Abstract
Background and Objectives:
Oklahoma has a fast-growing medical cannabis industry, showing a proliferation of industry marketing. While cannabis marketing exposure (CME) is a risk factor for cannabis use and positive attitudes about use, no studies have examined the impact of CME on attitudes and use behavior in a permissive cannabis policy environment, like Oklahoma.
Methods:
N = 5428 Oklahoma adults ages 18 and older completed assessments of demographics, past 30-day cannabis use, and past 30-day exposure to each of four types of cannabis marketing: outdoor (billboards, signs), social media, print (magazines), and Internet. Regression models examined associations of CME with positive attitudes towards cannabis use, cannabis harm perceptions, interest in obtaining a medical cannabis license (among nonlicensed participants), and past 30-day cannabis use.
Results:
Three quarters (74.5%) reported any past 30-day CME. Outdoor CME was most prevalent (61.1%), followed by social media (46.5%), Internet (46.1%), and print (35.2%). Correlates of CME included younger age, higher educational attainment and income, and medical cannabis license. In adjusted regression models, past 30-day CME and number of sources of CME were associated with current cannabis use behavior, positive attitudes about cannabis, lower cannabis harm perceptions, and greater interest in obtaining a medical cannabis license. Similar associations between CME and positive attitudes about cannabis were shown among noncannabis users.
Discussion and Conclusions:
Public health messaging should be employed to minimize the potential adverse impacts of CME.
Scientific Significance:
No studies have examined correlates of CME in a rapidly growing and relatively unrestrained marketing environment.
INTRODUCTION
While cannabis has some positive medical utility when used appropriately, such as for treating nausea and some types of epilepsy,1 unintended public health consequences may occur. Cannabis retailer density, a consequence of proliferating legal cannabis markets, has been linked to increases in cannabis-related hospitalizations,2 alcohol outlet density3 and alcohol use,4 and DUI arrests5 and crime in areas surrounding new cannabis dispensaries.3,6 Legalization also increases cannabis marketing exposure (CME), which has been shown to contribute to positive attitudes towards use,7–10 and use behavior.11
Oklahoma legalized medical cannabis in 2018 and legislated a unique and rather permissive medical cannabis policy environment.12 As of August 2022, there are 11,279 active medical cannabis business licenses approved by the Oklahoma Medical Marijuana Authority (OMMA) and 383,363 active patient licenses approved by OMMA, accounting for roughly 10% of the adult Oklahoma population. Oklahoma has more dispensaries than any other state (2335)13 in the United States that allows for legal medical or recreational cannabis, and ranks second-highest in dispensaries per capita (15.6/100,000 residents).14 As a result, cannabis advertising has proliferated.15 Furthermore, unlike other legalized cannabis states, there has been no cap on the number of dispensary licenses issued by the OMMA and fees for obtaining a patient cannabis license or a dispensary license are much lower than in other legal states, making access easy and affordable. For example, civilian patients with insurance pay a $100 nonrefundable application fee, or patients with proof of Medicare, Medicaid, or Veteran status with 100% disability pay only a $20 nonrefundable fee.16 Oklahoma also has minimal restrictions on locations and amounts of cannabis advertising.17 The only regulation on location for cannabis marketing prohibits dispensaries within 1000 feet of a school.17 Perhaps due to the rapid proliferation of patient and business licenses, the legislature issued a 2-year moratorium on new grower, processor, and dispensary licenses starting in late August 2022.
Oklahoma has long been a socially and politically conservative state, with a zero-tolerance attitude toward drug use, which is juxtaposed with the precipitous rise of cannabis dispensaries and patient license in the state. Little research exists on how exposure to cannabis marketing, in a permissive cannabis policy environment, is linked with cannabis use patterns, attitudes about use, and interest in obtaining a medical cannabis license. This study examined the impact of CME on perceptions of use and use behavior in a large sample of adults in Oklahoma. Oklahoma has quickly become a concentrated cannabis advertising environment and results can help identify marketing restrictions that may be most effective at preventing unintended health outcomes from medical cannabis markets.
METHODS
Study sample
Data were collected from English-speaking adults ages 18+ living in Oklahoma (verified by self-reported residential zip code) who completed one wave of a three-wave cross-sectional online survey. Waves were fielded 6 months apart (September 2020–September 2021). Respondents were randomly selected from a professionally maintained panel vendor, Lucid. Sampling quotas for age, gender, and race/ethnicity (using Oklahoma census data) were used to increase the likelihood that respondent demographics would be similar to the Oklahoma population. The demographic profile of Lucid panelists has been validated.18
A link to the survey was sent to potentially eligible individuals in the Lucid “marketplace” based on demographic information (state of residence, age) provided to panel suppliers. Participants completed the survey after providing consent and passing the screening that asked about age and state of residence. Surveys last 10–12 min and remained active and accessible until the sampling quota (n = 2000) was filled. Participants were compensated based on incentives provided by the panel to which they belonged (e.g., cash, gift cards, points to redeem reward prizes or gift cards; equating to roughly $1). More detail on the study methodology and data quality are published here.19 Procedures were approved by the IRB.
The sample size for each wave was based on budget as well as available respondents in the Oklahoma across Lucid panels (n = 2000 approximately). Demographic quotas were reached within a 3%–5% standard deviation of the state census data.20 Participants who completed more than one wave (n = 145) were identified and data from their most recent survey were retained for analyses.
