Tobacco advertising, particularly at the point of sale, is used to portray tobacco products in a positive light [
In 1998, the Massachusetts Department of Public Health conducted a statewide campaign, Operation Storefront, to document the extent of tobacco advertising in retail stores. Results showed that tobacco was the most advertised product visible to youth from outside retail stores. Each store had an average of five storefront ads that were visible from outside the retail store [
In 2009, the federal government enacted the Tobacco Control Act (TCA) which permitted state and local governments to further regulate tobacco ads in retail stores including licensing, placement, and categories of tobacco ads including power walls behind cash registers [
We conducted the first cross-sectional study of tobacco ads in Massachusetts following the wave of policies adopted by state and local governments after the 2009 Tobacco Control Act. We examined the extent and types of tobacco advertising. We also tested the association between the presence of tobacco ads, number of tobacco ad categories, and number of tobacco ads and comprehensiveness of local tobacco control policies. We hypothesized that jurisdictions with weaker tobacco control policies would have more tobacco ads and vice versa. Massachusetts has a progressive stance towards tobacco control policies relative to most other states in the US. For example, it has the 5th highest tobacco excise tax and was the first state to prohibit the sale of tobacco products in pharmacies. In addition, it reports the 7th lowest smoking rates in the country at 13.7%. The State and Community Tobacco Control Research Initiative (SCTC) has classified Massachusetts as one of the states that appear well-positioned to expand their efforts into the POS area. Thus, Massachusetts provides an excellent context for conducting tobacco advertising and compliance studies.
Data were collected from 419 retail stores within selected Massachusetts’ municipalities using a two-stage cluster sampling method from March to May 2017. We chose retail stores located in municipalities with varying comprehensiveness of tobacco regulations to ensure that retail stores in municipalities with and without comprehensive tobacco policies were eligible to be included in the sample.
The first stage involved selecting municipalities using probability-proportion to size (PPS); the selected municipalities were then split into seven groups according to the comprehensiveness of their tobacco regulations [
Following municipality selection across the seven groups noted (Figure
Sampled municipalities across local tobacco control policy strength.
Data on tobacco advertising were collected from the retail stores using a standardized observation instrument [
Data were collected electronically by the first author using an Android data collection software called Open Data Kit (ODK) (
Screenshot of the survey on mobile device using Open Data Kit (ODK) Software.
Measures of tobacco ads included (i) the presence or absence of tobacco ads, (ii) the number of tobacco ads, and (iii) the number of tobacco ad categories. The presence and absence of tobacco ads were recorded as binary: 1 for presence and 0 for absence. The number of tobacco ads was a count of all the tobacco ads that were on display at the store. The number of tobacco ad categories ranged from 0 to 8 including (1) external posters, (2) internal posters, (3) e-cigarette ads, (4) smokeless tobacco ads, (5) flavored tobacco ads (excluding menthol), (6) branded items, (7) power walls, (8) ads for discounts on tobacco products such as “Buy two get one free” or “Buy one get the other half-off”, and (9) backlit ads. Each tobacco product category was recorded as binary: 1 for presence and 0 for absence. External ads were recorded as conventional cigarette or cigar ads only visible from the outside of the store, e.g., on doors, windows, walls. Internal ads included posters, decals, hanging signs, and other kinds of signage of conventional cigarettes or cigars ads visible only within the store. All ad categories were mutually exclusive, as both external and internal posters excluded ads on smokeless tobacco, discounts on tobacco products, e-cigarettes, and flavored tobacco products. The specific location (inside or outside the store) was not recorded for the other ad categories: smokeless tobacco, discounts on tobacco products, e-cigarettes, and flavored tobacco products. Nine different categories of tobacco ads were assessed, but none of the retail stores had a backlit ad, resulting in the 0 to 8 range noted.
