Vaping Flavors and Flavor Representation: A Test of Youth Risk Perceptions, Novelty Perceptions, and Susceptibility

Abstract

Introduction

Whether novelty-flavored vaping devices should be available in the marketplace has been a hotly contested debate. From one perspective, the variety of different flavors, such as fruit and mint, may help adult cigarette smokers who are seeking to switch to reduced-harm nicotine products. However, these flavors are also wildly popular among youth, creating concerns about new nicotine product use among minors.

Aims and Methods

This experiment (n = 176) tests whether vaping flavors (tobacco vs fruit) and flavor representations on packages (flavor color, flavor image) influence how middle school youth perceive vaping products.

Results

While results show no difference in risk perceptions based on condition, novelty perceptions (eg, how fun, interesting) and susceptibility to vaping are highest among those who view the fruit-flavored vaping product with flavor color and flavor image. Those who viewed this condition reported higher novelty perceptions and susceptibility than those who viewed the fruit-flavored vaping product with no flavor color and no flavor image. Additionally, they reported higher novelty perceptions than those who viewed the tobacco-flavored vaping product with flavor color and flavor image. A post-hoc analysis in supplemental data shows that youth who report lower risk perceptions and higher susceptibility have higher behavioral intentions to vape in the next year.

Conclusion

Findings suggest that restricting flavor representation on packaging might reduce how fun and interesting youth perceive these products to be and how susceptible they are to using them.

Implications.

While there is debate over the regulation of flavored vaping products, this study suggests that restricting the use of the flavor color and flavor image on the packaging is associated with lower levels of novelty perceptions and susceptibility. This might be an alternative approach to minimize the appeal of vaping products among youth.

Introduction

In 2020, 4.7% of middle school youth (550 000) reported vaping in the past 30 days.¹ Youth rate fruit-flavored vaping devices as more preferable² and less risky than tobacco-flavored ones,^3–5 and they report they are more likely to try a fruit-flavored device.^3,6 This issue is timely, as the FDA has started to issue marketing denial orders for vaping products and has acknowledged that balancing the benefits to adult smokers and the risk to youth are part of the important considerations being made.⁷ Flavor representations on vaping packages, such as flavor color or flavor image, might influence how youth perceive vaping products.

In recent years, there have been efforts to restrict flavored electronic nicotine products because they are attractive to youth,⁸ however, flavors might also promote complete switching in adult smokers who are transitioning away from cigarettes,⁹ thereby resulting in harm reduction although not harm elimination. One question we might ask in relation to this tension between preventing youth uptake and facilitating cessation of smoking in adult smokers is whether restricting just the packaging itself, rather than the product, might reduce youth interest in the product, particularly as the U.S. Food and Drug Administration has been deemed the authority to regulate the marketing of these products.¹⁰

In this brief report, we present pilot data of an experiment conducted with middle school youth (ages 11–14 years). We varied the flavor and flavor representation on the vaping packages that youth viewed, and we measured their risk perceptions, novelty perceptions (eg, how fun, interesting), susceptibility to vaping (eg, want to try, be curious about), and behavioral intentions (try vaping in the next year?). This study follows a prior project with high school youth on perceptions of electronic nicotine product packages and flavors (tobacco vs fruit) that was conducted prior to the proliferation of nicotine salt pods and disposable flavored products.¹¹ In this earlier study, we found that youth who have tried vaping in the past had lower risk perceptions, higher novelty perceptions, and higher intentions to vape soon, while a fruit-flavored product (in comparison to a tobacco-flavored product) did not influence these key dependent measures.¹¹ Some questions our team had following that earlier study was whether middle school youth, with higher rates of never-vapers, would be influenced by flavor and flavor presentation and whether findings would be different for the current disposable, flavored products. It is essential to investigate the perceptions of middle school youth, as the rates of vaping increase tremendously as youth transition from middle school to high school.¹ The aim of this current study was to test the following hypotheses with middle school youth.

