Paradoxical Multilevel Relations Between Internalizing Symptoms and Solitary Drinking in Emerging Adults

OBJECTIVE: Theoretical models suggest that internalizing symptoms predispose individuals toward solitary drinking. However, data have not explicated whether longitudinal relations occur across or within individuals over time, which would have important yet distinct clinical implications. If relations exist across individuals, targeted prevention in individuals with higher internalizing symptoms may be most clinically effective. However, if relations exist within individuals, monitoring within-person deviations in internalizing symptoms and coupling increases in internalizing with adaptive interventions may prove most effective. METHOD: = 448; 43.4% female; ages 21-25) with a history of past-month binge drinking at baseline reported on depressive, anxiety, stress symptoms (i.e., "internalizing symptoms") and solitary drinking frequency every 6 months over 2 years. Autoregressive multilevel models tested whether (a) individuals with higher internalizing symptoms reported more frequent solitary drinking across the span of 2 years, and (b) within-person deviations in internalizing symptoms covaried with (and prospectively predicted) within-person deviations in solitary drinking. RESULTS: Internalizing symptoms predicted more frequent solitary drinking at the between-person level. Within-person deviations in internalizing symptoms were positively contemporaneously correlated with deviations in solitary drinking. However, within-person deviations in internalizing symptoms at a given time point predicted within-person decreases in solitary drinking 6 months later. CONCLUSIONS: Individuals with higher internalizing symptoms reported more frequent solitary drinking, yet within-person deviations in internalizing symptoms prospectively predicted decreases in solitary drinking. Findings suggest a complex, level-specific relation between internalizing symptoms and solitary drinking, which may have implications for preventive interventions.