LAUSR

OBJECTIVE Adolescents are exposed to alcohol marketing through traditional advertising and through newer digital media channels. Cumulative marketing exposure across channels is of concern but has been insufficiently studied. This study explores the measurement of alcohol marketing exposure across channels and whether cumulative recalled exposure is independently associated with underage drinking. METHOD Two hundred two New England adolescents (ages 12-17 years) were recruited from a general pediatrics clinic and completed an online survey. Recall of alcohol marketing across channels (e.g., Internet, magazines) was assessed, along with drinking behavior and relevant covariates (i.e., demographics, parental/peer drinking, smoking status, sensation seeking, Internet use, social media use, television use, and parental Internet monitoring). Confirmatory factor analysis was used to establish a latent construct of alcohol marketing exposure recall. Logistic regression tested associations between alcohol marketing recall and adolescent drinking, with covariates controlled for. RESULTS Adolescents reported recall of alcohol marketing across all marketing channels. Alcohol marketing recall items were significantly correlated, with α = .83. The latent measurement model of alcohol marketing recall provided excellent fit to the data, χ2(17, n = 202) = 27.402, p = .052; root mean square error of approximation (.000-.092) = .055; Tucker-Lewis Index = .960; comparative fit index = .976; standardized root mean square residual = .037). Adjusted cross-sectional logistic regression analyses demonstrated that the latent alcohol marketing recall construct was significantly associated with underage drinking (adjusted odds ratio = 4.08, 95% CI [1.15, 14.46]) when relevant covariates were accounted for. CONCLUSIONS The final measurement model provided support for construct validity of a novel alcohol marketing recall construct assessing cumulative cross-channel marketing exposure. Adolescent recall of alcohol marketing across channels was significantly associated with underage drinking, while associated factors such as peer/parental drinking were accounted for.