- The bulk of resilience research has focused on developmental outcomes in a single domain of functioning (e.g., academic achievement), overlooking the complex interplay of positive and negative developmental outcomes that may co-occur within an individual (Luthar, Cicchetti, & Becker, 2000). Resilience may strengthen over time and across domains (e.g., psychological well-being, physical health), thus a corresponding variation in functioning across domains of development is also likely (Wright, Fopma-Loy, & Fischer, 2005). A multidimensional approach to resilience is necessary to account for this variation in functioning. Therefore, the current research trend of examining a single outcome and then ascribing a unidimensional 'resilient vs. not resilient' label without specifying a domain of resilience may be masking the co-occurrence of positive and negative outcomes of at-risk individuals. To address this gap, this dissertation presents two studies exploring resilience as a multidimensional construct. Both studies examined data from 323 students who participated in the Chicago cluster randomized trial of the Positive Action (PA) program (Ji, DuBois, Flay, & Brechling, 2008). The first study identified distinct subcategories of resilience profiles in an at-risk group of adolescents. Latent class analysis results indicate that, in contrast to the unidimensional--i.e., resilient or not resilient--construct of resilience, there are multiple profiles of resilience among at-risk students. Four distinct groups of students with markedly different resilience patterns were identified: Consistently Resilient (35.5%)--high on positive and low on negative outcomes, Consistently Non-resilient (18%)--low on positive and high on negative outcomes, Socio-emotional Resilience with Academic Issues (29.6%)--high on Social-emotional and Character Development (SECD) and positive health behaviors, low/moderate on negative outcomes, but reporting very poor academic performance, and Academically Resilient with Mental Health Issues (16.9%)--high on positive outcomes (SECD was moderate), low/moderate on violence and substance use, but very high on anxiety and depression. The second study used multinomial logistic regression to examine individual-, family-, and community-level risk and protective factors at grade 3 that predict distinct resilience patterns among children at grade 5. Analyses predicted resilient group membership, indicating students with high levels of empathy, parental and teacher reports of responsibility, parental encouragement, and students’ feelings of safety at grade 3 were more likely to be resilient across multiple domains at grade 5. Together, these studies empirically support a more integrative, comprehensive, and multidimensional conceptualization of resilience and its predictors that is better-suited to identifying hidden strengths and vulnerabilities in children and adolescents.