Local tobacco control policies were based on the number of tobacco control policy areas covered beyond tobacco advertising at POS. Data for this variable were obtained from the Massachusetts Tobacco Control Program (MTCP). Municipalities were grouped according to how many of the five key policy areas had been adopted: (1) policies that limit tobacco retail sales permits, (2) minimum pricing for cigars, (3) policies regulating the sale of e-cigarettes and nicotine delivery products to minors, (4) a ban on all flavored tobacco products, and (5) a ban on the sale of tobacco products in pharmacies. So, municipalities with no policies were coded as 0, while municipalities with all five policy areas were coded as 5. Hence, each municipality’s score on this variable ranged from 0 to 5.
Median household income and the percentage of minority population in a municipality were acquired from the 2009–2013 American Community Survey (ACS) 5-year estimates. Percentage minority was defined as the percentage of the residents representing the minority population (minority defined as all other races excluding non-Hispanic white residents). Store types included convenience stores (
Data were exported to Excel through the ODK software and transferred to Stata 15.0 for analysis. Mixed-effects logistic regression, ordinal logistic regression, and negative binomial regression were used to assess the relationship between the different tobacco ad variables and the comprehensiveness of each municipality’s tobacco control local policies. Because the stores were clustered within the municipalities, all regression models were estimated using a multilevel approach. Robust standard errors were also used, which produce better variance estimates for clustered data.
The sample included 419 retail stores; more than half were convenience stores (29.1%) or gas stations (28.1%) (Table
Descriptive statistics of all study variables across all retail stores,
# (%)/mean (SD) | Minimum | Maximum | |
---|---|---|---|
Tobacco advertisements | |||
Presence of advertisements | 363 (86.7%) | ||
Number of tobacco advertisementsa | 6.69 (6.61) | 0 | 32 |
Number of tobacco advertisement categories (range) | 2.89 (1.84) | 0 | 7 |
Retail store type | |||
Convenience stores | 122 (29.1%) | ||
Gas station | 118 (28.2%) | ||
Liquor store | 73 (17.5%) | ||
Drug store | 10 (2.4%) | ||
Chain retail stores | 49 (11.7%) | ||
Nonchain retail stores | 16 (3.8%) | ||
Other store typesb | 31 (7.4%) | ||
Categories of tobacco advertisements | |||
Presence/number of ads (mean/SD) | |||
External ads | 204 (48.7%)/2.8 (4.3) | 0 | 26 |
Backlit ads | 0 (0.0%) | 0 | 0 |
Branded items | 14 (3.3%)/0.03 (0.2) | 0 | 1 |
Smokeless tobacco ads | 181 (43.2%)/0.5 (0.8) | 0 | 5 |
Flavored tobacco ads | 26 (6.2%)/0.1 (0.2) | 0 | 1 |
Power wall | 335 (80.0%)/0.8 (0.4) | 0 | 1 |
Ads of discounts on tobacco products | 91 (21.7%)/0.3 (0.5) | 0 | 6 |
E-cigarette ads | 234 (55.8%)/0.8 (1.0) | 0 | 8 |
Internal posters | 125 (29.8%)/1.4 (2.8) | 0 | 15 |
aSum of the number of tobacco advertisements across all stores = 2,804. bOther store types include tobacco shops, big-box stores, fashion stores, bars, and private clubs.
