Lesbian, gay, and bisexual (LGB) people are at much higher risk of tobacco use than their straight counterparts [1
]. In addition to more than 50% higher smoking prevalence than for straight people [2
], LGB people are more likely to smoke menthol cigarettes, flavored little cigars, filtered little cigars, and use e-cigarettes than heterosexual people [2
]. Data for transgender populations are scarce [6
], although similar disparities likely exist [7
]. The reasons for these disparities are only partially known, and research has focused primarily on the role of discrimination, stigma, and stress [8
]. Researchers have also suggested that the role of LGB bars as safe community spaces may promote tobacco use [9
], that the media environment may contribute as tobacco use is normative in the LGB print press and in LGBT-themed movies [11
], and that tobacco industry marketing targeted directly at LGB communities contributes to disparities [14
In one of the tobacco industry’s plans to market to gay men titled Project S
arket (SCUM), RJ Reynolds Tobacco Co. (Winston-Salem, NC, USA) planned to make its products and their marketing ubiquitous in a neighborhood considered to be a gay enclave, San Francisco’s Castro [14
]. Project SCUM called for better “in store presence”, better “store front presence”, and “consistent POS[point-of-sale]/PDI placements”, with an objective to “[p]enetrate fragmented/nontraditional outlets to increase Camel’s Distribution and presence” [17
]. Thus, in one of the clearest examples of tobacco industry targeted marketing to LGB people, Reynolds sought to make its marketing ubiquitous in tobacco retailers in a gay neighborhood.
In studies of the etiology of LGB health disparities in tobacco [8
], neighborhood-level marketing to LGB people has been largely ignored although some research has examined political and social environments at the county level and school level in relation to LGB youth smoking [18
]. However, there is emerging demographic evidence that internal migration of LGB people within the United States results in the concentration of same-sex couples (We use the terminology “same-sex couples” to describe patterns available from the U.S. Census, which does not report on individual sexual orientation. When discussing the literature or conceptual issues, we use LGB.) in certain types of neighborhoods, in regional cities, and in places where there are already more same-sex couples [19
]. These patterns are more complex than the common view of migration of LGB people to major cities [22
], reveal decision-making regarding migration that is informed by sexual orientation identity [23
], and indicate that LGB people have, as a population, unique spatial patterning [19
]. There are differences in patterning between same-sex female and male couples with men concentrating into fewer more urban neighborhoods and women in more suburban areas [20
]. Indeed, demographic research suggests that same-sex couples often live in lower income and more diverse neighborhoods [20
], especially male same-sex couples [25
]. Tobacco industry marketing is frequently found at greater volume at retailers in poorer and less White neighborhoods [27
]. Neighborhoods with more Black residents have more menthol marketing [29
], lower menthol prices [33
], and more little cigar marketing [35
]. E-cigarettes are more available in higher-income neighborhoods and neighborhoods with more White residents [36
]. Thus, same-sex couples are, as a population, more likely to live in a more racially/ethnically diverse neighborhood with lower income; these neighborhoods are likely to have disproportionate tobacco market due to industry targeting of neighborhoods with more Black residents and lower-income [27
Tobacco marketing at the POS is part of a broader marketing effort that is causally related to smoking behaviors [37
]. Both a National Cancer Institute monograph and the Surgeon General report highlight the importance of POS marketing to tobacco prevention and control [37
]. The tobacco industry spends the majority (85%) of its reported marketing dollars at the POS in the United States [40
]. Two systematic reviews have synthesized the evidence of the impact of POS marketing on tobacco-related health behaviors, suggesting sufficient evidence for policy intervention [42
]. Because tobacco use starts during adolescence [37
], POS tobacco marketing is relevant to our understanding of LGB tobacco use disparities through its role in stymied quit attempts. Greater volume of tobacco marketing at retailers in neighborhoods with more same-sex couples could delay quit attempts or make them more likely to fail [42
We aimed to examine the association between census tracts’ rate of same-sex couple households and tobacco retailers’ marketing. Because this is the first study to explore the rate of same-sex couples in neighborhoods in relation to retailers’ POS marketing, our hypotheses are driven by two competing approaches. First, same-sex couples tend to live in more diverse neighborhoods [20
], which are the same neighborhoods that are more likely to be targeted by the tobacco industry [27
]. These neighborhoods have smaller, non-chain stores with more marketing [49
]. Second, same-sex couples have been associated with neighborhood gentrification [52
] and rising home prices [54
], which are associated with less POS tobacco marketing [29
]. We proposed eight hypotheses based on the first approach, given the weight of demographic evidence [20
] but recognizing that this study is the first to investigate this relationship.
