Socioeconomic differences have been reported in adolescent substance use, with lower family socioeconomic status (SES) often linked to higher prevalence of regular and high-risk consumption. This study used data from the 2019 European School Survey Project on Alcohol and Drugs (ESPAD) to investigate whether SES inequalities in adolescent cannabis use are mediated by two proximal factors: perceived availability of cannabis and perceived risk of harm.
The analysis focused on two outcomes: current regular use (defined in the survey as ten or more uses in the past month) and past-year high-risk/problematic use. SES was operationalized using parental education (maternal and paternal levels). Models adjusted for age, sex, and country to account for demographic and contextual differences across the 35 European countries sampled.
Data source and population: The study is a cross-sectional, secondary analysis of the ESPAD 2019 survey, which collected self-reported information from students aged 15–17 years across 35 European countries. The authors used responses on lifetime and recent cannabis use, parental education, perceived availability of cannabis, and perceived risk of harm.
Outcome definitions: Two binary outcomes were modelled: (1) regular cannabis use, defined as reporting ten or more uses in the past month, and (2) past-year high-risk/problematic cannabis use (as defined in the ESPAD measures used by the authors).
Exposure and covariates: Parental education (maternal and paternal) was the main SES exposure. Analyses controlled for age, sex, and country. The study then introduced two potential mediators/moderators separately and jointly: perceived cannabis availability and perceived risk of harm, to evaluate their role in explaining SES differences.
Statistical approach: The authors modelled odds of the two cannabis outcomes as a function of parental education with sequential adjustment: first for demographic covariates (age, sex, country), then adding perceived availability and perceived risk of harm to assess mediation or attenuation of SES associations. Exact modelling details (e.g., regression type, handling of missing data, clustering by country/school) were reported in the full manuscript; summary results are presented in the abstract.
Association between parental education and cannabis use: Lower paternal education was associated with both higher odds of regular and high-risk cannabis use. Reported effect sizes for lower paternal education were an odds ratio (OR) of 1.55 (95% CI 1.20–2.0) for regular use and OR 1.77 (95% CI 1.49–2.11) for past-year high-risk use. A similar pattern of association was observed with maternal education, indicating that lower parental educational attainment relates to elevated cannabis use among adolescents.
Role of perceived availability: When perceived cannabis availability was introduced into the models, it increased the likelihood of higher use among adolescents with lower SES. The authors note this effect is related to differences in perceived availability across socioeconomic groups observed in the data.
Role of risk perception: Introducing perceived risk of harm substantially attenuated SES associations. After adjustment for risk perception, SES factors lost statistical significance for the regular-use outcome and reduced the odds ratios for high-risk use by approximately 40–50% relative to models without risk perception. These results suggest that differences in perceived harm explain a large portion of the SES gap, particularly for regular cannabis use.
Interpretation of mediation: The pattern of results indicates that SES gradients in cannabis consumption among adolescents are partly explained by socioeconomic differences in risk perception, with perceived availability also contributing but in a different pattern. The attenuation of SES effects after accounting for risk perception implies that lower perceived risk among adolescents from lower-education families helps explain their higher reported consumption.
Using a large, multi-country adolescent survey, the authors found that lower parental education is associated with higher odds of both regular and high-risk cannabis use among European adolescents. Differences in perceived risk of harm accounted for a substantial portion of these socioeconomic inequalities, particularly eliminating the SES association for regular use and markedly reducing it for high-risk use. Perceived availability also played a role, increasing likelihood of use among lower-SES adolescents in the adjusted models.
The authors highlight an important policy consideration: interventions or public messages that alter perceptions of cannabis risk could have differential effects across socioeconomic groups and might inadvertently widen consumption inequalities if they change risk perception more among some groups than others.
The analysis was a secondary analysis of anonymous ESPAD data. The authors reported that no new ethical approval or informed consent was required for this secondary analysis; ESPAD data use consent is referenced (ESPAD data use consent N. 2022_03_005). The authors declared no competing interests.
All data produced in the study are available from the ESPAD group upon reasonable request, per the authors' statement. Because this summary is based on the preprint abstract and metadata, readers interested in full methodological detail, sensitivity analyses, or country-level results should consult the full preprint and the ESPAD data and supplementary materials provided by the authors.