Abstract
Aims
To test whether the total consumption model and its extension, the theory of collectivity, apply to adolescent cannabis use. We estimated: (1) whether the distribution of cannabis use frequency was stable over time; (2) whether changes in average cannabis use were mirrored across consumption quantiles; and (3) whether higher average use corresponds with a higher prevalence of high-frequency users.
Design
Repeated cross-sectional study based on annual surveys. We analyzed trends in cannabis use frequency across consumption percentiles using Lorenz curves and Gini coefficients to assess distributional stability. We used ordinary least squares (OLS) regression to estimate the elasticities between overall mean frequency of use and mean frequency of use for five quantiles ranging from P25 to P95. The association between average use and the prevalence of high-frequency users was examined graphically and parametrically.
Setting
Sweden.
Participants
Ninth-grade students (aged 15–16) surveyed annually between 1990 and 2023 (excluding 2013), with a total of 180 059 respondents, and 2nd-year high school students (aged 17–18) surveyed annually between 2004 and 2023 (excluding 2013 and 2020), totaling 80 925 respondents. Our analyses were limited to individuals who reported cannabis use, resulting in analytical samples of 10 139 9th-grade students and 11 160 2nd-year high school students.
Measurement
Frequency of cannabis use was measured by a question on how many occasions the respondent has used hashish or marijuana. The seven response alternatives ranged from 0 to 50 times or more.
Findings
Mean frequency of cannabis use fluctuated statistically significantly across survey years among 9th-grade students (F-test = 7.647, P < 0.001) and marginally among 2nd-year students (F-test = 1.550, P = 0.068). Notwithstanding this variation in mean use, the distribution of frequency of cannabis use remained stable: Lorenz curves were consistent across years, and Gini coefficients showed no significant changes. Mean frequency in all five quantiles (25th–95th) were positively and statistically significantly associated with overall mean frequency, suggesting synchronized changes across user groups; e.g. the elasticity for P50 was estimated at 0.914 (P < 0.001) in 9th grade. Increases in mean frequency use were associated with a higher prevalence of high-frequency users. Thus, an increase in average frequency of cannabis use by 1 percentage point was associated with a 1.794 (standard error = 0.065, P < 0.001) percentage point increase in high-frequency users in 9th grade.
Conclusions
Adolescent cannabis use in Sweden appears to conform to key predictions of the total consumption model and its extension, the theory of collectivity.