Given the LGB tobacco disparities that exist (higher smoking prevalence, higher menthol use, and higher use of flavored little cigars) we hypothesized store marketing characteristics that would contribute to those disparities in hypotheses 1–7. We then hypothesized the likelihood of stores’ sale of e-cigarettes based on evidence that same-sex couples live in more diverse neighborhoods [36
]. The rate of same-sex couple households in census tracts is:
Positively associated with the presence of (H1) promotional offers and (H2) Newport-specific (i.e., menthol) promotional offers.
Not associated with the advertised price of Marlboro Red cigarettes at tobacco retailers.
Negatively associated with the price of Newport-brand mentholated cigarettes at tobacco retailers.
Positively associated with numbers of (H5) total marketing materials and (H6) total number of exterior marketing materials.
Positively associated with (H7) the likelihood of the sale of flavored cigars and (H8) negatively associated with the likelihood of the sale of e-cigarettes.
Overall, we found few significant relationships between the rate of same-sex couples in census tracts and eight measures of tobacco marketing in tobacco retailers within those tracts. Nonetheless, our study did find small but significant relationships in unexpected directions, including higher Newport prices for same-sex male couples and fewer ads for both male and female same-sex couples. Although these finding should be replicated in other data sources, it provides no evidence that the origin of LGB tobacco disparities lies in store-level differences in POS tobacco marketing. Nonetheless, our analysis does not take into account differences in retailer density and thus cannot determine if the total volume of marketing in neighborhoods
is associated with same-sex couple rates. In previous work, we found a small but significant positive association between same-sex couple rates and tobacco retailer density, with markedly higher density among the neighborhoods with the highest rates of same-sex couples [72
]. Greater density could indicate greater neighborhood-level tobacco marketing even in the absence of store-level differences in tobacco marketing.
Our results imply that there is no significant difference between marketing in neighborhoods with more same-sex couples compared to neighborhoods with fewer couples. However, in our data, as expected, same-sex couple rates at the tract level were positively associated with the proportion of tract residents reporting Black race (rs(n = 1696)
= 0.29, p
< 0.01; rs(n = 1696)
= 0.32, p
< 0.01) or Hispanic/Latino ethnicity (rs(n = 1696)
= 0.11, p
< 0.01; rs(n = 1696)
= 0.17, p
< 0.01) and negatively associated with median household income (rs(n = 1696)
= −0.18, p
< 0.01; rs(n = 1696)
= −0.15, p
< 0.01), respectively, for female and male same-sex couples. It may be that neighborhoods with higher same-sex couple rates are qualitatively different from those otherwise being targeted by the tobacco industry for their racial/ethnic diversity and lower income. Or, alternatively, processes of neighborhood change including gentrification may have attenuated the relationships between neighborhood demographics and tobacco industry targeting that have previously been documented [27
] by changing the composition of neighborhood stores and advertising. It is also possible that growing acceptance of same-sex couples has led to increasing integration of same-sex couples into a broader array of neighborhoods than in previous years. There is some evidence to support this; segregation indices for same-sex couples have declined between the 2000 and 2010 census [21
As for specific products and types of marketing, flavored cigars are disproportionately used by LGB people [4
] as are little filtered cigars [5
], and we expected to see a neighborhood-level association with the presence of flavored cigars (regular or little) in tobacco retailers. Not finding this association, we suggest further investigation using other more comprehensive measures of little cigar sales and examination of the potential for flavored little cigars use to be influenced by LGB identity-related preferences to enhance our understanding of neighborhood-level influence of the retail marketing environment.
