Changes in Tobacco Product Advertising at Point of Sale: 2015–2018

In late spring 2015, research staff visited all licensed tobacco retailers operating within a half-mile buffer of 41 New Jersey high schools that were randomly selected for the 2014 New Jersey Youth Tobacco Survey.¹⁴  A half-mile (2640 ft) buffer around high schools was chosen on the basis of the premise that it would be easy walking distance for students, and this has been used in other studies of tobacco retailer density around schools.¹¹  Of the 41 schools, 26 had at least 1 tobacco retailer operating within the half-mile buffer. Data collection was repeated annually at the same tobacco retailers each spring. This study covers a 4-year period: 2015 through 2018.

Although the initial sample in 2015 was 194, stores were dropped from the sample over time because of store closures and change of ownership to nontobacco retailers (n = 31).¹⁵  Additionally, for this study, given that liquor stores are age restricted, another 22 stores were dropped, resulting in a final analytic sample of 141 stores for which there are annual data points between 2015 and 2018. On the basis of their observations in the field, research staff categorized the licensed tobacco retailers by store type, including chain convenience stores, nonchain convenience stores, gas kiosks, dollar stores, drug stores (such as chain pharmacies), and others. For analyses focused on interior advertising, we excluded gas kiosks (n = 12) because the nature of these retailers generally does not allow for interior advertising; thus, the analytic sample for examining interior advertising was 129. There was a minimal number of interior advertisements in gas kiosks (Table 1), and we recoded them as exterior advertisements because they could be seen from the outside and functioned as such.

Trained raters working independently collected detailed information on product advertising. This information was unobtrusively recorded on cell phones by using Qualtrics survey software. There was a high level of interrater reliability denoted by Cohen’s κ statistic, ranging from 0.8 to 1.0 across all years and categorical measures. More detail regarding data collection methods is available elsewhere.^14,15 

We examined 3 indicators of tobacco advertising for 4 different tobacco products: cigarettes, cigars, smokeless tobacco, and e-cigarettes. Advertisements were defined as industry-made signs that included a company’s logo and/or image of the product. Only advertisements larger than the size of an index card (3 × 5 in) were counted to ensure that they were visible to patrons. The first indicator was observed presence of any tobacco advertisement by product type (yes or no), including whether a flavored product (eg, menthol for cigarettes) was advertised for each type of product. The second was observed tobacco advertising volume by product type (ie, mean number of advertisements across all stores). As the last indicator, we calculated for all products their share of advertising voice (SAV),¹⁹  or the proportion of advertising for each product type, compared with other products. SAV was calculated by dividing the number of advertisements for a specific tobacco product by the total number of advertisements for all tobacco products. We also report “none” because some retailers stopped advertising over time. All 3 measures were assessed for advertising on both a store’s exterior and interior.

Store type was recoded into the following categories: chain convenience stores, nonchain convenience stores, gas kiosks (only for analyses of exterior advertising), and other. Neighborhood poverty was determined by the percentage of students at the nearby high school who were eligible for the free and reduced-price lunch program: ≥75% was a high-poverty neighborhood, 26% to 74% was a midpoverty neighborhood, and ≤25% was a low-poverty neighborhood.

Descriptive statistics were used to describe the mean presence and mean volume of exterior and interior advertisements for each tobacco product at each time point as well as across all 4 years. We conducted one-way analysis of variance to compare advertisement presence and volume for each product by store type and by poverty level.

For all 3 indicators of tobacco advertising (presence, volume, and SAV), we used a general linear mixed regression model to examine changes in exterior and interior advertising by product type over time, controlling for store type and poverty. This method is best suited for this study because it does not pose issues related to sphericity and does not assume a normal distribution among variables but rather a dynamic change.²⁰  We examined both within-subject effects (ie, within-store changes over time) and between-subject effects (ie, differences across stores by store type and neighborhood poverty). We also tested for interaction effects to examine whether any within-store changes we observed over time were related to store type and poverty. Finally, we tested for polynomial contrasts for the main effect of time; because we had 4 time points, we tested for quadratic and cubic contrasts that would capture any nonlinear relationships. All analyses were conducted in SPSS version 25 (IBM SPSS Statistics, IBM Corporation).