Hypothesis 1: Participants who view a tobacco-flavored vaping product with a flavor color and flavor image on the package will report (H1a) higher risk perceptions, (H1b) lower novelty perceptions, and (H1c) lower susceptibility than participants who view a fruit-flavored vaping product with a flavor color and flavor image on the package. Hypothesis 1 addresses the impact of the vaping product’s flavor on youth’s perceptions and susceptibility.

Hypothesis 2: Participants who view a fruit-flavored vaping product with a flavor color and flavor image on the package will report (H2a) lower risk perceptions, (H2b) higher novelty perceptions, and (H2c) higher susceptibility than participants who view a fruit-flavored vaping product without a flavor color or flavor image on the package. Hypothesis 2 addresses the impact of the vaping product’s flavor representation on youth’s perceptions and susceptibility.

Methods

An experiment (parent–child panel data, middle school youth), approved by the University of Minnesota IRB and pre-registered through As Predicted, (https://aspredicted.org/blind.php?x=MBV_62B), was conducted to compare vaping flavor (tobacco vs fruit) and flavor presentation (flavor color, flavor image) on the variables of risk perceptions, novelty perceptions, and susceptibility. Hypotheses H1a, H1b, and H2b were pre-registered. The final sample included 176 middle school youth and their parents after participants were excluded for not completing the study in one sitting (n = 14) and not being 11–14 years old (n = 20). To reduce social desirability concerns, we told youth that their answers were a secret, and we would not share their responses with their parents.¹² Participants received points that equal monetary value from Qualtrics after completing the study.

Participants

Participants were recruited using Qualtrics’ national, quota-based sampling approach, and participants were invited to participate from standing online panels that work with Qualtrics Research Services. We requested that Qualtrics arrange segmentation by geographical distribution for a national sample (Midwest: 21.33%; Northeast: 18.02%; South: 37.27%; West: 23.38%), parental race (White: 60%; BIPOC [Black, Indigenous, and people of color]: 40%), youth grade (6th, 7th, and 8th), and youth sex (50% male/female). Among the youth participants, 91 participants (51.7%) defined themselves as female, and the average age was 12.4 years (S.D. = 1.0). There was an approximately equal distribution by grade. Youth conveyed their tobacco use status: tried cigarettes but not in the past 30 days (n = 21, 11.9%), past 30-day cigarette use (n = 15, 8.5%), tried vaping but not in the past 30 days (n = 20, 11.4%), and past 30-day vaping (n = 19, 10.8%). For a complete breakdown of the participants, see Supplementary Table 1 (parents) and Supplementary Table 2 (youth).

Procedure

Parents were invited to participate with their child in a study to investigate how to communicate risks associated with vaping, and they were informed that their child would view vaping packages. They completed demographic, tobacco use, and tobacco perception questions for themselves and viewed a debriefing. Next, parents were asked to hand the internet-enabled device to their child and to give their child privacy. Youth were asked to provide child assent and complete demographic and tobacco use questions. The youth were then randomly assigned to view one of the four flavor/flavor presentation conditions described below. Youth responded to dependent measures, including risk and novelty perceptions and susceptibility. Youth were then debriefed and provided with resources from the CDC, the FDA’s Real Cost Campaign, the Truth Initiative, and Smokefree.gov. They were informed that vaping is “unsafe for kids, teens, and young adults” and that “nicotine is addictive” and can “harm adolescent brain development.”¹³ A licensed fifth-grade teacher edited the youth materials for reading level. All youth confirmed that they responded to the survey honestly and correctly responded to an attention check question.