Measures
To assess CME, participants were asked whether they saw each of the following in the past 30-days: (1) “ads or promotions (e.g., billboards, posters) for a cannabis dispensary” (e.g., outdoor marketing); (2) “social media user-generated content related to cannabis products (e.g., social media posts, discussion threads on social media, photos, or videos)”; (3) “ads or promotions for cannabis products or dispensaries when reading a newspaper or a magazine”; and (4) “ads or promotions for cannabis products when using the Internet.” Two variables were created for analyses: any cannabis marketing in the past 30-days (yes/no), and the total number of sources of CME in the past 30-days (range: 0–4).
Positive attitudes towards cannabis use were measured with three items: (1) cannabis is not addictive; (2) cannabis is safer to use than prescription pain medications; and (3) cannabis is an effective treatment for general health conditions; with five response options for each item: (0) strongly disagree, (1) disagree, (2) neither agree nor disagree, (3) agree, and (4) strongly agree. Responses were averaged to create a total score, where higher scores indicated more positive attitudes (range: 0–4). Cronbach’s α = .80.
To assess cannabis harm perceptions, participants were asked: “how much do you think people harm themselves when they use cannabis” with four response options: (0) no harm, (1) a little harm, (2) some harm, and (3) a lot of harm. Responses were dichotomized such that higher scores indicated lower harm perceptions (no/little harm = 1 vs. some/lot of harm = 0).
Respondents who did not have a medical cannabis license issued by OMMA were asked: “please rate your level of interest in obtaining a medical marijuana card in the state of Oklahoma,” with five response options: (0) not interested, (1) somewhat interested, (2) moderately interested, (3) very interested, and (4) extremely interested.
Past 30-day cannabis use was asked differently in Wave 1 versus Waves 2 and 3. In Wave 1, past-year cannabis users were asked, in two separate follow-up questions, to report the number of days in the past 30-days they used cannabis for medical reasons (e.g., “like to treat or decrease symptoms of a health condition”) and for recreational reasons (e.g., “for pleasure or satisfaction, to have fun”). In Waves 2 and 3, past year cannabis users were asked to report the number of days they used cannabis in the past 30-days (for any reason). Those who reported using cannabis ≥1 day in the past 30-days were then asked to report, in two separate follow-up questions, the number of days in the past 30-days they used cannabis for medical and for recreational reasons. Across all waves of data, respondents who used cannabis ≥1 day for either recreational and/or medical purposes were coded as past 30-day cannabis users (1); those who did not use any cannabis in the past 30-days were coded as nonusers (0/reference).
Analyses
The final analytic sample consisted of 5284 participants ages 18 and older.
The prevalence and demographic correlates of those exposed and not exposed to CME in the past 30-days were assessed using unweighted cross-tabulations. Next, multivariable logistic regression models examined demographic correlates of any CME and each individual source of CME. Finally, multivariable linear or logistic regression models (depending on the outcome) were used to examine the association of each source of CME, any CME, and total number of sources of CME with the following outcomes: (1) positive attitudes towards cannabis use, (2) cannabis harm perceptions, (3) interest in obtaining a medical cannabis license (among nonlicense holders), and (4) past 30-day cannabis use (yes/no). All models controlled for age, sex, race/ethnicity, education, employment status, health insurance status, and income. Past 30-day cannabis use and possession of a medical cannabis license were also included as covariates in models where they were not the primary outcomes. The proportion of respondents reporting any CME did not differ across waves, so wave was not included as a covariate. Analyses were completed in SAS 9.4. Missing data comprised ≤5% for all analyses.
RESULTS
Table 1 shows the characteristics of the full sample and differences across those exposed and not exposed to CME. The sample was closely aligned (within 3%–5%) with the racial and ethnic demographics of Oklahoma.20,21 The sample was 70.4% non-Hispanic (NH) White, 7.9% NH Black, 7.5% American Indian/Alaskan Native, 9.5% Hispanic, and 4.7% NH Other. Relative to the state population (16%), there was a slightly lower proportion of participants in our sample aged 65+ (10.4%)and a slightly higher proportion of females (56.7% in the study sample vs. 50.2% in the state of Oklahoma). Slightly more than half of the sample was employed full- or part-time (53.7%) and just over a third (37.9%) had at least some college education/technical degree/associates degree (37.9%). Just over one-fifth had no health insurance (21.8% vs. 18% for state population), and a quarter (25.7%) reported an annual income below $20,000. Regarding cannabis use behavior and attitudes, 22.2% of participants possessed a medical cannabis license, and 34.8% reported past 30-day cannabis use. Among the full sample with complete data on medical and recreational cannabis use (n = 5239), 9.9% (n = 520) reported past 30-day medical but not recreational use, 65.5% (n = 3435) reported no past 30-day use of either, 4.1% (n = 217) reported past 30-day use of recreational but not medical, and 20.3% (n = 1067) reported past 30-day use of both medical and recreational cannabis. Most participants perceived “no” or “a little” harm from using cannabis (74.3%), and almost half with no medical cannabis license were at least “somewhat” interested in obtaining one (47.1%). The sample endorsed mostly positive attitudes about cannabis (M = 2.3 out of 4).
TABLE 1.
