Reevaluating the Link Between Psychopathology Symptoms and Alcohol and Cannabis Use: An Examination Across Intersectional Race/Ethnicity and Gender Identities.

Abstract

Objective: The internalizing (e.g., depression/anxiety) pathway to substance use is a prominent hypothesis, but its evidence has been critiqued for neglecting correlated dimensions of mental health such as externalizing (e.g., aggressive, rule-breaking) and attention-deficit/hyperactivity problems. Furthermore, most research has been conducted in predominantly White samples, limiting generalizability. We examine associations between substance use and psychopathology symptoms across intersectional race/ethnicity and gender identities.

Method: Participants (N = 1,187; 16.26% Black women, 17.61% Latina women, 15.67% White women, 16.68% Black men, 17.10% Latino men, and 16.68% White men) ages 18-26 completed online surveys. Alcohol and cannabis factors were indicated by six pattern-of-use items (e.g., frequency/quantity) and psychopathology symptoms via the Adult Self-Report. Differential item functioning was tested before making comparisons in univariate and multivariate models.

Results: Differential item functioning was detected; the degree of bias was modeled via partial metric invariance, allowing for latent inferences. At the univariate level, externalizing symptoms predicted alcohol and cannabis across all groups except Black and Latino men, whereas internalizing and attention-deficit/hyperactivity symptoms predicted cannabis use in White women and White men only. In a multivariate model, internalizing symptoms inversely predicted alcohol use for White women and White men, externalizing symptoms predicted alcohol use in all groups except Latino men, and externalizing symptoms predicted cannabis use in all groups except Black women and Latino men. Attention-deficit/hyperactivity showed no significant associations in the multivariate model.

Conclusions: The conceptualized structure of mental health has implications for research on substance use. Intersectionality-minded methods may provide a more generalizable scientific base.