Citation

Grigsby TJ, Forster M, Baezconde-Garbanati L, Soto DW, Unger JB. Addict. Behav. 2014; 39(3): 644-651.

Affiliation

University of Southern California, Keck School of Medicine, Department of Preventive Medicine, Institute for Health Promotion and Disease Prevention, United States. Electronic address: unger@usc.edu.

Copyright

(Copyright © 2014, Elsevier Publishing)

DOI

10.1016/j.addbeh.2013.11.021

PMID

24345310

Abstract

The present study compares statistical models for three conceptualizations of drug use in 11th grade (past 30day ever/never use, past 30day frequency of drug use and past 30day drug use consequences) with externalizing and internalizing problems in emerging adulthood when controlling for age, academic achievement and socioeconomic status in a Hispanic sample. Multivariate logistic regression models for the different drug use variables were compared when modeling weapon carrying, arrest, multiple lifetime sex partners, drug/alcohol use before sex and condom use in emerging adulthood. A multivariate linear regression model was used to model depression in emerging adulthood as a function of drug use measurement controlling for other covariates and depression in adolescence. Our findings suggest that any conceptualization of drug use will produce equitable results and model fit statistics when examining externalizing problems. However, when investigating internalizing problems, such as depression, lower frequency drug use-and not high frequency-was more strongly associated with depression whereas experiencing high levels of drug use consequences-and not low levels of consequences-was associated with depression in emerging adulthood despite similar model fit values. Variation between drug use and the experience of drug use consequences may lead to misspecification of "at-risk" subgroups of drug users. Implications and future directions are discussed.

Language: en