Characteristics of the study sample (n = 5428) and differences by past 30-day cannabis marketing exposure.
| Variable | N (%) or M (SD) | Any cannabis marketing exposure in the past 30 daysa | P | |
|---|---|---|---|---|
| No | Yes | |||
| (n=1370, 25.5%) | (n=4011, 74.5%) | |||
| Age | ||||
| 18–24 | 1123 (20.7%) | 271 (19.8%) | 837 (20.9%) | <.0001 |
| 25–34 | 1193 (22.0%) | 272 (19.9%) | 913 (22.8%) | |
| 35–44 | 1134 (20.9%) | 237 (17.3%) | 892 (22.2%) | |
| 45–54 | 779 (14.4%) | 199 (14.5%) | 574 (14.3%) | |
| 55–64 | 635 (11.7%) | 201 (14.7%) | 425 (10.6%) | |
| 65 and older | 564 (10.4%) | 190 (13.9%) | 370 (9.2%) | |
| Sex | ||||
| Male | 2350 (43.3%) | 564 (41.2%) | 1759 (43.9%) | .0830 |
| Female | 3078 (56.7%) | 806 (58.8%) | 2252 (56.2%) | |
| Race/ethnicity | ||||
| White, non-Hispanic (NH) | 3816 (70.4%) | 974 (71.1%) | 2814 (70.2%) | .1995 |
| Black, NH | 430 (7.9%) | 113 (8.3%) | 308 (7.7%) | |
| American Indian/Alaskan Native, NH | 408 (7.5%) | 100 (7.3%) | 305 (7.6%) | |
| Other, non-Htspanicb | 256 (4.7%) | 73 (5.3%) | 182 (43%) | |
| Hispanic | 513 (9.5%) | 110 (8.0%) | 399 (10.0%) | |
| Education level | ||||
| Less than 12 years | 403 (7.4%) | 129 (9.4%) | 267 (6.7%) | <.0001 |
| High school diploma or GED | 1470 (27.1%) | 420 (30.7%) | 1036 (25.8%) | |
| Some college/technical school/Associate’s degree | 2056 (37.9%) | 467 (34.1%) | 157 (39.2%) | |
| Bachelor’s degree or higher | 1496 (27.6%) | 353 (25.8%) | 1136 (28.3%) | |
| Employmentc | ||||
| Employed–full-time/part-time | 2914 (53.7%) | 641 (46.8%) | 2252 (56.2%) | <.0001 |
| Unemployed–looking for work | 589 (10.9%) | 155 (11.3%) | 425 (10.6%) | |
| Unemptoyed–not looking for work | 1767 (32.6%) | 535 (39.1%) | 1218 (30.4%) | |
| Other | 158 (2.9%) | 39 (2.9%) | 115 (2.9%) | |
| Health insuranced | ||||
| Private | 2048 (37.8%) | 440 (32.1%) | 1598 (39.9%) | <.0001 |
| Government | 2193 (40.4%) | 579 (42.3%) | 1591 (39.7%) | |
| No insurance | 1182 (21.8%) | 351 (25.6%) | 819 (20.4%) | |
| Household income | ||||
| Less than $20,000 | 1394 (25.7%) | 429 (31.3%) | 941 (23.5%) | <.0001 |
| $20,000–$49,999 | 1736 (32.0%) | 425 (31.0%) | 1299 (32.4%) | |
| $50,000–$99,999 | 1402 (25.8%) | 296 (21.6%) | 1099 (27.4%) | |
| $100,000 or more | 610 (11.2%) | 119 (8.7%) | 490 (12.2%) | |
| Refused to answer | 286 (5.3%) | 101 (7.4%) | 182 (4.5%) | |
| Past 30-day cannabis usee | ||||
| No | 3435 (65.5%) | 1064 (79.9%) | 2343 (60.5%) | <.0001 |
| Yes | 1810 (34.5%) | 267 (20.1%) | 1532 (39.5%) | |
| Medical cannabis license | ||||
| No | 4221 (77.8%) | 1198 (87.5%) | 2986 (74.5%) | <.0001 |
| Yes | 1203 (22.2%) | 172 (12.6%) | 1024 (25.5%) | |
| Positive attitudes toward cannabisf | 2.3 (1.1) | 2.1 (LI) | 2.4 (1.0) | <.0001 |
| Harm perceptions of cannabis use | ||||
| No harm | 2509 (46.3%) | 602 (44.0%) | 1889 (47.1%) | <.0001 |
| A little harm | 1526 (28.2%) | 358 (26.2%) | 1156 (28.8%) | |
| Some harm | 914 (16.9%) | 244 (17.8%) | 662 (16.5%) | |
| A lot of harm | 472 (8.7%) | 165 (12.1%) | 302 (73%) | |
| Interest in obtaining a cannabis licenseg | ||||
| Not interested | 2230 (52.8%) | 792 (66.1%) | 1419 (47.5%) | <.0001 |
| Somewhat interested | 700 (16.6%) | 153 (12.8%) | 544 (18.2%) | |
| Moderately interested | 400 (9.5%) | 93 (7.8%) | 302 (10.1%) | |
| Very interested | 342 (8.1%) | 58 (4.8%) | 280 (9.4%) | |
| Extremely interested | 548 (13.0%) | 102 (8.5%) | 440 (14.7%) | |
| Wave | ||||
| 1 | 1806 (33.3%) | 429 (31.3%) | 1368 (34.1%) | .1063 |
| 2 | 1826 (33.6%) | 461 (33.7%) | 1342 (33.5%) | |
| 3 | 1796 (33.1%) | 480 (35.0%) | 1301 (32.4%) | |
Note: Missing data was less than 5% for each variable.
