California became the first state to pass a comprehensive medical marijuana law in 1996, and as of 2017, 29 states in the United States and Washington, DC have legalized marijuana for medical purposes. Recent high-quality epidemiological studies have examined changes in overall marijuana use rates among adolescents before and after the passage of medical marijuana legalization laws in an attempt to examine whether marijuana use rates have increased, decreased, or stayed the same following legalization. Due to heterogeneity across studies and nuances in policy , there is no definitive conclusion . What research has shown, however, is a strong trend towards more positive views of marijuana among teens over the past 14 years . For example, more than 50% of 10th and 12th graders across the United States now endorse the belief that smoking marijuana regularly does not carry great risk . Research has also shown that more positive views about marijuana among teens are associated with increased marijuana use rates in this age group . Many of these positive beliefs about marijuana may come from exposure to marijuana positive messages on social media and through advertising ,how to dry cannabis which has increased as MMLs have passed . For example, a 2014 social media study showed that among people ages 17–19 years, the popular pro-marijuana Twitter handle @stillblazingtho was in the top 10% of all Twitter handles followed .
One recent cross-sectional study asked 742 young adult marijuana users about the number of times they had seen or heard information about advertisements/promotions, coupons or discounts for a dispensary or for buying marijuana in the past 30 days . Over half of those surveyed were exposed to marijuana advertising in the past month: 28% passively observed advertisements; 26% actively sought advertisements. Further, most respondents reported digital media sources for advertisements, and about half observed advertisements via print, television, radio, and on dispensary storefronts. Cross-sectional results also indicated that young adults seeking advertisements were more likely to report the medical use of marijuana and to use marijuana several times per day compared to those who did not actively seek out ads . Other work in this area has examined health claims made about marijuana use on Weedmaps, and demographics of Weedmaps’ followers on social media sites . Results indicated that 61% of retailers in Colorado and 44% of retailers in Washington made health claims about the benefits of using marijuana, including reduced anxiety and treatment for depression, insomnia, and pain/inflammation. The study also showed that most followers of Weedmaps on Twitter and Instagram were male and age 20–29 ; however, about 1 in 6 followers of Weedmaps on Twitter were under age 20. In the only longitudinal study to date to assess exposure to medical marijuana advertising among adolescents , D’Amico and colleagues found that middle school students’ exposure to MM advertising was related to both increased intentions to use marijuana and marijuana use one year later.
This work has been used to inform public policy surrounding advertising for this drug, and is cited in an act to amend Section 26152 of the Business and Professions Code relating to cannabis in California , and has also been used to inform a recent cannabis advertising ordinance for the city of Los Angeles , which both seek to regulate such advertising. Marijuana use during adolescence is associated with numerous issues, including poorer mental health and academic performance, increased delinquency, higher likelihood of abuse or dependence in adulthood, and neurocognitive deficits . It is likely that more states will pass both medical and recreational marijuana legalization laws in coming elections ; therefore, we must begin to look more closely at the longitudinal effects of MM advertising on marijuana use among adolescents so that we can better understand the extent to which youth are exposed to advertising and the effects that this advertising may have on their subsequent marijuana use and related outcomes. Thus, the current study is highly significant as it is the first to directly examine the conjoint longitudinal change in MM advertising and adolescents’ 1) marijuana use, 2) future intentions to use marijuana, 3) positive expectancies about marijuana use, and 4) negative consequences from marijuana use. These associations are examined over a seven-year period using parallel process growth curve models. Furthermore, this analysis will be informative for other states that may want to examine the effects of legislation on outcomes and must do so in the context of a fast-changing marketing landscape.
