Self-reported factors predicted substance use modestly, but failed to…

Predicting substance use behaviors: A multi-data set mega-analysis.

Journal of psychopathology and clinical science • • Relevant
🤖

AI Summary

This mega-analysis asked how well self-reported information can predict alcohol, tobacco, and cannabis use. It combined data from 44,862 participants and assessed 197 harmonized predictors spanning demographic, behavioral, familial, and psychosocial factors, using linear regression and machine-learning methods.

Methods designed for many correlated predictors—random forest and Elastic Net—performed similarly, predicting about 14% of variation at baseline and about 4% of change over time. In exploratory tests, models did not predict outcomes in independent held-out data sets (R² near zero), suggesting the predictive associations were largely specific to the data sets studied. Externalizing behaviors were informative especially for baseline use, while demographic and familial factors became more apparent when predicting change. This abstract-based summary cannot establish causal relationships or reliable prediction across new populations; it does not report cannabis-specific predictive results separately.

💡 Key Findings

1
Across 44,862 participants, models used 197 self-report predictors to estimate alcohol, tobacco, and cannabis use.
High
95%
2
Random forest and Elastic Net showed comparable performance, explaining about 14% of baseline variation and 4% of change over time.
High
95%
3
In exploratory leave-one-data-set-out tests, models did not generalize to independent data sets (R² near zero), indicating context-specific predictive associations.
High
95%
4
Externalizing behaviors were predictive particularly at baseline; demographic and familial factors became more apparent in models of change over time.
High
85%

📄 Original Abstract

Predicting substance use behaviors is a critical public health challenge. Although self-report survey data have been central to understanding substance use, their longitudinal predictive utility has yet to be quantified. We conducted a "mega"-analysis of 44,862 participants, integrating 197 harmonized self-report predictors across demographic, behavioral, familial, and psychosocial domains. Specifically, we employed both linear regression and machine learning approaches to evaluate the extent to which alcohol, tobacco, and cannabis use could be predicted. Models that handled high-dimensional, correlated predictor sets substantially outperformed unregularized linear regression, reflecting the known limitations of fitting unregularized regression to large, collinear predictor sets rather than a general advantage of machine learning. Random forest and regularized regression (Elastic Net) yielded comparable predictive performance (R² ≈ 14% at baseline and ≈4% for longitudinal change). In exploratory leave-one-data-set-out analyses, models did not generalize to independent held-out data sets (R² ≈ 0), indicating that predictive associations were largely context-specific. Externalizing behaviors predicted a significant proportion of substance use variability, particularly at baseline, whereas the predictive power of demographic and familial factors became apparent in models predicting change over time. Combining large-scale harmonization with computational modeling clarifies both the potential and limits of forecasting substance use trajectories. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Explore More Research

Stay informed about the latest cannabis science.

Your stash, decoded.