Participants (n = 47) who had any missing data for cannabis marketing exposure were not included in analysis.
Other included Asians (n = 65), Native Hawaiians or other Pacific Islanders (n = 16), and those who reported more than one race (n = 175).
Employed included participants who reported full-time (n = 2273) and part-time work (n = 641). Not employed (not looking for work) included participants who reported not looking for work (n = 105), homemakers (n = 390), students (n = 232), unable to work or disabled (n = 578), and the retired (n = 462).
Government insurance included participants who reported having Medicare (n = 916), Medicaid (n = 989), and Military or Veteran’s Affair Insurance (n = 288).
Participants who were inconsistent in their responses about cannabis use (N = 34) were not included in analyses that included cannabis use as a covariate. Total number of respondents who provided complete data on past 30-day cannabis use was n = 5394.
Positive attitudes about cannabis scores (n = 5413) were based on responses to three items: (1) cannabis is not addictive; (2) cannabis is safer to use than prescription pain medications; and (3) cannabis is an effective treatment for general health conditions. Participants were given five response options for each question: (0) strongly disagree, (1) disagree, (2) neither agree nor disagree, (3) agree, and (4) strongly agree. Total scores based on participants’ responses were calculated and averaged, with higher scores indicating more positive attitudes toward cannabis use (min = 0, max = 4).
Among nonlicense holders; n = 4221.
Prevalence and correlates of cannabis marketing exposure
Three-quarters of the sample (74.5%) reported past 30-day CME. In Table 2 of logistic regression model results, compared to adults ages 65+, adults ages 18–24 (aOR = 1.57), 25–34 (aOR = 1.36), and 35–44 (aOR = 1.46) had greater odds of reporting past 30-day CME. Adults who completed some college education/graduated with an Associate’s degree, or completed technical school, had greater odds of reporting CME than adults with less than 12 years of education (aOR = 1.47). Adults with no health insurance (vs. private insurance) had lower odds of reporting past 30-day CME (aOR = 0.70). Compared to adults income <$20,000, those earning $20,000 and $49,999 (aOR = 1.35), $50,000 and $99,999 (aOR = 1.63), or $100,000+(aOR = 1.74) had higher odds of reporting past 30-day CME. Those who reported past 30-day cannabis use (aOR = 2.36) and those with a medical cannabis license had higher odds of reporting CME (aOR = 1.31).
TABLE 2.
Logistic regression model of the associations of demographic characteristics with any cannabis marketing exposure in the past 30-days.
| Variable | Any cannabis marketing exposure in the past-30 days AOR (95% Cl) |
|---|---|
| Age | |
| 18–24 | 1.57 (1.20, 2.05) |
| 25–34 | 1.36 (1.05, 1.77) |
| 35–44 | 1.46 (1.13, 1.90) |
| 45–54 | 1.20 (0.91, 1.57) |
| 5–64 | 0.98 (0.76, 1.28) |
| 65 and older | Ref |
| Sex | |
| Male | Ref |
| Female | 1.00 (0.88, 1.15) |
| Race/ethnicity | |
| White, non-Hispanic (NH) | Ref |
| Black, NH | 0.88 (0.69, 1.13) |
| American Indian/Alaskan Native, NH | 1.06 (0.82, 1.36) |
| Other, NH | 0.83 (0.61, 1.11) |
| Hispanic | 1.08 (0.85, 1.37) |
| Education level | |
| Less than 12 years | Ref |
| High school diploma or GEO | 1.10 (0.85, 1.43) |
| Some college/technical school/Associate’s degree | 1.47 (1.13, 1.91) |
| Bachelor’s degree or higher | 1.24 (0.93, 1.65) |
| Employment | |
| Employed–full-time or part-time work | Ref |
| Unemployed–looking for work | 0.95 (0.75, 1.19) |
| Unemployed–not looking for work | 0.86 (0.73, 1.02) |
| Other | 1.40 (0.92, 2.14) |
| Health Insurance | |
| Private | Ref |
| Government | 0.87 (0.73, 1.03) |
| No insurance | 0.70 (0.57, 0.84) |
| Household income | |
| Less than $20,000 | Ref |
| $20,000–$49.999 | 1.35 (1.13, 1.61) |
| $50,000–$99.999 | 1.63 (1.32, 2.00) |
| $100.000 or more | 1.74 (1.32, 2.30) |
| Refused to answer | 0.89 (0.67, 1.19) |
| Past 30-day cannabii uvc | |
| No | Ref |
| Yes | 2.36 (1.97, 2.83) |
| Medical cannabb license | |
| No | Ref |
| Yes | 1.31 (1–06, 1.62) |
Note: Bolded text indicates significance (p < .05).
Abbreviations: AOR, adjusted odds ratio; CI, confidence interval.
