Methods: 288 foster youth (mean age=17.31, SD=.61) were assessed and randomized to intervention condition (146 control, 142 treatment), and were re-assessed at 12 months (post-intervention, n=243) and at 24 months (n=226, with 78% retention overall). The baseline sample is 53% female, 19% Hispanic/Latino, and 46% White, 16% Black or African American, 19% multi-racial, and 19% other. At baseline, 59% received high school special education services, including those with developmental disability (22% of sample). 44% were participating in ILP services, 45% were in a non-relative foster home, 23% in a relative/kin placement, and 32% were in more restricted settings. Outcomes were analyzed with a two (treatment group) by three (time point) repeated measures design, using the GLMM growth curve model to determine the main effects of group (averaged across time) and time (averaged across groups), and the overall group-by-time interaction. Measures include the Arc Self-Determination Scale (ARC), the Career Decision Self-Efficacy (CDSE) Scale, the My Life Activity Checklist (MLAC) and My Life Self-Efficacy Scale (MLSES), and coded interview data assessing youth self-determination skills development.
Results: The overall group-by-time effect for the ARC was not quite statistically significant (p=0.053, one-tailed), though the groups did significantly diverge between Times 1 and 2 (p=.030) and between Times 1 and 3 (p=.025), with a small effect size (<.30). For the CDSE, there were not significant group-by-time effects overall, but the contrast between T1 and T3 was significant (p=.03). There were group-by-time differences for the MLAC overall (p=.01), where the treatment group increases more over time, especially between T1 and T2 (p=.026), with a trend-level effect from T1-T3 (p=.095). There were also treatment group gains across a range of MLAC and CDSE subdomains (career, education, etc.). There was a trend-level (p=.08) T1-T3 treatment group gain on the MLSES. Lastly, there were significant findings on coded variables from the self-determination skills interview, including group-by-time differences on youth identification of accomplishments, steps to achieve goals, and strategies for stress management.
Conclusion/Implications: This study informs child welfare policy and practice by investigating whether a theory-based intervention model helps youth learn and apply skills for increased goal-setting and self-determined transition planning. Overall, there is a consistent pattern of hypothesized effects on proximal outcomes related to self-determination and self-efficacy, which are the focus of the My Life model. However, we expected these measures to be attenuated by risk factors associated with being in foster care, as described in a separate but related analysis of moderators of intervention effectiveness on the outcome measures reported here.