Overall, 363 retail stores (86.6%) had tobacco advertising (Table
Descriptive statistics of all study variables at the municipality level,
Mean (SD) | Minimum | Maximum | |
---|---|---|---|
Tobacco advertisements | |||
Presence of advertisements | 8.6 (4.6) | 4 | 36 |
Number of tobacco advertisementsa | 6.7 (2.8) | 1.6 | 14.5 |
Number of tobacco advertisement categories (range) | 2.9 (0.8) | 1.3 | 5 |
Retail store type | |||
Convenience stores | 7.4 (8.3) | 1 | 24 |
Gas station | 4.0 (1.9) | 1 | 8 |
Liquor store | 2.6 (1.2) | 1 | 5 |
Drug store | 1.2 (0.4) | 1 | 2 |
Chain retail stores | 1.9 (0.7) | 1 | 3 |
Nonchain retail stores | 1.3 (0.4) | 1 | 2 |
Other store typesb | 1.9 (1.0) | 1 | 4 |
Categories of tobacco advertisements | |||
Presence of external ads | 5.0 (3.1) | 1 | 18 |
Presence of backlit ads | 0 (0.0) | 0 | 0 |
Presence of branded items | 1.3 (0.9) | 1 | 4 |
Smokeless tobacco ads | 4.4 (2.1) | 1 | 10 |
Presence of flavored tobacco ads | 2.6 (1.4) | 1 | 5 |
Presence of a power wall | 8.0 (4.1) | 4 | 32 |
Presence of discounts on tobacco products | 3.8 (4.2) | 1 | 20 |
Presence of e-cigarette ads | 5.6 (2.0) | 1 | 12 |
Presence of internal posters | 4.0 (3.5) | 1 | 19 |
On average, retail stores had three different categories of tobacco ads (mean = 2.98, SD = 1.84) (Table
Of the 2804 tobacco ads, the highest proportion was found in gas stations (
Descriptive and bivariate statistics for tobacco advertisements categories by retail store type,
Total ( |
Convenience ( |
Gas station ( |
Liquor store ( |
Drug store ( |
Chain retail ( |
Nonchain ( |
Other ( |
|
|
---|---|---|---|---|---|---|---|---|---|
Presence of tobacco ads | 363 (100%) | 107 (87.8%) | 115 (97.5%) | 63 (86.3%) | 8 (80.0%) | 38 (77.6%) | 12 (75.0%) | 20 (64.5%) | 21.1 (1), |
Total no. of tobacco ads (count) | 2804 | 1016 | 1096 | 309 | 13 | 187 | 78 | 105 | 3.7 (28), |
Average no. of tobacco ads | 6.7 (6.6) | 8.3 (7.5) | 9.3 (6.3) | 4.2 (4.6) | 1.3 (1.2) | 3.8 (4.9) | 4.9 (5.0) | 3.4 (5.9) | 2.2 |
Average no. of tobacco ad categories (range) | 2.9 (1.8) | 3.4 (2.0) | 3.7 (1.3) | 2.2 (1.5) | 1.2 (0.9) | 2.0 (1.7) | 2.4 (2.0) | 1.5 (1.9) | 24.1 (6), |
External ads | 204 (100%) | 72 (59.0%) | 77 (65.3%) | 23 (31.5%) | 0 (0.0%) | 14 (28.6%) | 7 (43.8%) | 11 (35.5%) | 46.5 (6), |
Backlist ads | 0 (100%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | |
Branded items | 14 (100%) | 7 (5.7%) | 4 (0.3%) | 3 (0.04%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 6.0 (6), |
Smokeless tobacco ads | 181 (100%) | 63 (51.6%) | 72 (61.0%) | 18 (0.3%) | 0 (0.0%) | 15 (30.6%) | 7 (43.8) | 6 (19.4%) | 47.0 (6), |
Flavored tobacco ads | 26 (100%) | 11 (9.0%) | 7 (5.9%) | 7 (9.6%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 1 (3.2%) | 8.5 (6), |
Power wall | 335 (100%) | 101 (82.8/%) | 109 (92.4%) | 51 (69.9%) | 8 (80.0%) | 38 (77.6%) | 12 (75.0%) | 16 (51.6%) | 32.6 (6), |
E-cigarette ads | 234 (100%) | 74 (60.7%) | 92 (78.0%) | 27 (37.0%) | 3 (30.0%) | 21 (42.9%) | 7 (43.8%) | 10 (32.3%) | 49.1 (6), |
Ads of discounts | 91 (100%) | 49 (40.2%) | 23 (19.5%)) | 9 (12.3%) | 1 (10.0%) | 3 (6.1%) | 4 (25.0%) | 2 (6.5%) | 40.7 (6), |
Internal posters | 125 (100%) | 44 (36.1%) | 48 (40.7%) | 20 (27.4%) | 0 (0.0%) | 9 (18.4%) | 2 (12.5%) | 2 (6.5%) | 26.8 (6), |
Results from the mixed-effects logistic and negative binomial regression models are presented in Table
Bivariate and multiple regression models showing the association between tobacco advertising variables and the strength of local tobacco control policies.
