Methods: The sample includes 423 youth who completed the first two CalYOUTH interviews (age 17 and age 19) and who were still in care at age 19. Online surveys were administered to the 265 child welfare workers who supervised the care of these 423 youth around the time youth participated in the wave 2 interview. The two dependent variables (high school/GED completion and days homeless since last interview) were both collected from the wave 2 interviews. The main county-level predictors (caseworkers’ perceptions of service availability and quality of cross-system collaboration) were drawn from the caseworker surveys. Control variables included youths’ demographic characteristics, which were collected during the wave 1 interviews, and characteristics of youth’s foster care and maltreatment histories, which were obtained from California child welfare administrative data. Logistic regression was used to examine associations between high school/GED completion and county-level secondary education support availability and quality of collaboration with education system actors (the sample for this analysis excluded 51 youth with a high school diploma/GED at wave 1, n=372). Poisson regression was used to explore whether housing option availability and quality of collaboration with local housing services predicted the total number of days youth were homeless since wave 1 (n=423). Multiple imputation was used to address missing data. Both regression models controlled for a wide range of youth demographic characteristics, baseline outcomes, and foster care and maltreatment history characteristics.
Results: In terms of education, 74.3% of youth had earned a high school diploma or obtained a GED between age 17 and 19. In the logistic regression analyses, higher levels of perceived collaboration with education systems predicted increased odds of high school/GED completion between ages 17 and 19 (OR=1.69, p<.01). In terms of homelessness, youth experienced an average of 4.6 days of homelessness between waves 1 and 2. Poisson regression results found that caseworkers’ perceptions of increased housing availability deceased the estimated number of days youth were homeless (RRR=0.46, p<.05).
Conclusions and Implications: This study found that county-level service availability and cross-system collaboration increased the expected odds of completing high school/GED and decreased the expected number of days youth were homeless. Findings underscore the roles that differences between counties in service availability and cross-system collaboration play in promoting positive outcomes for foster youth.