Of the 141 stores in our sample, 24 were chain convenience stores, 78 were nonchain convenience stores, 12 were gas kiosks, and 27 were other store types. Approximately half of the stores were high poverty (52%), 33% of the stores were midpoverty, and 16% were low poverty.

Overall, cigarettes were advertised on the outside of more than half of the visited tobacco retailers (55%), with menthol cigarettes being advertised in 46% of the retailers (Table 1). Over the years, however, the presence of any cigarette advertising decreased, from 61% of stores advertising cigarettes on the exterior in 2015 to 49% in 2018 (P = .009); advertising for menthol cigarettes remained stable (Table 2). In contrast, there was a significant linear increase in stores with exterior advertising for cigars (11% to 23%; P = .009). Advertising for flavored cigars also increased (9% to 17%; P = .03), showing a significant quadratic trend. E-cigarette advertising showed significant linear and quadratic trends, declining in the first 3 years from 36% to 16% but increasing again in 2018 (P < .001). Advertising for flavored e-cigarettes followed similar, albeit not significant, trends. Exterior advertising for smokeless tobacco (including flavored) did not change over time. There was a significant 3-way interaction between time, store type, and poverty for exterior advertising for smokeless tobacco, including flavored smokeless tobacco. Exterior advertising was more prevalent in areas of low poverty in all store types except for “other” store types, which had no advertising regardless of poverty rates.

There was a significant between-subject effect of store type for all tobacco products such that chain convenience stores had the most exterior advertising compared with nonchain convenience stores, gas kiosks, and other stores. There was a significant between-subject effect of poverty for e-cigarettes and flavored smokeless tobacco such that retailers in areas of high poverty had the least presence of advertising for these products each year.

The volume of exterior advertisements across all tobacco products did not significantly change over time (overall mean = 3.4) and were mostly driven by chain convenience stores (Table 1). Despite the decline in stores advertising cigarettes on the exterior, cigarettes still remained the most promoted tobacco product at the point of sale across all 4 years, with the highest volume of advertisements (overall mean = 2.5) and no significant change over time (Table 2). There was an increase in the mean number of cigar advertisements, from 0.2 in 2015 to 0.5 in 2018 (P = .03). The mean number of e-cigarette advertisements declined from 0.7 in 2015 to 0.3 in 2017 but increased back to 0.4 in 2018 (P < .001). The volume of advertisements for smokeless tobacco remained constant.

Again, there was a significant between-subject effect of store type for all products such that chain convenience stores had the highest volume of advertisements for cigarettes, e-cigarettes, and smokeless tobacco, whereas gas kiosks had the highest volume of advertisements for cigars. There was a significant between-subject effect of poverty for cigars such that retailers in areas of high poverty had the highest volume of advertisements for these products.

Over time, cigarettes declined in SAV (from 50% to 40%; P = .04), whereas the SAV for cigars seemed to increase over time (albeit not significantly; Table 2). The SAV for e-cigarettes decreased from 12% in 2015 to 5% in 2017 but began to increase again in 2018 (P = .003). Smokeless tobacco had a minimal SAV across all years with no significant changes. Interestingly, the SAV for no advertisements significantly increased over time, from 34% to 44% (P = .003), indicating a longitudinal shift toward less exterior advertising of tobacco products in general.