Variables | Presence of tobacco ads1 | Presence of discounts | Presence of flavored tobacco ads | Range of tobacco ad categories2 | Number of tobacco ads3 | |||||
---|---|---|---|---|---|---|---|---|---|---|
OR | AOR | OR | AOR | OR | AOR | OR | AOR | IRR | AIRR | |
Independent variable | ||||||||||
Strength of other tobacco policies | 0.84+ | 0.84+ | 0.99 | 1.01 | 0.71 | 0.78 | 0.89 |
0.88 |
0.91 |
0.91 |
Ban on discounts | 0.44 | 0.63 | ||||||||
Ban on flavored OTP | 0.10 |
0.05 |
||||||||
Control variables | ||||||||||
Retail store level | ||||||||||
Store type | ||||||||||
Convenience stores (REF)# | ||||||||||
Gas station | 5.49 |
0.76 | 0.54 | 1.22 | 1.23 | |||||
Chain retail/liquor stores | 0.50 |
0.21 |
0.37 | 0.19 |
0.46 |
|||||
Nonchain retail/drug store/other store types4 | 0.45 | 0.11 |
1 | 0.22 |
0.49 |
|||||
Municipality level | ||||||||||
Percent minority (% of minority population) | 0.98 | 1.09 |
1.01 |
1.00 | 1.00 | |||||
Median household income ($1,000) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
|||||
Fit statistics | ||||||||||
|
||||||||||
N | 419 | 419 | 419 | 419 | 419 | 419 | 419 | 419 | 419 | 419 |
AIC | 330.0 | 311.7 | 369.7 | 345.1 | 164.1 | 161.8 | 1629.1 | 1554.5 | 2485.0 | 2421.9 |
BIC | 342.1 | 344.0 | 385.8 | 381.4 | 180.3 | 193.1 | 1665.4 | 1611.0 | 2501.2 | 2458.3 |
The odds of a retail store having seven tobacco ad categories, the highest number of ad categories (Table
On the other hand, the estimated rate of having an additional tobacco ad in a chain retail/liquor store or a nonchain retail/drug store/other store type was 54% and 51% lower (IRR = 0.46,
This study reveals that the sampled retail stores in Massachusetts had an average of seven tobacco ads and three tobacco ad categories. Like previous studies, convenience stores and gas stations had most of the tobacco ads while drug stores and nonchain retail stores had the least tobacco ads [
As mentioned earlier, external and internal ads in this present study, which were only cigarette and cigar ads, excluded ads on smokeless tobacco, discounts, e-cigarettes, and flavored tobacco products, which were recorded separately, thereby potentially underestimating the number of all external ads. Some of the e-cigarette ads, smokeless tobacco ads, flavored tobacco ads and ads of discounts on tobacco products may have been located at the storefront. This suggests that the range of ads visible from outside the store may range from 2.8 (external ads) to 4.4 (all tobacco ads excluding indoor ads, power walls, and branded items). This range is generally consistent with the average number of external ads reported in previous surveys pre-TCA, including 5 external cigarette, cigar, and smokeless tobacco ads in 1998 [
Notably, over half of the retail stores in the sample had e-cigarette ads. All the sampled retail store types had an average of one e-cigarette ad. While over 60% of the convenience stores and gas stations had e-cigarette ads, over 30% of the drug stores and other store types also had e-cigarette ads. The high visibility of e-cigarette ads in retail stores may signal a parallel surge with increasing use of e-cigarettes [
Although several municipalities in Massachusetts have adopted policies limiting or banning tobacco discounts, there is still widespread use of this advertising avenue. Approximately 22% of all the retail stores sampled had ads of discounts on tobacco products. Moreover, these categories of ads were present in each retail store type, ranging from 10% or less in drug and other retail stores to approximately 40% in convenience stores and gas stations. Discounts on tobacco products are mostly harmful to youth and adolescents who are more easily swayed by price promotions than older smokers [
Smokeless tobacco was also heavily advertised, with ads appearing in 43.2% of the retail stores sampled. Smokeless tobacco manufacturers increased their advertising expenditures by about $160 million from 2014 to 2016 [
