In addition, emerging research suggests that adolescents’ exposure to retail marketing is associated with greater curiosity about smoking cigars15 and higher odds of ever smoking blunts.The Table summarizes descriptive statistics for store type and for schools as well as mixed models with these covariates. Nearly half of the LCC retailers near schools were convenience stores with or without gasoline/petrol. Overall, 61.5% of LCC retailers near schools contained at least one type of marijuana co-marketing: 53.2% sold blunt wraps, 27.2% sold cigarillos marketed as blunts and 26.0% sold blunt wraps, blunts or other LCC with a marijuana related “concept” flavor. After adjusting for store type, marijuana co-marketing was more prevalent in school neighborhoods with lower median household income and with a higher proportion of school-age youth.Nearly all LCC retailers sold cigarillos for $1 or less. The largest pack size at that price contained 2 cigarillos on average . The largest packs priced at $1 or less were singles in 10.9% of stores, 2-packs in 46.8%, 3-packs in 19.2%, 4-packs in 5.5%, and 5 or 6 cigarillos in 5.5%. After adjusting for store type, a significantly larger pack size of cigarillos was priced at $1 or less in school neighborhoods with lower median household income and near schools with a lower proportion of Hispanic students .In California,trim tray pollen 79% of licensed tobacco retailers near public schools sold LCCs and approximately 6 in 10 of these LCC retailers sold cigar products labeled as blunts or blunt wraps or sold cigar products with a marijuana-related flavor descriptor. A greater presence of marijuana co-marketing in neighborhoods with a higher proportion of school-age youth and lower median household income raises concerns about how industry marketing tactics may contribute to disparities in LCC use.

The study results also suggest that $1 buys significantly more cigarillos in California school neighborhoods with lower median household income. Policies to establish minimum pack sizes and prices could reduce the widespread availability of cheap cigar products and address disparities in disadvantaged areas.After Boston’s 2012 cigar regulation, the mean price for a grape-flavored cigar was $1.35 higher than in comparison communities.The industry circumvented sales restrictions in some cities by marketing even larger packs of cigarillos at the same low price, and the industry’s tipping point on supersized cigarillo packs for less than $1 is not yet known. The retail availability of 5- and 6-packs of LCCs for less than $1 observed near California schools underscores policy recommendations to establish minimum prices for multi-packs .A novel measure of marijuana co-marketing and a representative sample of retailers near schools are strengths of the current study. A limitation is that the study assessed the presence of marijuana co-marketing, but not the quantity. The protocol likely underestimates the prevalence of marijuana co-marketing near schools because we lacked a comprehensive list of LCC brands and flavor varieties. Indeed, state and local tobacco control policy research and enforcement would be greatly enhanced by access to a comprehensive list of tobacco products from the US Food and Drug Administration, including product name, category, identification number and flavor. Both a routinely updated list and product repository would be useful for tobacco control research, particularly for further identifying how packaging and product design reference marijuana use. This first assessment of marijuana co-marketing focused on brand and flavor names because of their appeal to youth.However, the narrow focus is a limitation that also likely underestimates the prevalence of marijuana co-marketing. Other elements of packaging and product design should be considered in future assessments.

Examples are pack imagery that refers to blunt making, such as the zipper on Splitarillos, as well as re-sealable packaging for cigarillos and blunt wraps, which is convenient for tobacco users who want to store marijuana. Coding for brands that are perforated to facilitate blunt making and marketing that refers to “EZ roll” should also be considered. Future research could assess marijuana co-marketing across a larger scope of tobacco/nicotine products. The same devices can be used for vaping both nicotine and marijuana. Advertising for vaping products also features compatibility with “herbs” and otherwise associates nicotine with words or images that refer to marijuana . Conducted before California legalized recreational marijuana use, the current study represents a baseline for understanding how retail marketing responds to a policy environment where restrictions on marijuana and tobacco are changing, albeit in opposite directions.The prevalence of marijuana co-marketing near schools makes it imperative to understand how tobacco marketing capitalizes on the appeal of marijuana to youth and other priority populations. How marijuana co-marketing contributes to dual and concurrent use of marijuana and tobacco warrants study, particularly for youth and young adults. In previous research, the prevalence of adult marijuana use in 50 California cities was positively correlated with the retail availability of blunts.Whether this is correlated with blunt use by adolescents is not yet known. Consumer perception studies are necessary to assess whether marijuana co-marketing increases the appeal of cigar smoking or contributes to false beliefs about product ingredients. Research is also needed to understand how the tobacco industry exploits opportunities for marijuana co-marketing in response to policies that restrict sales of flavored tobacco products and to policies that legalize recreational marijuana use. Such assessments are essential to understand young people’s use patterns and to inform current policy concerns about how expanding retail environments for recreational marijuana will impact tobacco marketing and use.In the United States, heightened levels of cultural polarization and political partisanship have magnified the role of political identity in shaping social behavior .

