Parental knowledge/monitoring and adolescent substance use: Exploiting pandemic-induced, within-family fluctuations to disentangle correlation and causation

ABSTRACT

Purpose: Parental knowledge/monitoring of youth's whereabouts/activities is a well-established risk factor for adolescent substance use. However, findings linking parental knowledge/monitoring to adolescent substance use are almost entirely correlational the observed association may be due to pre-existing differences between families with low vs. high monitoring, rather than a causal relation. The current study aims to disentangle correlation and causation by exploiting pandemic-induced, within-family fluctuations in parental monitoring over a period of several months. Methods: During the COVID-19 pandemic, 7,856 youth at 21 sites across the United States (https//abcdstudy.org/) completed three surveys spaced ~40 days apart. Youth (ages 10.8-14.8 years) reported on past month substance use, level of parental knowledge/monitoring;and whether they were separated from a close family member due to COVID-19 (presumably reducing monitoring). To examine the associations among these measures, we used an econometric technique called first differencing, which involves transforming longitudinal data and fitting regressions to the timeseries of within-subject changes. The advantage of this approach is that it rules out all time-invariant factors (e.g., shared genetics, SES, youth temperament) as explanations for covariation between changes in the constructs of interest. To rule out other potential confounding variables, we adjusted for several time-varying covariates. Results: Month-over-month, within-family changes in parental knowledge/monitoring were associated with month-over-month changes in youth substance use (p < 0.01), as were month-over-month changes in separation status (p < 0.05). Youth were 90% more likely to initiate substance use in months with marked decreases in monitoring and 10% more likely to desist use in months with marked increases in monitoring. Similarly, youth were more 38% more likely to initiate substance use in months when they experienced a COVID-19-induced separation from a close family member and 64% more likely to desist in substance use in months when they reunited with a close family member. Conclusions: In a nationwide sample, within-family changes in parental knowledge/monitoring over a period of several months were robustly associated with adolescents' transitions into and out of using substances. This association remained strong after ruling out all time-invariant potential confounders of the association and adjusting for several time-varying potential confounders (e.g., youth stress). Findings support a causal interpretation of the association between parental knowledge/monitoring and adolescent substance use.