Stimuli

At the time of this study, disposable vaping products typically featured flavor color on the package, and the flavor images were either directly on the box (as represented in our stimuli) or visually surrounding the product on the websites. Participants each viewed three brands of vaping box fronts (in the order of Puff Bar, Viigo, and Posh), with the FDA warning label, in their assigned condition. We selected Puff Bar, as the leading brand, and it is important to note that the stimuli were created before Puff Bar moved to using tobacco-free nicotine. The other brands were selected because of their similarity to Puff Bar and because they were marketed alongside Puff Bar on websites with similar visual features. Each image stayed on the screen for a minimum of 15 seconds before an option to advance the survey was provided. As shown in Supplementary Figure 1, participants were randomized to one of four conditions: 1. tobacco with flavor color and flavor image on the package, 2. fruit flavor with no flavor color and no flavor image, 3. fruit flavor with flavor color and no flavor image, and 4. fruit flavor with flavor color and flavor image (reference condition).

Measures

Risk Perceptions

We averaged participants’ responses to three items (pre-registered) measured on a 4-point scale (1 = definitely not risky, 2 = probably not risky, 3 = probably risky, 4 = definitely risky): How risky is vaping?, Are the products you viewed risky for a middle school student’s health?, and How much do you think middle school students harm themselves when they use the products you viewed? (M = 3.56, SD = .58, α = .83).

Novelty Perceptions

We averaged participants’ responses to four items (pre-registered) on a 5-point scale (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, 5 = strongly agree). We asked whether they thought middle school students would think the products are: fun, interesting, tasty, and like gum or candy (M = 3.04, SD = 1.20, α = .90).

Susceptibility

We averaged participants’ responses to three items measured on a 4-point scale (1 = definitely not, 2 = probably not, 3 = probably yes, 4 = definitely yes): I think students in middle school would want to try the products I viewed, I think students in middle school would be curious about the products I viewed, and I think students in middle school would be interested in trying the products I viewed (M = 2.80, SD = .91, α = .93).

Behavioral Intentions

To measure behavioral intentions to vape, we asked participants, Do you think you will try vaping nicotine in the next year? Participants responded on a 4-point scale, 1 = I definitely will not try; 2 = I probably will not try; 3 = I probably will try; 4 = I definitely will try (M = 1.68, SD = .94).

Statistical Analysis

We fit linear regressions with robust standard errors with the vaping package condition as the predictor of interest and risk perceptions, novelty perceptions, and susceptibility as the dependent measures. We controlled for the following variables: parent past 30-day smoking (yes/no), parent past 30-day vaping (yes/no), youth self-reported sex (male/female), and youth vaping (never/not recently/past 30 days). While the general analysis plan was pre-registered, the covariates were added post-hoc to increase the precision of the analysis, as well as estimate the associations between participant characteristics and the dependent measures.¹⁴P-values for the pairwise mean differences in outcomes between the fruit flavor with flavor color and flavor image condition and the other three vaping package conditions were adjusted using Holm’s procedure to control the family-wise error rate for each of the three dependent measures. For the post-hoc analysis, we fit a linear regression with robust standard errors with risk perceptions, novelty perceptions, and susceptibility as the predictors and behavioral intentions as the dependent measure, with adjustment for the randomized condition and covariates specified above.

Results

Risk Perceptions

We hypothesized that vaping flavor (H1a) and package flavor representation (H2a) would influence risk perceptions, but there was no significant difference in risk perceptions related to the viewed vaping package condition (p = .87). Instead, risk perceptions had a ceiling effect, with the lowest risk perception estimated marginal mean at 3.51/4 (Supplementary Figure 2). However, youth vaping status was associated with risk perceptions (p = .002). Compared to never vapers, past 30-day vapers had 0.43 points lower mean risk perceptions (95% confidence interval [CI]: 0.10–0.76 points lower, p = .01) and previous (but not past 30 days) vapers had 0.45 points lower mean risk perceptions (95% CI: 0.12–0.78 points lower, p = .01).