Overall, cigarettes were advertised inside more than two-thirds of the visited tobacco retailers (68%), with menthol cigarettes being advertised in 58% of the retailers (Table 1). Advertising for cigarettes, including menthol cigarettes, did not increase or decrease over time. On the other hand, there was a significant increase in stores advertising cigars (19% to 30%; P = .008), including those that were flavored (8% to 22%; P < .001; Table 3). Mirroring the pattern of e-cigarette exterior advertising, e-cigarette interior advertising also showed significant linear and quadratic trends, declining in the first 3 years from 42% to 23% but increasing in 2018 (P < .001). Interior advertising for flavored e-cigarettes followed similar, albeit not significant, trends. Advertising for smokeless tobacco did not change over time, but advertising for flavored smokeless tobacco increased from 7% in 2015 to 14% in 2018 (P < .001).

Interestingly, there was a 3-way interaction for time, store type, and poverty for interior advertising of flavored cigars, smokeless tobacco, and flavored smokeless tobacco. In chain convenience stores, interior advertising for flavored cigars decreased over time in areas of low poverty but increased over time in areas of high poverty. There was a significant between-subject effect of store type for all tobacco products except cigarettes such that chain convenience stores had the most interior advertising compared with nonchain convenience stores and other stores.

The volume of interior advertisements across all tobacco products fluctuated over time (P = .02), showing a significant cubic trend. Cigarette advertising contributed most to the total volume of interior tobacco advertising and remained constant across all 4 years, whereas the mean number of cigar advertisements and smokeless tobacco advertisements increased over time (Table 3). The mean number of e-cigarette advertisements significantly declined from 1.1 in 2015 to 0.4 in 2017 but increased back to 0.6 in 2018 (P < .001).

Again, there was a significant between-subject effect of store type for all products except cigars such that chain convenience stores had the highest volume of advertisements for cigarettes, e-cigarettes, and smokeless tobacco compared with nonchain convenience and other types of stores.

Of the different tobacco products, only e-cigarettes experienced any changes in SAV, declining from 11% in 2015 to 5% in 2017 and increasing again to 7% in 2018 (P = .001; Table 3). Although the SAV for cigarettes, cigars, and smokeless tobacco did not change significantly over time, they all displayed an increasing pattern while the SAV for no advertisements displayed a decreasing pattern, suggesting an overall increase in interior tobacco product advertising over time.

Using longitudinal data over 4 years (2015–2018), this study found meaningful changes in point-of-sale tobacco advertisements in retailers that were likely to be frequented by youth. Advertising for each tobacco product seems to directly respond to changes in youth use rates for the corresponding product,²¹  with chain convenience stores immediately responding to the marketplace in particular. Taken together with previous research showing that point-of-sale exposure further contributes to youth tobacco use,³  these findings point to a vicious cycle between point-of-sale activities and youth tobacco use and the need for continuous monitoring and regulation. Future studies making direct associations between advertising trends and use patterns, both in New Jersey as well as nationally, would provide further evidence for regulatory actions.

In our study, we consistently observed a quadratic trend in e-cigarette advertising that perfectly mirrored national youth e-cigarette use. E-cigarette advertising both inside and outside of retailers experienced a consistent linear decline until 2017 but began to increase in 2018. Similarly, data from the National Youth Tobacco Survey show that e-cigarette use among high school students decelerated from 16% in 2015 to 11.7% in 2017 but increased by 78% in 2018 to 20.8%.²²  This recent surge in youth e-cigarette use coincides with the increased popularity of JUUL, an e-cigarette that resembles a Universal Serial Bus (or USB) flash drive and whose unit sales have increased by >500% since its launch in 2015.²³  Although the US Food and Drug Administration has initiated efforts to address the growing popularity of e-cigarettes, including JUUL, among youth, our findings emphasize the importance of additionally regulating point-of-sale advertising activities in places that are easily accessible to youth and within close proximity to schools at the local and state levels.

