Proximity to dispensaries in neighborhoods has been shown to have a positive association with more frequent marijuana use among adults , but the effect of proximity to marijuana outlets to schools on adolescent marijuana use is unknown. Research Question #4 was: “Is the effect of dispensary bans on student marijuana use dependent on the distance of the nearest dispensary from a student’s high school? The hypotheses associated with RQ4 propose that H4.1) dispensary bans are associated with longer distances between dispensaries and schools; H4.2) the proximity of dispensaries is positively associated with students’ likelihood of using marijuana; and H4.3) the relationship between city dispensary bans and high school students’ marijuana use is mediated by dispensary bans effectiveness at keeping dispensaries a greater minimum distance from schools than city policies that allow dispensaries. Table 7.12 presents the results of the mediation analysis assessing the mediating effect of the distance to the nearest dispensary within LA County on lifetime marijuana use. As reported earlier, the relationship between dispensary bans and lifetime marijuana use was negative and non-significant . Whether a city had a dispensary ban was regressed on the distance in miles from the participant’s school to the nearest unlicensed dispensary using an HLM model controlling for the covariates and with a random intercept for city. The association between dispensary bans and the distance between the students’ high school and the nearest unlicensed dispensary was positive and statistically significant ,plant racks which supports H4.1 for lifetime marijuana use, i.e., my theory that dispensary bans would be associated with a greater average distance from schools than in cities that allow dispensaries.
The distance in miles from the school to the nearest unlicensed dispensary was then regressed on the lifetime marijuana use, which revealed a statistically significant negative association between distance and reports of lifetime use, meaning that students in schools located further from unlicensed dispensaries were less like likely to report lifetime marijuana use , confirming H4.2., that shorter distances between the participants’ schools and the nearest dispensary were associated with higher prevalence of lifetime marijuana use. Finally, the regression analysis of dispensary bans and lifetime marijuana use was repeated including the distance in miles from the school to the nearest unlicensed dispensary. Adding the distance measure increased the effect of dispensary bans and in the expected direction, i.e., it was hypothesized that dispensary bans would be negatively associated with marijuana use and adding the variable quantifying the distance to the nearest dispensary into the regression strengthened the negative association between dispensary bans and lifetime marijuana use, but the overall relationship still fell just short of statistical significance . This result indicates that indirect mediation was occurring, where the effectiveness of dispensary bans was partially dependent on how far away they keep unlicensed dispensaries from high schools compared to city policies that allow dispensaries. The above steps were repeated for the recent marijuana use outcome, with similar results. The focal relationship between dispensary bans was negative and not statistically significant, as reported earlier. Path a, the association between dispensary bans and the distance to the nearest dispensary was the same as described above for recent use. Path b, the effect of the distance between the nearest unlicensed dispensary and recent marijuana use was negative and statistically significant , indicating the further away unlicensed dispensaries were located from the participants’ schools, the less likely they were to report recent marijuana use .
Path c’, the relationship between dispensary bans and recent marijuana use was strengthened substantially and in the expected direction but fell short of statistical significance, which indicated partial mediation. Hypothesis 4.3 was therefore partially supported. In Table 7.14 I present the results of a sensitivity analysis conducted a series of multilevel logistic regression models on increments of one mile, continuing to increments of a quarter mile for recent marijuana use, where the distance to the nearest MMD was not statistically significant at one mile. For each measure of marijuana use I started with a distance of over five miles and worked inward in increments of one mile, until I reached the point at which the distance to the nearest MMD had a statistically significant association with student marijuana use. For lifetime marijuana use, this occurred at a distance of one mile. For recent marijuana use, the distance from the school to the nearest MMD was not statistically significant at one mile, so I continued using smaller distances in quarter mile increments until I reached the point where the distance between the school and the nearest unauthorized MMD became statistically significant, which occurred at a distance of between ½ mile and ¾ mile of the nearest unauthorized MMD and the participant’s school. This analysis also revealed that the distance from the school to the nearest MMD became significantly negatively associated with student marijuana use at longer distances. The association of distance becomes protective against lifetime marijuana use at a distance of over 5 miles from the participants’ school and for recent marijuana use this occurs between 3 and 4 miles. As the distance used was the distance from school to only the nearest MMD from the participant’s school, these distances were not cumulative.
