College of Liberal Arts & Social Sciences Theses and Dissertations

Graduation Date


Document Type


College/Department Conferring Degree



stress; protective factors; symptoms; moderator; low-income urban youth


Research examining the etiology of psychopathology has largely focused on stress as a predictor of psychological distress. Particular attention has been paid to exploring the mechanisms that explain the influence of stress on psychopathology. However, few studies have examined these constructs in the context of the low-income urban environment (Grant et al., 2005). Recent literature suggests that psychopathology may manifest uniquely in the most stressful situations, found in the low income urban environment. The current research examines the differential effects that family support demonstrates on internalizing and externalizing symptoms. Further, the study demonstrates the strengths of using structural equation modeling (SEM) in examining moderating effects in developmental psychopathology research, applying techniques recommended by Holmbeck to Structural Regression (1997). Further, specific situations are detailed that demonstrate how apparent "Net Suppressor" effects may be best explained through specification error, in some cases. Finally, the current research demonstrates how mixed methods may be integrated into advanced statistical techniques, such as SEM. Methods The current same consists of 392 urban youth, ranging in grade from 6th to 9th (64% female) and of predominantly African American and Latino ethnicity. Symptoms were assessed using multiple reporters with the Youth Self Report (YSR; youth/child report). Child Behavioral Checklist (CBLC; parent report), and the Children's Depression Inventory (CDI; youth report; Achenbach, 1991a; Kovaks, INSERT YEAR). Qualitative interview and coding procedures were used to collected endorsement of family support and youth perceived severity of stress (Grant et al., 2003). Empirically supported survey instruments were also used measure exposure to violence (Richters & Martinez) and Daily Hassles/ Major Life Events (Allison et al.,2003). Results Results indicate demonstrated a well-fitting model of stress, internalizing, and externalizing psychopathology that utilized multiple indicators of each construct. Models using multiple indicators demonstrate in the literature to reduce the effects of measurement error and eliminate the need for product terms (Holmbeck, 1997). Moderator analyses demonstrate significant moderating effects for youth endorsed family support, although in an unexpected manner. Youth that did not endorse family support demonstrated a stronger positive (and significant) association between stress and externalizing symptoms than youth that endorsed family support, consistent with extant literature. In contrast, youth that endorsed that family support demonstrated a significantly stronger (and positive) association between stress and internalizing symptoms, compared to their non-endorsing counter-parts. These findings may support extant literature that suggests that low-income urban youth may be more likely to externalize (Grant et al, 2005), but this effect may be influenced by family interactions, leading to traditional internalizing symptom expression. Implications for interventions are discussed.