Association of Retail Tobacco Marketing With Adolescent Smoking

Abstract

A survey of 2125 middle-school students in central California examined adolescents’ exposure to tobacco marketing in stores and its association with self-reported smoking. Two thirds of sixth-, seventh-, and eighth-grade students reported at least weekly visits to small grocery, convenience, or liquor stores. Such visits were associated with a 50% increase in the odds of ever smoking, even after control for social influences to smoke. Youth smoking rates may benefit from efforts to reduce adolescents’ exposure to tobacco marketing in stores.

The tobacco industry spends more on point-of-purchase marketing than on all other forms of cigarette advertising combined ($9.5 billion vs $1.7 billion in 2001).1 Since the ban on billboard advertising, promotional materials for tobacco have increased significantly on store windows and around the counter.2,3 Three out of 4 teenagers shop at a convenience store at least once a week,4 staying an average of 16 minutes per visit—twice as long as adults.5 In recent surveys, all 15- and 16-year-olds reported seeing point-of-purchase marketing for cigarettes,6 and teenaged smokers preferred whichever brand (Marlboro or Camel) was advertised most heavily in the convenience store nearest their school.7 Experimental studies suggest that retail tobacco marketing exposure distorts adolescents’ perceptions about the availability, use, and popularity of tobacco8,9—normative beliefs that are precursors of smoking.10–12

The only study to correlate adolescents’ self-reported smoking with exposure to retail tobacco marketing found that experimentation was 38% more likely for seventh-grade students who said that they visited a convenience, liquor, or small grocery store at least weekly and remembered seeing advertisements for cigarettes sometimes or a lot.13 Because cigarette advertising is more noticeable to experimenters than to abstinent youths,14,15 this finding may be attributable to perceived exposure alone. Alternatively, store exposure may be associated with smoking for reasons other than the influence of advertising, such as a propensity for risk taking or lack of after-school supervision, which are established risk factors for smoking.11,16,17 After we controlled for such risk factors and social influences to smoke, we examined whether adolescent smoking is associated with exposure to stores that contain widespread tobacco marketing.

METHODS

We administered the Survey of Teen Opinions about Retail Environments (STORE Study) in all 3 middle schools in Tracy, Calif (population 62500), in spring 2003. Of the initial sample of sixth-, seventh-, and eighth-grade students (N=2731), 396 students did not return parental consents, 95 were absent for the survey, and 115 parents refused permission, yielding a response rate of 78%. The final sample (N=2125) was 53% female and 42% Hispanic, 27% White, 11% Asian, 6% African American, 1% American Indian, 1% Pacific Islander, and 12% multiethnic.

Each survey included photographs and addresses of 12 retail tobacco outlets in the school catchment area that were identified by student focus groups as popular destinations for purchasing snacks. Exposure to retail tobacco marketing was coded for students who reported at least weekly visits to convenience, liquor, or small grocery stores, either in response to questions about the pictured stores or in response to questions about visiting any such stores in the past month. Exposure to other forms of tobacco marketing was coded for students who reported owning a cigarette promotional item, sometimes or often seeing tobacco advertisements in magazines in the past week, or sometimes or often seeing people smoking on television or in movies in the past week. Exposure to family and peer smoking was coded, separately, for students who reported having a parent or other household member who currently smokes and for students who reported having at least one friend who currently smokes. The survey also included self-reported grades, a 3-item risk-taking measure,11 and 4 items of the Authoritative Parenting Index18 that measure maternal supervision.

We used GLIMMIX for SAS, Version 8.0 (SAS Institute Inc, Cary, NC), to examine the potential effect of school clustering with multilevel modeling, but results did not differ from the standard multiple regression reported here. Odds ratios were computed to test the association between adolescents’ exposure to retail tobacco marketing and ever smoking a cigarette, even just a puff. This was the primary study outcome because the prevalence of current smoking was quite small: 2.6% of sixth-, 6.0% of seventh-, and 7.6% of eighth-grade students reported any cigarette smoking in the past 30 days.

RESULTS AND DISCUSSION

Two thirds of the students said that they visited a convenience, liquor, or small grocery store on their way to or from school at least weekly (Table 1 ▶); about one fourth of the students reported visiting such stores practically every day. Exposure to retail tobacco marketing was more prevalent among boys; Latino youths; and students who reported low maternal supervision, high risk taking, exposure to social influences to smoke, and exposure to other forms of tobacco marketing.