The ranges of distances were treated as bands, where if the nearest dispensary was located within a certain range of distance from the participants’ school distance from their school the value for that range of distance was coded as ‘1’ and all other rangers of distance were coded as ‘0’. For the distances within one mile I chose increments of a quarter mile because it is hard to visualize distance by feet at longer distances. I created variables using the distance to the nearest MMD from each student’s school to indicate whether the nearest MMD to the student was located within 1 mile , between three quarters of a mile and one mile, and between three quarters of mile and a half mile . There were too few MMDs located within a half mile from schools to produce statistically significant and reliable results without including all of them instead of only the nearest , so I stopped at a distance of a half mile from the school. Students that did not have MMDs located within any of these distances had a zero value for each of these variables. As these were binary variables I present odds ratios in Table 7.14. The sensitivity analysis indicates that there was no “safe distance” within a mile that unauthorized MMDs could be located near schools without having an association with significantly higher rates of lifetime marijuana use among students. The distance between the school and the nearest unauthorized MMD was significantly associated with greater prevalence of recent marijuana use at a distance of a ½ mile to a ¾ mile. The association between the distance between participants’ schools and the nearest unlicensed dispensaries was remarkably consistent for lifetime marijuana use with the exception of a non-statistically-significant result for distance between ½ and ¾ mile and lifetime use. Overall, there is a clear relationship with the distance between participants’ schools and the nearest unlicensed MMD,plant growing trays where shorter distances were associated with significantly greater odds of marijuana use. The association with distance to the nearest unlicensed dispensary then decreased in size and lost statistical significance at longer distances until eventually becoming associated with lower odds of marijuana use as distances increased. The distances at which this research showed statistically significant associations with student marijuana use are much further away from schools than the state requirement of 600 feet, but it is important to note the same associations did not apply for licensed dispensaries. In contrast to the associations with distance from schools found for unlicensed dispensaries, there was not a consistent association with the distance that licensed dispensaries were located from participants’ schools and their odds of reporting lifetime or recent marijuana use . It is therefore possible that smaller distances such as 600 feet are sufficient to prevent licensed outlets from being associated with greater prevalence of marijuana use among students, as licensed marijuana outlets seem to have less of an effect on student use than unlicensed outlets. As presented in Table 7.15, the associations between the distance from the participants’ schools to the nearest unlicensed MMDs were largely inconsistent or were not statistically significant.
For lifetime marijuana use the associations were consistently negative for distances below a mile, meaning that shorter distances to licensed MMDs were associated with lower odds of lifetime marijuana use among students, but these associations were not statistically significant below a mile and were inconsistent at distances greater than a mile. The inconsistent results for distances greater than a mile may indicate confounding effects from city borders, indicating that measurements of distance to authorized outlets greater than mile should be interpreted with caution and may be of limited utility in studies of the effects of distances between schools and dispensaries on high school students’ marijuana use behavior. Similarly, the association between the distance from participants’ schools to the nearest licensed MMDs and recent marijuana use was significant and protective at one mile, but was not statistically significant below one mile and inconsistent with any discernable pattern at distances greater than one mile. Rather than concluding that the presence of licensed MMDs is protective against marijuana use among high school students at distances of one mile, I think it is safer to conclude that there is not a clear relationship between the distance to licensed MMDs and student marijuana use and that any future studies of the effect of this distance should be bound by city borders to avoid any confounding effects of authorized dispensaries being found only within cities that allow them, while unlicensed dispensaries were found both in cities that allowed MMDs and cities with bans. That there was not a clear pattern of association with the distance from schools to the nearest licensed MMD when the association was clear for unlicensed dispensaries is surprising. This finding has important implications for marijuana regulation at a city level, as it suggests that any associations between the proximity of MMDs to areas young people frequent and their marijuana use behavior may be driven by unlicensed outlets. It also reflects a need for further study of why the effects of unlicensed and licensed MMDs should be so noticeably different. It is evident from the results presented above that the continuous distance between schools and unlicensed dispensaries is a mediator of effectiveness for dispensary bans, while there appears to be little association between student marijuana use and the distance to the nearest licensed dispensary. Although I found in Chapter 5 that dispensary bans were not effective on their own, they were associated with a longer distance between participants’ schools and the nearest unlicensed dispensary, which was in turn strongly correlated with a lower prevalence of student marijuana use. It is therefore important to consider what is actually driving adolescent marijuana use and how best to prevent it using city policy. This research suggests that it may need not necessarily be a ban. Although dispensary bans were associated with longer distances to the nearest unlicensed marijuana outlet, strict enforcement of distance requirements and closing down unauthorized outlets while allowing some dispensaries to operate far from sensitive areas could possibly achieve the same aim. The cities in LA County that allow dispensaries have use local ordinances that specify the conditions under which dispensaries can operate in the city. All six of the cities that allowed dispensaries in September of 2016 required them to be at least 500 feet from schools. Current State regulations require a minimum distance of 600 feet, but at the time of data collection, the distances dispensaries were required to be kept from schools in these six cities ranged between 500 – 1,000 feet . Distances specified in dispensary ordinances may be based on somewhat arbitrary criteria, as no empirical research has established what a “safe” distance is. I suspected the presence and number of dispensaries within a larger radius could still be influential and therefore tested the mediating role of the number of dispensaries located 2,000 feet of the study participants’ schools.