This study focuses on two cohorts of youth who were in 6th and 7th grade in 2008 and were followed until 2017 . Participants were initially recruited from 16 middle schools across three school districts in Southern California . Responses are protected by a Certificate of Confidentiality from the National Institutes of Health, and procedures were approved by schools and the institution’s internal review board. Schools were selected to obtain a diverse sample and have similar alcohol and other drug use rates at baseline. Schools were matched to their nearest neighbor school based on the squared Euclidean distance measure, estimated using publicly available information on ethnic diversity, approximate size, and standardized test scores . Detailed procedures are reported in the original prevention trial and other trajectory work . Briefly, adolescents completed waves 1 through 5 in middle school during PE class ; follow-up rates ranged from 74% to 90%, excluding new youth that could have come in at a subsequent wave. Adolescents transitioned from 16 middle schools to over 200 high schools and were re-contacted and re-consented to complete annual web-based surveys. At Wave 6, 61% of teens participated in the follow-up survey. We retained 80% of the sample from wave, 91% of the sample from wave, and 89% of the sample from wave. If a participant did not complete a wave of data collection, they were still eligible to complete all subsequent waves. That is,how to cure cannabis they did not “dropout” of the study once they missed a survey wave; rather we fielded the full sample at every wave so that all participants had an opportunity to participate in each survey. Failure to complete a certain wave was not significantly associated with demographics or risk behaviors, such as drinking and marijuana use . The current study focuses on wave 4 through 9 . We began to collect data on exposure to MM advertising at wave 4 because a proposition to legalize marijuana was being discussed in the California Senate in January 2010 and was added to the California ballot in November 2010 . The mean age of the sample at wave 4 was 13. Youth are ethnically and racially diverse , and rates of marijuana use across waves are comparable to national samples . Specifically, in Monitoring the Future, 16.4% of eighth graders reported lifetime marijuana use in 2011 compared with 15.8% in our 8th grade sample. The trajectory sample comes from a sample of youth who were in 6th or 7th grade at wave 1. As noted above, we use waves 4 through 9, and adolescents were in 7th or 8th grade at wave 4. We used parallel process latent growth modeling in a structural equation modeling framework to assess multivariate change over time. The majority of youth completed two or more survey waves . We examined the association between dual trajectories of MM advertising exposure over-time with 1) trajectories for marijuana use, 2) intentions to use marijuana, 3) positive marijuana expectancies, and 4) marijuana negative consequences. The conceptual model is presented in Fig. 1. This framework uses observed scores to estimate latent growth factors, which in turn are used to model an individual’s scores at each time point. In LGM, two growth factors are estimated: one latent factor for the intercept , and another latent growth factor for the slope, or rate of change.
A parallel process LGM is a relatively straightforward approach that simultaneously models multiple growth trajectories. A key feature is the ability to test associations among growth factors such that the growth factors in one process can be related to the growth factors of another. Thus, this model provides a powerful method for investigating change across time in multiple variables. This method was chosen over alternative techniques such as cross lagged models as many of the limitations of the CL model arise from an emphasis on inter-individual differences and not on intra-individual change. Therefore, the CL model, while useful, must be considered with the caveat that it cannot easily incorporate a theory of intra-individual change, and that cross lagged effects are not specific to the type of individual-level change observed over time . Additionally, the CL model is not ideal when the objective is to model the functional form of growth or evaluate intra-individual change. An important advantage of LGMs is that the model allows one to examine inter-individual differences in intra-individual growth in longitudinal studies, with inter-individual differences being captured by the variances of the growth factors. Moreover, this model can assess the functional form over-time. All models were estimated in Mplus v8 with the weighted least squares with mean and variance adjusted estimator , which can handle missing data and provide consistent and unbiased estimates. This estimation method provides model fit indices, which guide in the evaluation of overall model fit. For each model, we report on a combination of model fit indices including both relative and absolute indices. While χ2 is traditionally reported, in large samples, it can be overpowered and detect even small deviations between the observed and model-implied covariance matrix . Thus, we also report Root Mean Square Error of Approximation , Comparative Fit Index , and WRMR to provide a complete picture of model fit. In LGM and its parallel process extension, the model intercept represents the predicted value of the outcome when the predictor is equal to zero. Because assessment waves were not evenly spaced across years , we set this to zero at 3 years after the first assessment. That is, time was centered at the middle of the overall assessment waves . Specifically, between waves 4 and 9, there were 6 total years, treated as follows in the growth models: wave 4 = − 3 years, wave 5 = − 2 years, wave 6=0 year, wave 7 = 1 year, wave 8 = 2 years, wave 9 = 3 years. Thus, for each LGM, the intercept can be interpreted as the average MM advertising exposure, marijuana use, intentions to use, positive expectancies and consequences. The slope represents the rate of change in the aforementioned outcomes. Covariates in each model included age, gender, race/ethnicity, and intervention status. This analytic framework allowed us to test whether changes in MM advertising exposure and marijuana use, intentions to use, positive expectancies, and consequences were associated over time. Within each latent growth process, covariation between intercept and slope was evaluated. Although cross-process covariances were estimated , we limit our discussion to within-process growth factors associations as these are of primary interest. Within each construct, with the exception of consequences, intercept and slope co-varied significantly such that the average for the construct was positively associated with rate of change . That is, for youth with greater marijuana use, intentions to use, and positive expectancies, there was a greater increase over time than for those who reported lower marijuana use, intentions to use, and positive expectancies. Across waves 4–9, there was significant variability in each of the trajectories. Fig. 2 provides data for a random sample of 1000 individuals across MM advertising exposure and use. The variability around intercepts reflects that among individuals, there were varying degrees of average exposure to MM advertising and marijuana use. Note that plots also look similar for intentions to use marijuana, positive marijuana expectancies, and marijuana negative consequences. Moreover, for slopes, in addition to the rate of change for each outcome being significant indicating a nonzero slope, there was also significant variability around the slopes, which highlights that individuals had differential rates of change on each of the outcome measures.This is the first longitudinal study to examine effects of exposure to MM advertising across seven years on adolescents’ marijuana use, intentions to use marijuana, positive expectancies about marijuana, and negative consequences from using marijuana.