| dc.description.abstract | For decades, students with emotional or behavioral disabilities (EBD) have had consistently poor high school graduation rates and postschool outcomes in terms of employment, postsecondary education enrollment, and involvement with the criminal justice system (Wagner, Newman, Cameto, & Levine, 2005). In addition, compared to students who are White and/or whose families have high incomes, outcomes are generally worse for youth with disabilities who are African-American or Hispanic and/or whose families have low incomes (Newman, Wagner, Cameto, & Knokey, 2009). Using school and postsecondary data on students with EBD (n = 450) from the National Longitudinal Transition Study-2 (NLTS2), this study used cross tabulations and mean comparisons to examine significant demographic differences among students with EBD in engagement (i.e., family and student engagement) during their school years and in outcomes after high school. In addition, logistic regression analyses were used to determine the predictive value of four categories of predictor variables on key postschool outcomes. Predictor variables included (a) demographic variables (i.e., race/ethnicity, family income, and gender); (b) negative student engagement (e.g., being bullied, being suspended); (c) family engagement at home and in school; and (d) positive student engagement (e.g., student connectedness to school, grades, involvement in extracurricular activities). Four criterion variables were examined: high school graduation, postsecondary education, fulltime employment, and involvement with the criminal justice system (i.e., arrest). High school graduation status was included in the models for postsecondary education, fulltime employment, and arrest. Results showed that positive student engagement was associated with higher odds of graduating from high school and enrolling in postsecondary education, and negative student engagement was associated with higher odds of arrest. High school graduation status was found to be a significant predictor of postsecondary enrollment and arrest, whereas family engagement predictors contributed little to the models in most instances. Results differed depending on the criterion variable (i.e., high school graduation, postsecondary enrollment, fulltime employment, and arrest): For example, student engagement had little statistical value for predicting fulltime employment. In addition, some significant differences in outcomes based on demographic variables (i.e., race/ethnicity, family income, and gender) were found to be reduced or negated when student or family engagement variables were included in logistic models; other differences persisted after the inclusion of the engagement variables. Limitations, directions for additional research, and practical implications are discussed. | |