Novelty Perceptions

Vaping package condition influenced novelty perceptions (p = .03). Participants who viewed the fruit-flavored vaping packages with flavor color and flavor image had 0.59 points higher novelty perceptions (95% CI: 0.13–1.06 points higher, p = .04) than those who viewed the tobacco-flavored packages with flavor color and flavor image (H1b supported), 0.62 points higher novelty perceptions (95% CI: 0.14–1.10 points higher, p = .04) than those who viewed the fruit flavor with no flavor color and no flavor image (H2b supported), and 0.41 points higher novelty perceptions (95% CI: 0.04 points lower to 0.86 points higher, p = .08) than those who viewed the fruit flavor with flavor color and no flavor image (Figure 1). Additionally, males had 0.41 points higher novelty perceptions (95% CI: 0.07–0.75 points higher, p = .02) than females, and those whose parents were past 30-day vapers had 0.53 points higher novelty perceptions (95% CI: 0.02–1.04 points higher, p = 0.04) than those whose parents were not. Lastly, compared to never vapers, past 30-day vapers had 0.82 points higher novelty perceptions (95% CI: 0.38–1.26 points higher, p < .001) and previous (but not past 30 days) vapers had 0.81 points higher novelty perceptions (95% CI: 0.35–1.26 points higher, p < .001).

Figure 1.

Estimated marginal mean novelty perceptions (with 95% CIs) by flavor and flavor representation.

Susceptibility

Participants who viewed fruit-flavored packages with flavor color and flavor image scored 0.33 points higher on susceptibility (95% CI: 0.05 points lower to 0.70 points higher, p = .18, not significant) than those who viewed tobacco-flavored packages with flavor color and flavor image, 0.52 points higher (95% CI: 0.12–0.93 points higher, p = .03) than those who viewed fruit-flavored packages without flavor color or flavor image (H2c supported), and 0.24 points higher (95% CI: 0.13 points lower to 0.61 points higher, p = .20, not significant) than those who viewed fruit-flavored packages with flavor color and no flavor image (Figure 2). Additionally, compared to never vapers, past 30-day vapers reported mean susceptibility scores that were 0.56 points higher (95% CI: 0.13–0.98 points higher, p =.01) and previous vapers (not past 30 days) reported 0.35 points higher susceptibility, although non-significant (95% CI: 0.14 points lower to 0.85 points higher, p = .16). Lastly, those whose parents were past 30-day vapers scored 0.37 points higher on the susceptibility scale (95% CI: 0.03–0.72 points higher, p = .04) than those whose parents were not.

Figure 2.

Estimated marginal mean susceptibility (with 95% CIs) by flavor and flavor representation.

The results comparing the mean outcomes between the vaping package conditions from the analyses without adjustment for covariates were very similar to those from the analyses presented above (Supplementary Figure 3). In a post-hoc analysis, we found that risk perceptions and susceptibility were associated with behavioral intentions, with higher risk perceptions associated with lower intentions to vape and higher susceptibility associated with higher intentions to vape (Supplementary Figure 4).

Discussion

Review of Findings

In this study, we tested whether risk perceptions, novelty perceptions, and susceptibility differ depending upon whether participants viewed tobacco- or fruit-flavored vaping products and whether they viewed a package with flavor color and a flavor image. While controlling for covariates, we found that risk perceptions were the same regardless of which flavor and flavor representation participants viewed. There was a ceiling effect for risk perceptions, such that these middle school participants in all conditions, on average, rated the product as between probably risky and definitely risky.

Participants who viewed fruit-flavored packages with flavor color and a flavor image reported higher novelty perceptions than those who viewed tobacco-flavored packages with flavor color and a flavor image. This finding suggests that middle school youth view fruit flavors as more fun and interesting than tobacco flavors.

Additionally, those who viewed fruit-flavored vaping packages with the flavor color and flavor image had the highest novelty and susceptibility ratings. Novelty perceptions were higher among males, those whose parents were current vapers, and youth who were current and previous vapers. Susceptibility was higher among youth who were current vapers and whose parents were current vapers. Our post-hoc analysis showed lower risk perceptions and higher susceptibility were associated with higher intentions to vape.