In Supporting Information: Table S1, outdoor marketing was the most common source of CME in the past 30-days (61.1%), followed by social media marketing (46.5%), internet marketing (46.1%), and print marketing (35.2%). Younger age groups were more likely to report social media CME (aOR = 2.03–3.50) and internet CME (aOR = 1.36–1.62) compared to adults 65+. Females (aOR = 1.14) were more likely to report outdoor CME and less likely to report internet CME (aOR = 0.87). Past 30-day cannabis users were more likely to report exposure to all sources of cannabis marketing (aOR = 1.44–1.91), and those with a medical cannabis license were also more likely to report to exposed to all sources of cannabis marketing (aOR = 1.21–1.31), except outdoormarketing. Compared to NH White respondents, Hispanic respondents (aOR = 1.32) and NH Other (aOR = 1.41) respondents were more likely to report print CME. Those with some college education (aOR = 1.47), and those with higher household incomes (aOR = 1.28–1.45) were more likely to report outdoor CME, while Hispanic respondents (aOR = 0.79) and individuals with government health insurance (aOR = 0.75) or no health insurance (aOR = 0.74) were less likely to report outdoor CME. Being unemployed and not looking for work was associated with lower odds of reporting print CME (aOR = 0.77).
Associations of CME with cannabis attitudes and behavior
In Table 3, any past 30-day CME and exposure to a greater number of cannabis marketing sources were associated with more positive cannabis attitudes, greater interest in obtaining a medical cannabis license (among nonlicense holders), and increased odds of past 30-day cannabis use (p’s < 0.0001). Exposure to a greater number of cannabis marketing sources was associated with increased odds lower cannabis harm perceptions (aOR = 1.05).
TABLE 3.
Linear and logistic regression models of associations of cannabis marketing exposure with attitudes and use behavior in the full sample.
| Variable | Positive attitudes about cannabis usea | Interest in obtaining a medical cannabis licenseb | Cannabis harm perceptions (no/a little vs. some/a lot)a AOR (95% CI) | Past 30-day cannabis usec AOR (95% CI) | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| b | SE | β | P | b | SE | β | P | |||
| Any source of marketing | ||||||||||
| No | Ref | Ref | Ref | Ref | ||||||
| Yes | 0.18 | 0.03 | .07 | <.0001 | 0.26 | 0.04 | .08 | <.0001 | 1.15 (0.99,1.34) | 2.65 (2.27, 3.10) |
| Number of marketing sources | 0.09 | 0.01 | .11 | <.0001 | 0.10 | 0.01 | .10 | <.0001 | 1.05 (1.00, 1.10) | 1.30 (1.25, 1.36) |
| Specific sources of cannabis marketing exposure d | ||||||||||
| Outdoor marketing | ||||||||||
| No | Ref | Ref | Ref | Ref | ||||||
| Yes | 0.12 | 0.03 | .05 | <.001 | 0.05 | 0.04 | .02 | .24 | 1.34 (1.14, 1.56) | 1.38 (1.19, 1.60) |
| Social media marketing | ||||||||||
| No | ||||||||||
| Yes | 0.15 | 0.03 | .07 | <.001 | 0.23 | 0.05 | .08 | <.001 | 1.22 (1.03, 1.44) | 1.83 (1.58, 2.12) |
| Print marketing | ||||||||||
| No | ||||||||||
| Yes | −0.07 | 0.03 | −.03 | .03 | 0.00 | 0.05 | .00 | .98 | 0.72 (0.61, 0.85) | 1.10 (0.96, 1.28) |
| Internet marketing | ||||||||||
| No | ||||||||||
| Yes | 0.09 | 0.03 | .04 | <.001 | 0.12 | 0.05 | .04 | .01 | 1.02 (0.86, 1.21) | 1.04 (0.89, 1.21) |
Note: Bolded text indicates significance for AOR (p < .05). Cannabis harm perceptions coded as 1 = no/a little and 0 = some/a lot. Abbreviations: b, unstandardized beta; β, standardized beta.
Model included age, sex, race/ethnicity, education, income, health insurance, employment status, past 30-day cannabis use, and possession of medical cannabis license as covariates.
Examined only among respondents without a medical cannabis license. Model included age, sex, race/ethnicity, education, income, health insurance, employment status, and past 30-day cannabis use as covariates.
Model included age, sex, race/ethnicity, education, income, health insurance, and employment status as covariates.
Cannabis marketing sources were analyzed in the same model.
When controlling for all sources of cannabis marketing in the same model, outdoor (β = .05), social media (β = .07), and internet CME (β = .04) were all associated with greater positive cannabis attitudes (p’ < .01). Print CME was associated with lower positive cannabis attitudes (β = −.03, p < .05). Social media (β = .08) and Internet CME (β = .04) were both associated with greater interest in obtaining a medical cannabis license (p’s < .01). Outdoor and social media CME were both associated with lower cannabis harm perceptions (outdoor aOR = 1.34 and social media aOR = 1.22) and past 30-day cannabis use (outdoor aOR = 1.38 and social medial aOR = 1.83). Print marketing exposure was associated with lower cannabis harm perceptions (aOR = 0.72).
Post hoc analyses
To determine if the associations between CME and cannabis use attitudes were primarily due to current cannabis use, models were reanalyzed among noncannabis users only. Results were similar to those of the full sample (Table 4), wherein any past 30-day CME and a greater number of sources of marketing exposure were positively associated with all outcomes of interest, controlling for demographics.
TABLE 4.
