Crossing the Finish Line: Predictors of College Degree Completion By Age 30 Among Foster Care Youth

Background and Purpose: Although nearly 80 percent of older adolescents in foster care aspire to complete college (Courtney et al., 2014; McMillen, 1999), only 2-10 percent earn a two-year or four-year college degree by their mid-20s (Pecora et al., 2006; Courtney et al., 2011). To date, few quantitative studies have evaluated college outcomes for foster youth, and many are limited by sampling issues, investigation of few predictors, and short follow-up times. The current analysis examines a broad range of predictors of college completion up to age 30 among a representative sample of youth in foster care as adolescents.

Methods: Data come from the Midwest Study, which included a representative sample of adolescents in foster care in three Midwestern states at ages 17/18 in 2002. The baseline response rate was 95% and the four subsequent interview waves spaced two years apart each had response rates of over 80%. The analytic sample includes 329 participants who had enrolled in college and who could be observed for at least six years. College enrollment and completion data were obtained from the National Student Clearinghouse and supplemented by self-report to address underreporting in NSC data. Over two dozen predictors were evaluated. These included baseline measures of demographic characteristics, academic history and performance, and foster care history characteristics. Other predictors included several pairs of risk and protective factors measure before (pre-entry) and after (post-entry) participants entered college, as well as college-level characteristics. Multivariable logistic regression was used to evaluate the likelihood of credential attainment, and bivariate probit models were run with the full Midwest Study sample (n=732) as a robustness check to investigate possible selection effects. Multiple imputation was used to address missing data.

Results: Males were significantly (p<.05) less likely than females to earn a degree (OR=.41) as were youth who first entered college after age 20 relative to youth who entered before age 19 (OR=.19). Post-entry parenthood (OR=.37), number of economic hardships (OR=.67), and full-time employment (OR=.22) each decreased the odds of completion. Higher levels of pre-entry social support positively predicted completion (OR=1.91). Several college-level factors increased the odds of completion, including selectivity, % of students receiving a Pell grant, expenditures on academic services and on student support. Bivariate probit models yielded similar findings as logistic regression models.

Conclusions: Some of the strongest impediments to finishing college involved life circumstance after participants enrolled in college. Financial hardships and the need to work hindered completion. Males, individuals who enter school later, and students who are parents are subgroups that may need particular support. Bolstering youths’ social support networks, beginning before they enter college, may connect them with resources promotive of college completion. The institutions that youth attend also play a critical role; schools with higher expectations and more resources to support students appear to increase foster youths’ chances of graduating. Given their multiple barriers, it is argued that early, targeted interventions that remain in place as other foster care supports phase out will be integral to support these young people through college.