Interestingly, in our previous study with high school youth, which was conducted prior to the proliferation of nicotine salt pods and disposable-flavored products, before reports of EVALI, and before the focus on electronic nicotine products by The Real Cost Campaign,¹⁵ we did not find this ceiling effect on risk perceptions, and the product flavor did not have a direct influence on risk or novelty perceptions.¹¹ It is hard to parse whether the differences in this current study are due to increased public health education on the risks of vaping for youth or the focus on middle school youth, but it is useful to consider how this information can be harnessed to reduce uptake among youth.

Limitations and Future Research

It is important to acknowledge limitations associated with this study. Our risk perception and novelty perception measures were adapted from our previous research with high school youth,¹¹ and our susceptibility scale used measures from Pierce et al. and Seo et al.^16,17 However, in this current study we asked about how a “middle school student” would perceive the product rather than asking participants to reflect on themselves or how “one” might perceive them. We made this intentional choice because in our previous research we found that youth are bounded by their own behavioral consideration set, answering based on whether they have previously used electronic nicotine products.¹¹ With vaping behavior, many middle schoolers might not have it as part of their self-schema, and thus, the advantage of asking self-referential questions diminishes.¹⁸ By asking them to interpret what middle school students think, we are able to obtain a response outside of this self-referential prism. However, when we asked youth about their behavioral intentions, we were directly seeking to measure a self-referential concept in relation to their own behavioral plans rather than their product perceptions, and therefore, we asked them directly about their own plans to vape in the next year. Future research on this topic should investigate risk perceptions in different ways. For example, recent literature has identified 10 best practices for measuring cigarette risk perceptions among adult smokers and nonsmokers,¹⁹ and many of these strategies can be applied to studies of other tobacco products and youth. The purpose of preregistration is to clarify for the reader any ways in which the final presentation of the data differs from the initial plans for research, and changes were made in order to enhance clarity or add additional context to the findings.

Tobacco Regulatory Implications

As explained above, the packages with flavor color and flavor image are most like those in the marketplace, although in some cases the flavor image surrounds the package in the online presentation of the product rather than directly on the box. This study suggests that limiting the flavor options or restricting the packaging might reduce how fun and interesting youth perceive these products to be and might reduce their interest in trying them. As mentioned above, there has been some concern in recent years about the importance of balancing the needs of adult smokers who are seeking these flavored products to support cessation from combustible cigarettes with the concern that youth are attracted to these flavors. This study suggests that restricting the use of the flavor color and flavor image on the package might be an alternative approach to minimize the appeal of vaping products among youth, while still keeping the products available in the marketplace for adults seeking to use them in place of cigarettes. Manufacturers may object to packaging regulations, and the FDA may need to demonstrate that they are applying the least restrictive regulation to achieve the intended public health benefit.²⁰ This study demonstrates that restricting both the flavor image and the flavor color is necessary to significantly lower novelty and susceptibility among youth.¹⁰ That said, additional research with adult smokers is needed to determine how they would respond to variations in flavor presentation on electronic nicotine products and whether it would influence whether or not they are likely to view these products as alternatives to cigarettes. Future research should also continue to investigate other approaches not tested in this study, such as restricting the descriptors of flavors and selling these and other products in adult-only brick-and-mortar shops.^21,22

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

Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health and the FDA Center for Tobacco Products (CTP) under Award Number R21CA246602. Research reported in this publication was also supported by NIH grant P30CA077598 utilizing the Biostatistics and Bioinformatics Core shared resource of the Masonic Cancer Center, University of Minnesota and by the National Center for Advancing Translational Sciences of the National Institutes of Health Award Number UL1-TR002494. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Food and Drug Administration.

Declaration of Interests

There are no conflicts of interest to report.

Data Availability

The dataset analyzed for this study is available from the corresponding author on reasonable request.