Linear and logistic regression models of the associations of cannabis marketing exposure with cannabis use outcomes among noncurrent cannabis users.
| Variable | Positive attitudes about cannabis use | Interest in obtaining medical cannabis licensea | Cannabis harm perceptions (no/a little vs. some/a lot) aOR (95% CI) | ||||||
|---|---|---|---|---|---|---|---|---|---|
| b | SE | β | p | b | SE | β | p | ||
| Any source of marketing | |||||||||
| No | Ref | Ref | Ref | ||||||
| Yes | 0.21 | 0.04 | .10 | <.0001 | 0.25 | 0.04 | .10 | <.0001 | 1.19 (1.01, 1.40) |
| Number of marketing sources | 0.09 | 0.01 | .13 | <.0001 | 0.10 | 0.01 | .13 | <.0001 | 1.07 (1.01, 1.13) |
Note: All models included age, sex, race/ethnicity, education, income, health insurance, employment status. Bolded text indicates significance (p < .05).
Abbreviations: aOR, adjusted odds ratio; b, unstandardized beta; β, standardized beta; CI, confidence interval.
Analysis restricted to those without a medical cannabis license.
DISCUSSION
Three quarters of our sample reported past month exposure to some form of cannabis marketing, which is higher than other published work. A repeated cross-sectional online survey conducted by the Oregon Public Health Division in 2015 and 2016 showed that just over 50% of their sample of n = 3885 adults reported exposure to cannabis advertising in the past month.22 Rup and colleagues (2020) found that 56.7% of adults in United States cannabis legal states (n = 7292) reported exposure to some form of cannabis advertising in the past month.23 Higher rates of exposure in Oklahoma could be due to the greater density of dispensaries in the state, leading to an increased need for advertising in a highly competitive environment.14 Despite Oklahoma’s continuing trend toward social and political conservatism, there is a saturation of cannabis marketing in the state, even though cannabis remains illegal at the federal level.
Consistent with previous studies, outdoor marketing was the most common source of exposure,22,23 reported by 61.1% of the sample. The most significant sociodemographic correlates of any past 30-day CME were younger age, higher income and educational attainment, private health insurance, past 30-day cannabis use, and possessing a medical cannabis license. Younger individuals may report more CME due to the variety of locations they visit in everyday life relative to older adults whose daily activities may be limited to work and home. Those with a medical cannabis license and current cannabis users may more easily recognize marketing content making it more memorable to them, or may come into contact with marketing because of regular visits to dispensaries. It was surprising that lower-income individuals, those with lower educational attainment, and individuals with no health insurance were less likely to be exposed to CME, given that these individuals are more likely to report current cannabis use.24 They may have been more likely to live in rural areas, and thus less likely to be exposed to outdoor marketing, as marketing tends to cluster in densely populated areas.25 It could also be possible that the legal cannabis industry targets higher-income residents, and does so primarily online. Lower-income respondents may also have been less likely to have access to a car and were thus less likely to be exposed to marketing through daily travel. Lastly, we did not see differences across sex, racial and ethnic groups, or employment status on CME, which is consistent with previous work.22,23,26
Even after controlling for a variety of other factors correlated with cannabis use, the associations between any CME in the past 30-days and number of cannabis marketing sources with attitudes and use behavior remained significant. This suggests, not surprisingly, that cannabis marketing is an effective tool at influencing consumers, much like marketing in the alcohol and tobacco industries. It is likely that even cannabis naïve individuals desire to purchase cannabis products because they are inundated with marketing. While interest may not convert to behavior, it can be a proxy for future use.27 CME maintained a robust association with cannabis-related attitudes (e.g., low perceptions of harm, positive attitudes about use) even after adjusting for license status and when restricting analyses to noncurrent cannabis users. This suggests the link between CME and positive perceptions about cannabis is not entirely due to current use behavior. Because our study is cross-sectionalwe do not know if observed associations indicate that cannabis marketing is influencing attitudes and use behavior, or that those who use cannabis and those who are more likely to be positively disposed to medical cannabis use are more attentive to pro-cannabis messaging in their environment.
In regression models that simultaneously controlled for the potential overlap among all CME sources, those exposed to outdoor marketing and social media marketing reported more positive attitudes about cannabis use and were more likely to report past 30-day cannabis use and lower cannabis harm perceptions. This latter finding is particularly alarming, as studies consistently cite a link between lower cannabis harm perceptions and greater likelihood of problematic cannabis use.28 Finally, exposure to print marketing was negatively associated with positive attitudes about cannabis and lower perceptions harm. In our sample, multivariable regression models showed that print marketing exposure was most common among mid-life adults (ages 25–44), who may have less positive attitudes about cannabis use overall, and who also have lower rates of internet and social media use compared to young adults.29 Cannabis content and marketing has proliferated on social media30,31 and a recent systematic review shows that substance use content on social media frequently promotes engagement in, and positive sentiment about substance use.32 Thus, age may be driving the relationship between exposure to print marketing (vs. other marketing sources) with attitudes about cannabis use.
Several implications can be gleaned from our findings. First, the high rates of past 30-day CME in our sample are not surprising, given the high number of dispensaries in Oklahoma relative to other states,13 which may contribute to highly prevalent marketing. Licensing caps should be considered in emerging cannabis markets to curb the impact of marketing. Restrictions on social media marketing may be difficult to implement given ongoing debate about how to balance First Amendment rights with the portrayal of potentially harmful content on these platforms.33 Increases in revenue from state and local sales tax of cannabis products also make it difficult to restrict marketing.34 Government officials and dispensary business owners may be concerned that marketing restrictions will negatively impact economic growth. Lastly, messaging campaigns that “inoculate” the public against cannabis marketing persuasion may help counteract the abundance of pro-cannabis marketing in the environment.
