Parent alcohol information was derived from reports by the parent (s) themselves or a first-degree relative and linked to offspring via a unique family identifier. At least one parent of a subject interviewed in COGA’s prospective wave would have been assessed for AUD during an earlier phase of the COGA study. Dummy variables representing biological mother-only AUD (father unaffected), biological father-only AUD (mother unaffected), and both biological parents affected by AUD were created, using families with no parental AUD as the referent. We also included dummy variables where parent alcohol information was missing, diagnostically subthreshold, or incomplete as covariates in the analysis.