Between 2013 and 2014, spending on cigarette tobacco ads inside stores increased from $55.7 million to $238.2 million [
The mixed results on the association between tobacco ads and its corresponding bans support existing studies on tobacco regulations [
Also, the study supports existing studies regarding factors associated with tobacco advertisements by finding an inverse relationship between advertising (presence, number of categories, and number) and the comprehensiveness of its local tobacco policies, such as a ban on the sale of e-cigarettes to minors, ban on the sale of tobacco products in pharmacies, and a limit on tobacco permits for retail stores. Similar to the previous studies, these results suggest that more comprehensive tobacco control policies might be more effective in reducing the extent of tobacco advertising relative to weaker policies [
This is one of the few studies assessing the extent of tobacco retail advertising in Massachusetts and its association with local tobacco control policies. However, there are several limitations worth noting. The external and internal ads were cigarette and cigar ads only, limiting this study’s ability to make comprehensive assessments of more tobacco product ads in reference to the ad location. Similarly, this study did not distinguish the location of the other tobacco ad categories that were not external ads. Apart from the power walls and branded items, the study did not indicate where the flavored tobacco ads, ads of discounts on tobacco products, smokeless tobacco ads, and e-cigarette ads were located (whether outside or inside the store). With this limitation, we were unable to make exact comparisons with the previous studies on retail store tobacco advertising. Also, we collated the regulatory provisions into an ordinal scale and as such were unable to assess the effect of each regulation.
Notably, the study is only representative of the subset of retail stores in Massachusetts municipalities with ten or more retail stores, as municipalities with less than ten retail stores were excluded from the sampling frame. Additionally, we were unable to conduct interrater reliability because there was only one data collector.
Despite the Surgeon General’s warning that tobacco advertising and promotion cause smoking initiation and the Tobacco Control Act that allows state and local governments to regulate retail tobacco advertising, retail store tobacco advertising continues in Massachusetts. This study reveals the extent to which young people are potentially exposed to retail tobacco advertising in Massachusetts and confirms that this potential exposure is reduced in municipalities with comprehensive local tobacco control policies. In particular, tobacco companies are still channeling promotion through multiple avenues including external ads, internal ads, power walls, e-cigarettes ads, and smokeless tobacco ads. These results suggest that local policies governing advertising should be encouraged across all municipalities to limit the extent of tobacco advertising in their area, hence reducing smoking rates. Other municipalities might follow Boston’s example in enacting a ban on flavored tobacco products but include a ban on menthol-flavored products as well. Also, municipalities without a ban on the sale of tobacco products with discounts could incorporate a ban to further reduce the use of discounts to sell tobacco products. Federal and state governments are also encouraged to adopt policies that are already promoted on a municipality level such as a ban on all discount types to purchase tobacco products. These findings imply that municipalities with weak tobacco control policies should strengthen them, in the expectation that less exposure to tobacco advertising is likely to reduce smoking uptake among young people [
The data used to support the findings of this study are available from the corresponding author upon request.
The authors declare that they have no conflicts of interest.