Fallin et al.
report that LGB smokers were more likely to use menthol than their heterosexual peers in the National Adult Tobacco Survey [3
]. The reasons for this discrepancy are unclear. We note that document research about LGB targeting by the tobacco industry has not reported specific targeting of mentholated products [14
]. Menthol is also disproportionately used by Black and African American smokers [74
]; researchers have consistently suggested that this is due to heavy targeting by the tobacco industry [31
] and linked declines in menthol in Australia to declining marketing [75
]. Because we did not find differences in retail marketing of menthol products, other avenues to explain this disparity should be explored. Physiological differences in sensory experience of menthol, which have been suggested by some researchers [76
], seem unlikely to explain differences for LGB people. However, menthol use is associated with being health conscious and a desire to quit [77
], and we believe future research should examine the role of these as potential reasons for LGB menthol disparities, given gendered and cultural differences in menthol preferences [77
Last, our research focuses on the neighborhood level of the social-ecological framework and the potential role of POS tobacco marketing therein on LGB tobacco disparities. Further research is needed on other sources of influence on these disparities, such as policies increasing the per-unit cost of tobacco products, other forms of tobacco marketing (e.g., print media, corporate sponsorship), media effects, and differential effects of tobacco use cessation interventions.
There are important limitations to this study. First, the census may underestimate same-sex households and only captures information on same-sex couples. Individual LGB people are more likely to live in more urban areas than same-sex couples [26
]. Individual sexual orientation data, which is not available in the census, would have strengthened our study, which is not generalizable to LGB individuals.
Second, reliability on some audit measures was low to moderate; we believe this is due to an up to six-week gap between audits. This may reflect expected changes in product promotions (and reliability was negatively related to the length of time between audits); audit questions such as ours generally show good reliability [78
]. But lower reliability makes it harder to detect a true effect. Given our largely null findings, this is a cause for concern.
Third, because of the national scope of this study, it was not financially viable to visit retailers who were not known to be tobacco retailers. Our phone verification protocol may have biased our study against the inclusion of smaller, independent retailers who may be in higher minority and lower income neighborhoods [49
]. Indeed, at the county level, there are differences in phone verification rates by the proportion of county population reporting African American race (rs(n = 9)
= −0.21, p
= 0.04), Hispanic ethnicity (rs(n = 97)
= −0.37, p
< 0.001), and same-sex couple households (rs(n = 97)
= −0.28, p
= 0.01). Phone verification rates do not differ significantly by median county household income (rs(n = 97)
= −0.14, p
Fourth, this study, because of its sampling strategy, was conducted in largely urban counties. Geographers have noted that assessing differences within higher density LGB areas reduces our ability to see differences across the country [79
]. That is, by focusing in urban areas, where there are overall higher concentrations to same-sex couples [20
], we may have attenuated our ability to detect differences.
Fifth, we had two limitations from our measures of product marketing. Mentholated products were limited to advertised price of a leading mentholated cigarette brand and the presence of price promotions for that same brand. Future research should examine same-sex couple rates in relation to the volume of marketing for mentholated products. Our measure of flavored cigars was not specific to little cigars and included regular cigars.
Nonetheless, this is one of the largest national audit studies to date, and past systematic reviews [78
] have identified no other studies assessing retail tobacco marketing in relation to LGB people or same-sex couples.
In a 2013 systematic review, Blosnich and colleagues noted that the minority stress model is the most frequently used conceptual approach to explaining the origin of LGB tobacco disparities [8
]. We attempted to extend this line of research to include spatial patterning of same-sex couples; however, our findings suggest that tobacco industry marketing at the store level is not disproportionately greater in neighborhoods with more same-sex couples. This is not to say existing marketing is not meaningful; even without disproportionate exposure, tobacco industry marketing may have a greater impact on LGB people than heterosexual people possibly due to LGB community appreciation for being recognized [16
]. Further research is needed to assess density in relation to marketing because greater density [72
] could cause total ads per neighborhood to be higher even with no store-level differences. Although we hope others will replicate and extend this study, it suggests that the store-level physical marketing environment may play a limited or role or no role in the origin of LGB tobacco disparities, use of mentholated products, use of flavored little cigars, and use of e-cigarettes.