Bridging Disciplinary Boundaries (January 11 - 14, 2007)

Understanding the trajectories of behavior problems across adolescence will help identify high risk subgroups and factors that might become targets for intervention. This paper identifies discrete subgroups of behavioral trajectories and then examines the predictors of these subgroups.

Methods. Approximately 400 urban and reservation American Indian adolescents were interviewed yearly from 2001 to 2004. The Youth Self Report (YSR) constitutes the base measure for the behavior change trajectories, with personal and environmental variables serving as independent variables. Analyses were based on group based modeling via Proc Traj, a SAS macro (Jones, 2001), and multivariate logit modeling.

Results. A model containing five trajectory groups provided the best fit. Youths who started with a YSR score less than the clinical cut-off point had two different trajectories: low stable (n=142) or low improving (n=175). Youths who began with scores over the clinical cut-off point fell into 3 different trajectory groups: very high chronic (n=5), high chronic (n=33), and high improving (n=30). The high improving trajectory was curvilinear, with youths improving over 4 years until they scored close to the best group.

We examined differences between the two diverging high problem groups (high chronic and high improving), whose initial average YSR scores were similar. A multinomial regression model predicted the odds of being in the high improving group versus the high chronic group were almost 6 times higher for reservation youths. A one point increase in family satisfaction was associated with a 14 % increase in the odds of being in the high improving group, and each point higher on the school problems scale decreased the odds of being in the high improving group by about 16 %.

Youths in the high improving group made several psychological and environmental changes over the four years: substance use, symptoms of alcohol and drug dependence/abuse and depression symptoms dropped, family satisfaction increased, fewer parents and family members had mental health or addictions problems; fewer peers used substances, and neighborhood problems and stressors decreased.

Although the total number of services that the youths received (after controlling for need) did not differ in 2001, the high improving group received over 3 times as many school services.

Implications. We discuss reservation community changes (improved financial opportunities, social programs, and addictions interventions) that might have contributed to the changes in the largely reservation based high improving group. We also discuss reservation reported drops in arrest rates, addicted newborns, and increases in high school graduation over the same four year period. Our study has import both for what it demonstrates about trajectories and also for the directions it indicates for influencing that behavior change in youths.

Jones, B., Nagin, D., & Roeder, K. (2001). A SAS procedure based on mixture models for estimating developmental trajectories. Sociological Methods and Research, 29, 3, 374-393.

Funding: NIMH K02 MH01797-01A1, and NIDA R24DA13572-0 and R01 DA13227-01