Prior work has found that eco-friendly language contributes to reduced product harm perceptions,^29,30 so although this study did not find an overall effect on ad receptivity, it is nonetheless still a tactic worth monitoring. It is possible that eco-friendly language was not integrated well-enough into the selected advertisements and the potential discordance between the insertion of eco-friendly tactics and the overall message or design of the ad resulted in null effects.

There are several policy implications regarding these findings. The Family Smoking Prevention & Tobacco Control Act grants FDA the authority to regulate the advertising, marketing, and promotion of tobacco products,³¹ and in particular, marketing that appeals to adolescents and children.³² Although the FDA has previously restricted advertising practices, such as banning the use of tobacco brand names in sponsorship of sporting, entertainment, and cultural events,³³ cigarette advertising regulation is not listed as part of the FDAs current Comprehensive Plan for Tobacco and Nicotine regulation while bans on adolescent targeted advertising for other tobacco products, like electronic nicotine delivery systems are included.³⁴ The FDA could make cigarette advertising a priority by restricting the use of specific youth-appealing advertising features, such as those identified in this research, accompanied by an evidence-based description of features that have been empirically shown to increase ad receptivity among young people.

However, banning specific advertising tactics may not do enough to prevent tobacco companies from creating advertisements that appeal to youth, particularly because the tobacco industry may quickly pivot to replace restricted tactics with other similar, not yet restricted tactics. For example, when FDA restricted use of the terms “natural” and “additive-free” to market American Spirit cigarettes,³⁵ research on Natural American Spirit advertising before and after the implementation of this agreement provided evidence that such specific restriction were insufficient and the brand was still able to communicate reduced harm using alternative terminology, including “Tobacco and Water” and “Real. Simple. Different.”³⁶ Similarly, banning the specific terms “light,” “mild,” and “low” resulted in an industry pivot toward using different color to communicate those same concepts of reduced harm.^37–39 Even if FDA were willing to prohibit the use of flora imagery and sweepstakes in advertising, the immense resources of the tobacco industry could likely be used to identify additional features that could have the same effect while still following the restrictions.

As such, there are several alternative regulations on cigarette advertising that are likely to be more effective and feasible to enforce and implement. One alternative is to follow the precedent set by local jurisdictions (eg, La Mesa, California, Long Beach, New York), and enact content neutral advertising bans, which place restrictions on all types of advertising regardless of content and place limits on the amount, time, location or manner of advertisements and could result in lower adolescent exposure to all tobacco advertisements, regardless of feature.^40,41 Additionally, restricting tobacco retailer licensing or tobacco outlet density could also reduce advertising exposure and avoid First Amendment issues that may result in tobacco industry litigation on advertising content restrictions.^41,42 Counter-marketing campaigns may be a useful supplement to any single or combination of suggested policies. Such interventions could aim to inoculate adolescents and young adults against strategies used by the industry that research has identified as increasing their receptivity to tobacco producing, including the use of flora imagery and sweepstakes in ads.

Limitations

First, data for this study were collected prior to the December 2019 national Tobacco 21 regulation. As such, the sampling plan and quotas do not account for the federal increase in the legal age of sale from 18- to 21 years old. Because the advertisements used in this study were first run prior to the national Tobacco 21 regulation (when it was legal for 18 year olds to purchase tobacco and thus permissible for tobacco companies to market to this group), when conducting the analysis we accordingly split our age groups between those aged 15–17 and those aged 18–24 (those allowed to purchase cigarettes at the time the ad was first run). Second, while the concept of receptivity is useful, it does not comprehensively capture the range of advertising effects.⁴³ It is possible that the advertising tactics investigated in this study may engender additional positive responses among youth not captured via our outcome measure. In particular, future research should assess behavioral response to the ad in addition to perceptions of ad effectiveness (see Noar et al., 2019 for a review of perceived and actual message effectiveness measures). The number of statistical tests run to assess interactions may also have increase the likelihood of a type one error, and significant interaction findings should be taken as exploratory. Finally, this study exposed participants to only one ad per brand which does not capture the on-going relationship between consumer and brand.⁴³ The manipulation of ads to remove the presence of each ad feature of interest resulted in ads that may have appeared incomplete to some participants. The decision not to use alternative language to fill blank space was driven by the goal of isolating the effects of only the presence or absence of the ad features. This would not have been possible if alternative text or imagery had been used, as it could have been responsible for observed effects. Future research should investigate these ad features against a variety of alternative features. Maximum Difference Scaling might be well-suited for this purpose.⁴⁴ Strengths of this study include its randomized controlled design, use of a novel brand, and focus on youth.

Conclusion

Results from this study provide evidence that the presence of flora imagery in cigarette advertisements increase receptivity to the advertisements among youth and young adults, while the presence of sweepstakes has a marginal positive effect on receptivity among these groups. Limiting the ability of the tobacco industry to use these advertising tactics could help decrease the appeal of cigarettes to youth and young adults.

Supplementary Material

A Contributorship Form detailing each author’s specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.

Funding

This work was funded by the National Institute on Drug Abuse and Food and Drug Administration Center for Tobacco Products (K01DA037903; PI: Moran). CW’s effort is supported by the National Cancer Institute (T32 CA009314; PI: Platz). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or Food and Drug Administration.

Declaration of Interests

MBM has served as a paid expert witness in litigation sponsored by the Public Health Advocacy Institute against RJ Reynolds. This arrangement has been reviewed and approved by the Johns Hopkins University in accordance with its conflict of interest policies.

Data Availability

Data may be made available upon request to the study contact author.