Adolescent romantic relationships and substance abuse: A multi-level analysis


Lewis, Lisa; McConnell; Aalsma, Matthew; C; & Tong, Yan (2008). Adolescent romantic relationships and substance abuse: A multi-level analysis. 2008 Add Health Users Conference. Bethesda, MD.


Objective: The purpose of this analysis is to examine the longitudinal association between supportive romantic relationships (i.e., reciprocated romantic partners at Wave I) and substance abuse in adolescents. Methods: Participants were 534 individuals (267 dyads) from Waves I, II and III of Add Health. Included within the 267 dyads were 80 dyads who stayed in the same relationship in Waves I and II. The Actor Partner Interaction Model (APIM) was used to predict future substance use. Model fit was based on change in AIC. Results: Predictors include Wave I individual and romantic partner substance use (alcohol, tobacco, marijuana), gender and age. Additional derived measures from Wave I, based on APIM, included dissimilarity (Male – Female behavior; higher absolute scores indicate greater dissimilarity) and dyad interaction (Male x Female behavior; higher scores indicate greater dyad behavior). As expected, individual effects significantly predicted Wave III alcohol use (beta =.53), tobacco use (beta =.71) and marijuana use (beta =.39). Dyad effects were significant with alcohol use (beta =.27) and partner effects were noted with tobacco use (beta =.10), indicating that partner behavior at Wave I predicted Wave III behavior. CONCLUSIONS: In this study, we noted the importance of dyads in the prediction of alcohol and tobacco use but not marijuana use. This may suggest that the decision to smoke may be more an individual decision-making process as opposed to the choice to use alcohol which may be more impacted by social process within the adolescent.


Reference Type

Conference proceeding

Book Title

2008 Add Health Users Conference


Lewis, Lisa
Aalsma, Matthew
Tong, Yan

Year Published


City of Publication

Bethesda, MD

Reference ID