Recent work has found political identity to increasingly influence behavior in unprecedented ways, frequently dictating choices of one’s personal and online social networks , whom one would consider dating , and even prompting many Americans to relocate to regions more aligned with their political sympathies . Political identity is shaped by a multiplicity of factors including age, ethnicity, demography, culture, and increasingly, information sources across both traditional and social media. There is mounting evidence that political identity in the United States has impacted safety responses to the COVID-19 pandemic as well as how that risk pertains to their group identity versus individual identity . A Pew Research survey found 35% of Republicans were ”very” or ”somewhat” concerned that they would become infected with COVID-19 . In the same survey, 29% of Republicans said that people in their community should ”always wear a mask” in public. Our research is rooted in the intersection of public health and the vast and growing literature in political identity. This literature includes earlier work , and seminal work by Akerlof and Kranton on the economics of identity,indoor garden table as well as recent work that highlights the the sharpening and polarization of U.S. political identity under the Trump Administration . With respect to the pandemic, Hornsey et al. , for example, study the effect of presidential tweets on vaccine hesitancy in the U.S. while Collins et al. find political identity to exhibit a stronger effect over people’s views of the pandemic than personal impact from COVID-19. In this research, we present empirical estimates showing how political identity has shaped COVID-safety responses during the first year of the pandemic; estimate the health costs of political identity in terms of COVID cases and deaths; and test the extent to which these COVID behavioral responses and outcomes are associated with a political identity specifically tied to support for former president Donald Trump relative to more traditional strains of American conservatism. We merge U.S. county-level socioeconomic, demographic, and political data to estimate the effect of conservative political identity on COVID-safety behaviors, reported COVID cases, and deaths attributed to the virus in the first 20 months of the pandemic. After controlling for a host of county-level characteristics, employment, and demographic variables, we estimate that a 10 percentage point increase in the county popular vote for President Trump during the 2020 election to be associated with a 3.9 percentage point decrease in the number of people stating that they wear masks ”all of the time” in public, a 5.1 percentage point decrease in the COVID vaccination rate, and a 0.23σdecline in a COVID-safety behavior index. Estimates show differences in political identity during the first year of the pandemic significantly related to differences in COVID cases and deaths. We find a 10 percentage point increase in the county Trump vote to be associated with 1,394 additional COVID cases and 27.COVID-related deaths per 100,000 county residents. Moreover, the statistical relationship that we find between decreased mask-wearing and elevated COVID cases from differences in political identity is remarkably close to estimates of the average treatment effects of mask-wearing on symptomatic COVID infection obtained in the most extensive randomized controlled trial on the effects of mask-wearing . We test whether observed differences in COVID-safety behaviors, cases, and deaths can be better explained by either of two strains of traditional conservatism in the United States: American social conservatism index we construct consisting of state-level prevalence of abortion clinics, restrictions on same-sex marriage, and support for prayer in schools and American libertarian conservatism .

We find that these traditional strains of American conservatism have little systematic explanatory power over COVID-safety behaviors, cases and deaths relative to 2020 Trump voter support, which retains very strong significance over these outcomes. Our estimates on the impact of political identity on COVID-safety behaviors, cases, and deaths are robust to inclusions of different sets of control variables, demeaning and interactions of controls to address the potential for fixed-effects bias under heterogeneous effects , regularization of controls through a machine-learning algorithm, the use of Conley spatially correlated errors across states, and Oster bounds tests for endogeneity. Our COVID-19 cases and deaths data span roughly the first twenty months of the U.S. experience of the pandemic through October 2021. County-level cases, deaths and mask wearing data are taken from the New York Times COVID-19 database. Cases and deaths are reported from state and county-level health jurisdictions and generally taken from a person’s residence rather than where a person was tested or died . Mask-wearing data in the database originate from online interviews that were conducted by the global data and survey firm Dynata. The survey consists of 250,000 responses between July 2, 2020 and July 14, 2020, after the politicization of mask-wearing responses to the pandemic had taken root. Each survey participant was asked: ”How often do you wear a mask in public when you expect to be within six feet of another person?” and our data reflect the percent of respondents by county who responded ”all of the time.” We also incorporate GPS location data from a large number of mobile devices collected by the company Safe Graph to calculate the median number of devices that remained ”at home” in each county from March 1, 2020 to February 15, 2021 relative to the median that remained ”at home” during the year 2019. The mobility data provides daily observations for the total percent of devices always at home in a given census block group during the first year of the pandemic, in which citizens in many regions were often requested or required to shelter at home. We first take the median percent of devices at home for each county by day. Based on the daily median, we calculate the median percent of devices at home by month. To get the change in devices between the pandemic and pre-pandemic time, we subtract the median percent of devices at home between the pandemic and pre-pandemic periods by month. We then use this difference to obtain the change in the median percent of devices remaining at home during the pandemic months compared to pre-pandemic 2019. Our county-level vaccination data comes from the Centers for Disease Control along with the CDC’s guidelines that we use to establish our vector of control variables that are associated with heightened levels of risk for COVID infection. This county-level data is taken from the U.S. Census Bureau and includes median age, median income, population density, and percent Latino, African-American, and Asian-American in the county population. We also use the percent of county-level employment in manufacturing, services and retail to control for occupations of essential workers. It is important to control for co-morbidities in our analysis, and to do this we use the University of Wisconsin Population Health Institute county health rankings data, where each county receives a percentile score for baseline health.