TABLE 1—

Characteristics of Study Participants and Prevalence of Exposure to Retail Tobacco Marketing: 2125 Middle-School Students, Tracy, Calif, 2003

	% of Full Sample (N = 2125)	Prevalence of Exposure to Retail Tobacco Marketing, %	Unadjusted OR (95% CI)
Grade
6tha	33.2	68.0
7th	29.8	66.1	0.9 (0.7, 1.2)
8th	37.0	65.2	0.9 (0.7, 1.1)
Gender
Femalea	53.3	63.9
Male	46.7	69.1	1.3 (1.1, 1.5)
Ethnicity
Not Latino or Hispanica	57.9	59.6
Latino or Hispanic	42.1	76.0	2.2 (1.8, 2.6)
Self-reported grades
Above mediana	52.6	60.2
Below median	47.4	73.1	1.8 (1.5, 2.2)
Parent or household smoking
Noa	55.5	60.3
Yes	44.5	74.0	1.9 (1.6, 2.3)
At least 1 friend smokes
Noa	72.3	61.9
Yes	27.7	77.9	2.2 (1.7, 2.7)
Maternal supervision
Above meana	50.6	62.7
Below mean	49.4	70.0	1.4 (1.2, 1.7)
Risk taking
Below meana	46.3	56.6
Above mean	53.7	74.6	2.3 (1.9, 2.7)
Own cigarette promotional item
Noa	84.8	63.5
Yes	15.2	82.8	2.8 (2.0, 3.7)
See cigarette advertisements in magazines
Never or rarelya	48.8	55.4
Sometimes or often	51.2	76.9	2.7 (2.2, 3.2)
See smoking on television or in movies
Never or rarelya	32.0	55.9
Sometimes or often	68.0	71.4	2.0 (1.6, 2.4)
Exposure to retail tobacco marketing
< Once/weeka	33.8	. . .	. . .
At least once/week	66.2	. . .	. . .

Note. OR = odds ratio; CI = confidence interval.

^aReference category.

Table 2 ▶ summarizes the bivariate and multivariate associations of ever smoking with demographic attributes and psychosocial risk factors. After we controlled for all other variables in the model, weekly or more frequent exposure to retail tobacco marketing was associated with a 50% increase in the odds of ever smoking—nearly as much as the effect of exposure to a parent or household member who smokes. Although Latino youths were more likely to be exposed to retail tobacco marketing, they had no greater risk for smoking when other variables were controlled.

TABLE 2—

Unadjusted and Adjusted Odds Ratios (ORs) From Stepwise Logistic Regression for Ever Smoking: 2125 Middle-School Students, Tracy, Calif, 2003

	Unadjusted OR	CI	Adjusted ORa	CI
Grade
6thb
7th	2.7	2.0, 3.7	2.0	1.4, 2.9
8th	4.5	3.3, 6.1	3.0	2.1, 4.3
Gender
Femaleb
Male	1.1	0.9, 1.4	NS
Ethnicity
Not Latino or Hispanicb
Latino or Hispanic	1.4	1.1, 1.7	NS
Self-reported grades
Above medianb
Below median	2.7	2.2, 3.4	1.5	1.2, 2.0
Parent or household smoking
Nob
Yes	2.6	2.1, 3.2	1.8	1.4, 2.3
At least 1 friend smokes
Nob
Yes	6.6	5.2, 8.3	3.1	2.4, 4.0
Maternal supervision
Above meanb
Below mean	1.9	1.5, 2.4	1.5	1.1, 1.9
Risk taking
Below meanb
Above mean	6.7	5.1, 8.9	3.4	2.5, 4.7
Own cigarette promotional item
Nob
Yes	6.0	4.7, 7.8	2.7	2.0, 3.6
See cigarette advertisements in magazines
Never or rarelyb
Sometimes or often	2.5	2.0, 3.1	1.4	1.1, 1.8
See smoking on television or in movies
Never or rarelyb
Sometimes or often	2.2	1.7, 2.8	NS
Exposure to retail tobacco marketing
< Once/weekb
At least once/week	2.8	2.1, 3.6	1.5	1.1, 2.1

Note. CI = confidence interval; NS = not selected by stepwise selection method.

^aAdjusted OR controls for school differences and all other model variables simultaneously.

^bReference category.

Because we controlled for confounders such as risk taking, maternal supervision, and self-reported grades, the association between adolescents’ store visits and their smoking behavior appears more likely an effect of advertising exposure than an artifact of idle hands or hanging out with the wrong crowd. Relative to other forms of tobacco marketing, retail marketing exposure was second only to owning a cigarette promotional item in increasing the odds of ever smoking. However, the smaller association observed for exposure to cigarette advertisements in magazines and no association for depictions of smoking on television or in movies may be an artifact of redundancy within exposure measures.

This cross-sectional study could not confirm a causal role for retail tobacco marketing in the uptake of smoking but provided stronger evidence for causality than previous studies have by controlling several potential confounds.

The effect of retail tobacco marketing on adolescent smoking may be particularly potent in states with comprehensive tobacco control programs where tobacco advertisements and promotions are more numerous at the point of sale.19,20 Because the tobacco industry has relatively few regulations on advertising in stores,21–25 further research, including longitudinal studies, is needed to determine how the proliferation of tobacco marketing in stores affects youths.