A prospective study of alcohol involvement and the dual‐systems model of adolescent risk‐taking during late adolescence and emerging adulthood

Abstract Aims To estimate interactions and unique effects of behavioral approach and behavioral control on alcohol involvement hypothesized by dual‐systems models, during adolescence and emerging adulthood. Design In a longitudinal study, behavioral approach and behavioral control were examined in relation to alcohol involvement, cross‐sectionally and prospectively. Setting and participants A total of 846 general population twins born in Colorado, USA were assessed twice (mean = 17.3 and 22.8 years; female = 51.4%; white = 91.8%). Measurements Behavioral approach was measured by self‐report questionnaires of sensation‐seeking and subjective effects of alcohol. Behavioral control was measured by self‐reported lack of planning and nine executive functioning (EF) tasks. Interviewers administered semi‐structured clinical interviews to assess alcohol use and disorder (AUD). Findings None of 36 interaction effects was statistically significant (β = −0.16–0.14, P > 0.06), suggesting that dual systems are additively related to alcohol involvement. In multiple regression models, behavioral approach and behavioral control explained independent variance in alcohol use quantity (β = 0.09–0.33, P < 0.04) and frequency (β = −0.11–0.29, P < 0.03) at both waves. During adolescence, only subjective effects (β = 0.27–0.28, P < 0.001) explained independent variance in AUD. Moreover, measures of the same construct explained independent variance in alcohol involvement: for behavioral control, lack of planning and EF were associated with alcohol frequency in adolescence (β = −0.11–0.25, P < 0.02) and AUD in emerging adulthood (β = −0.09–0.16, P < 0.03). For the behavioral approach, subjective effects were associated with all measures at both waves (β = 0.20–0.33, P < 0.01) and sensation‐seeking was associated with all measures in emerging adulthood (β = 0.09–0.11, P < 0.04). In prospective models, adolescent alcohol involvement was associated with later lack of planning (β = 0.12–0.18, P < 0.03) and lack of planning in adolescence was associated with later alcohol involvement (β = 0.12–0.14, P < 0.02). Conclusions Both the behavioral approach and behavioral control components of dual‐systems models explain alcohol involvement during adolescence and adulthood, and different measures of the same system assess separate risk processes. The relations between alcohol involvement and the dual systems appear to be bidirectional.