This study had several limitations. First, we cannot determine from our study whether CME impacts attitudes and use or whether attitudes and use predispose one to notice more marketing. It is worth noting that the CME’s associations with positive attitudes about cannabis use and lower s harm perceptions were significant even when models examined only nonusers. Second, we did not examine changes over time in cannabis use and attitudes as a function of CME or policy changes that may have affected cannabis marketing over time. Third, spatial factors that could impact exposure were not assessed, such as the density of cannabis dispensaries in someone’s neighborhood. Fourth, it is possible that some individuals driving by or seeing a dispensary reported this as advertising/promotion, however, we did not assess differences in exposure to outdoor advertising for a cannabis dispensary versus a billboard. Finally, the method of sampling via an existing panel produced an over-representation of medical cannabis licenses (22% vs. 10% in state), and nearly a third used cannabis in the past month, higher than the prevalence rate for Oklahoma (14.19%) reported in the 2019–2020 National Survey of Drug Use and Health (NSDUH).35 The NSDUH, however, queries about cannabis use more generally, while we asked about both medical and recreational cannabis use in the past 30-days, which may have resulted in higher self-reports. Given limitations on recruiting such a large sample in Oklahoma, the sample could not be fully representative of the state, although our demographics for age group, sex, race/ethnicity, income, and insurance were closely aligned with state estimates (±3%–5%).20
CONCLUSIONS
Given the ensuing rise in advertising for cannabis products in medical cannabis markets, the potential impacts of CME on increased appeal and desire to use should be considered in regulation (e.g., content, density, location) and in public health messaging to minimize unintended public health impacts. Before implementing a new cannabis market, local authorities should consider caps on the number and location of dispensaries and restrictions on billboards and other outdoor signage to help deter problematic cannabis use. While first amendment rights prohibit most forms of regulation around content of outdoor advertising, buffers around sensitive areas, like schools, churches, and parks, may be more aptly considered by policymakers.
Supplementary Material
ACKNOWLEDGMENTS
This study was support in part by the University of Oklahoma Health Sciences Center, Oklahoma Tobacco Settlement Endowment Trust (TSET) contract # R22-03 and NCI grant P30CA2255520 awarded to the Stephenson Cancer Center. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Funding information
Division of Cancer Prevention, National Cancer Institute, Grant/Award Number: P30CA2255520; Oklahoma Tobacco Settlement Endowment Trust, contract # R22-03
Footnotes
CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.
SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.
REFERENCES
- 1.National Academies of Sciences E, and Medicine. The Health Effects of Cannabis and Cannabinoids: The Current State of Evidence and Recommendations for Research. National Academies Press; 2017. [PubMed] [Google Scholar]
- 2.Mair C, Sumetsky N, Kranich C, Freisthler B. Availability of medical cannabis dispensaries and cannabis abuse/dependence-related hospitalizations in California. Addiction. 2021;116:1908–1913. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Subica AM, Douglas JA, Kepple NJ, Villanueva S, Grills CT. The geography of crime and violence surrounding tobacco shops, medical marijuana dispensaries, and off-sale alcohol outlets in a large, urban low-income community of color. Prev Med. 2018;108: 8–16. [DOI] [PubMed] [Google Scholar]
- 4.García-Ramírez G, Paschall MJ, Grube JW. Retail availability of recreational marijuana and alcohol in Oregon Counties and co-use of alcohol and marijuana and related beliefs among adolescents. Subst Use Misuse. 2020;1:1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hunt P, Pacula RL, Weinberger G. High on crime? Exploring the effects of marijuana dispensary laws on crime in California counties IZA Institute of Labor Economics, Discussion Paper Series; 2018.
