Self-control demands and alcohol-related problems: Within- and between-person associations

Abstract

This study tested a multilevel structural model of associations between two aspects of self-control (effortful control and reactivity), self-control demands, alcohol consumption, and alcohol problems and related risk behaviors using daily diary data from 196 young adults (4,177 person-days). Self-control demands were hypothesized to be positively associated with alcohol consumption and alcohol problems and related risk behaviors both within- and between-persons. At the between-person level, self-control demands were hypothesized to mediate the association between trait self-control and alcohol problems and related risk behaviors. At the within-person level, self-control demands had a direct positive effect on alcohol problems and related risk behaviors, over and above alcohol consumption. However, contrary to expectation, self-control demands were inversely associated with alcohol consumption. In contrast, self-control demands were positively associated with alcohol consumption at the between-person level and partially mediated the positive effects of reactivity on consumption and alcohol problems and related risk behaviors. That is, reactivity was associated with higher perceived self-control demands, which in turn predicted higher rates of consumption and alcohol problems and related risk behaviors. Effortful control was not significantly associated with alcohol consumption or self-control demands. The pattern of self-control demand effects at the within-person level suggest that young adults are less likely to drink when struggling to manage their day-to-day behavior, yet if they do drink they are more susceptible to negative consequences. Trait effects suggest that individual differences in self-control may be associated with alcohol use patterns in part attributable to development of, and response to, structured daily routines. (PsycINFO Database Record

Figure 1.

Multilevel structural model depicting within-and between-person effects. Unstandardized coefficients with standard errors (in parentheses). N = 196 persons, 4177 person days. Gender is a covariate but paths are omitted for clarity. *p < .05, **p < .01, *** p < .001.