There was also a notable increase in both exterior and interior advertising of cigars, building on previous study findings by Giovenco et al.¹⁵  Cigar consumption remains prevalent among young people, with 7.6% of high school students reporting current cigar use in 2018,²¹  and the continued growth in cigar advertising reflects these use rates. The sharp increase in advertising of flavored cigars is particularly troubling because youth report flavors to be a primary reason for using cigars.^24–26  In contrast, the percentage of retailers with cigarette advertisements on the outside of stores has decreased over the years, with no change in the presence of interior cigarette advertising. This reflects the decline in national youth use rates of combustible cigarettes.²¹  Nevertheless, cigarette advertisements continue to dominate both the interior and exterior advertising space. Thus, when examining advertisement prevalence across all combustible products (ie, both cigarettes and cigars), we see a lack of growth in advertising that mirrors the trends for cigarette advertising and fails to acknowledge the surge in cigar advertising. Grouped findings, such as by combustible and noncombustible products, deemphasize product-level differences, highlighting the importance of examining marketing trends by individual tobacco product.

Furthermore, although cigarettes currently make up most of the advertising proportion both outside and inside of retailers, the SAV for cigars and e-cigarettes are nonnegligible and likely to continue increasing in the near future on the basis of the pattern of results. In other words, cigar and e-cigarette advertisements will likely occupy more retailer advertising space, replacing cigarette advertisements. SAV is an important indicator of potential marketing influence on youth because research has shown that SAV amplifies the effect of youth exposure to advertisements on tobacco use.¹⁹  After all, the more advertisements there are for a product compared with other products within a store, the more youth are likely to notice and absorb the advertising for that product. As such, youth will notice that there are more advertisements for cigars and e-cigarettes compared with cigarettes, further contributing to the way they pay attention to the advertisements for these products. Thus, researchers should continue to account for the advertising proportion of different tobacco products inside and outside of stores.

Further analyses found that retailers near low-income schools had the least advertising for e-cigarettes and flavored smokeless tobacco and the most advertising for cigars. In fact, chain convenience stores near low-income schools increased advertising for flavored cigars over time. These findings are in line with previous research that has shown that areas with lower income levels and larger proportions of nonwhite residents have more advertising and availability of inexpensive, combustible products, such as cigars,^13,27,28  and less advertising of noncombustible products, such as smokeless tobacco and e-cigarettes.^12,27,29  Our finding that such advertising differences exist near schools is particularly concerning because they may introduce tobacco-related disparities at an earlier age, potentially persisting into adulthood and exacerbating existing inequities. Thus, these neighborhood-level differences warrant focused attention as regulations are considered.

Potential limitations of our study include an inability to generalize findings to other parts of the United States because data were specific to retailers within close proximity of New Jersey high schools. Furthermore, the study specifically examined retailers surrounding schools under the assumption that these places afforded youth convenient access to tobacco and repeated exposure to tobacco marketing but did not account for retailers near other youth activity zones, such as parks or recreation centers. Future studies replicating these findings in a broader sample of retailers in other parts of the country or at a national level could lead to more robust conclusions about the overall point-of-sale environment surrounding US youth. This study did not include race as a covariate, which some may consider a limitation. However, our indicator of poverty was highly correlated with our measure of race (ie, percentage of whites [r = −0.88]), and controlling for both poverty and race would have led to overadjustment of variables. Finally, this study did not take into account new tobacco retailers that started selling near schools since the 2015 baseline and visited the same cohort of retailers since baseline.

The current study is the first, to our knowledge, to provide evidence for longitudinal changes in point-of-sale advertising of different tobacco products near schools. Findings show a high prevalence and even increases in exterior and interior point-of-sale advertising for tobacco products near high schools with increased opportunity for youth exposure to such advertisements. Youth are affected by exposure to industry marketing, including at the point of sale. Thus, it is imperative that changes in and the effects of the point-of-sale environment continue to be assessed, building an evidence base for regulatory efforts that can minimize youth exposure to industry marketing and ultimately reducing youth tobacco use.

e-cigarette

electronic cigarette

SAV

share of advertising voice