- 6.Hughes LA, Schaible LM, Jimmerson K. Marijuana dispensaries and neighborhood crime and disorder in Denver, Colorado. Justice Q. 2020;37(3):461–485. [Google Scholar]
- 7.Trangenstein PJ, Whitehill JM, Jenkins MC, Jernigan DH, Moreno MA. Cannabis marketing and problematic cannabis use among adolescents. J Stud Alcohol Drugs. 2021;82(2):288–296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.D’Amico EJ, Rodriguez A, Tucker JS, Pedersen ER, Shih RA. Planting the seed for marijuana use: changes in exposure to medical marijuana advertising and subsequent adolescent marijuana use, cognitions, and consequences over seven years. Drug Alcohol Depend. 2018;188:385–391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Shih RA, Rodriguez A, Parast L, et al. Associations between young adult marijuana outcomes and availability of medical marijuana dispensaries and storefront signage. Addiction. 2019;114:2162–2170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Shih RA, Tucker JS, Pedersen ER, et al. Density of medical and recreational cannabis outlets: racial/ethnic differences in the associations with young adult intentions to use cannabis, e-cigarettes, and cannabis mixed with tobacco/nicotine. J Cannabis Res. 2021;3(1):28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dai H Exposure to advertisements and marijuana use among US adolescents. Prev Chronic Dis. 2017;14:170253. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Schroyer J Why red Oklahoma is home to a booming medical marijuana market. MJBizDaily. Accessed September 10, 2021. https://mjbizdaily.com/why-red-oklahoma-is-home-to-a-booming-medical-marijuana-market/ [Google Scholar]
- 13.Scavelli M Oklahoma cities top list of most medical marijuana dispensaries across the US. Fox 25. Accessed January 23, 2020. https://okcfox.com/news/local/oklahoma-cities-top-list-of-most-medical-marijuana-dispensaries-across-the-us
- 14.Cooper J Oklahoma has the 2nd highest number of marijuana dispensaries per capita. Oklahoma’s Own News on 6. Accessed February 18, 2020. https://www.newson6.com/story/5e6279d7cd4aa89d1b92f272/oklahoma-has-the-2nd-highest-number-of-marijuana-dispensaries-per-capita [Google Scholar]
- 15.Dulaney J In Oklahoma, cannabis billboards irk some. The Oklahoman. https://tulsaworld.com/news/local/marijuana/in-oklahoma-cannabis-billboards-irk-some/article_3f4cdb8b-67e3-570d-9583-c85b93732b21.html [Google Scholar]
- 16.Oklahoma Medical Marijuana Authority. Patient Licenses. Accessed August 5, 2022. https://oklahoma.gov/omma/patients-caregivers/patient-licenses.html#minor-patient-licenses
- 17.Oklahoma State Department of Health. Title 310. Oklahoma State Department of Health Chapter 681. Medical Marijuana Regulations; 2020:38. [Google Scholar]
- 18.Coppock A, McClellan OA. Validating the demographic, political, psychological, and experimental results obtained from a new source of online survey respondents. Res Politics. 2019;6(1): 2053168018822174. [Google Scholar]
- 19.Kendzor DE, Ehlke SJ, Kharazi Boozary L, Smith MA, Cohn AM. Characteristics of adults with a medical cannabis license, reasons for use, andperceptions of benefit following medical cannabis legalization in Oklahoma. Prev Med Rep. 2022;27:101777. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.U.S. Census Bureau. Annual County Resident Population Estimates by Age, Sex, Race, and Hispanic Origin: April 2010 to July 1, 2020. Population Division; 2021. Release date: June 2021. [Google Scholar]
- 21.United States Census Bureau. Oklahoma: 2020 Census. Accessed August 18, 2022. https://www.census.gov/quickfacts/OK
- 22.Fiala SC, Dilley JA, Firth CL, Maher JE. Exposure to marijuana marketing after legalization of retail sales: oregonians’ experiences, 2015–2016. Am J Public Health. 2018;108(1):120–127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Rup J, Goodman S, Hammond D. Cannabis advertising, promotion and branding: differences in consumer exposure between ‘legal’ and ‘illegal’ markets in Canada and the US. Prev Med. 2020; 133:106013. [DOI] [PubMed] [Google Scholar]
- 24.Hasin DS, Shmulewitz D, Sarvet AL. Time trends in US cannabis use and cannabis use disorders overall and by sociodemographic subgroups: a narrative review and new findings. Am J Drug Alcohol Abuse. 2019;45(6):623–643. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Lowery B Outdoor advertising. Interdiscip J Signage Wayfinding. 2019;3(2):21–28. [Google Scholar]
- 26.Trangenstein PJ, Whitehill JM, Jenkins MC, Jernigan DH, Moreno MA. Active cannabis marketing and adolescent past-year cannabis use. Drug Alcohol Depend. 2019;204:107548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Pierce JP, Choi WS, Gilpin EA, Farkas AJ, Merritt RK. Validation of susceptibility as a predictor of which adolescents take up smoking in the United States. Health Psychol. 1996;15(5):355–361. [DOI] [PubMed] [Google Scholar]
- 28.Mariani AC, Williams AR. Perceived risk of harm from monthly cannabis use among US adolescents: National Survey on drug use and health, 2017. Prev Med Rep. 2021;23:101436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Pew Research Foundation. Internet/Broadband Fact Sheet. https://www.pewresearch.org/internet/fact-sheet/internet-broadband/?menuItem=480dace1-fd73-4f03-ad88-eae66e1f4217
- 30.Rutherford BN, Sun T, Johnson B, et al. Getting high for likes: exploring cannabis-related content on TikTok. Drug Alcohol Rev. 2022;41(5):1119–1125. [DOI] [PubMed] [Google Scholar]
- 31.Whitehill JM, Trangenstein PJ, Jenkins MC, Jernigan DH, Moreno MA. Exposure to cannabis marketing in social and traditional media and past-year use among adolescents in states with legal retail cannabis. J Adolesc Health. 2020;66(2):247–254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Rutherford BN, Lim CC, Johnson B, et al. # TurntTrending: a systematic review of substance use portrayals on social media platforms. Addiction. 2022;118(2):206–217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Hooker MP. Censorship, free speech & facebook: applying the first amendment to social media platforms via the public function exception. Wash JL Tech Arts. 2019;15:36. [Google Scholar]
- 34.Oklahoma Medical Marijuana Authority. Licensing Reports, Tax Revenue Reports, and County Totals. Accessed August 12, 2022. https://oklahoma.gov/omma/help/licensing-and-tax-data.html
- 35.Substance Abuse and Mental Health Services Administration. 2019–2020 National Survey of Drug Use and Health Model-Based Prevalence Estimates (50 States and the District of Columbia). 2021. https://www.samhsa.gov/data/sites/default/files/reports/rpt35339/NSDUH%20Attachments/NSDUH%20Attachments/NSDUHsaePercents2020.pdf
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.