Epidemiology of community college completion: Applications of epidemiological methods to solve a mystery in higher education

Community colleges have increased post-secondary educational access for disadvantaged youth, but few graduate. Half of community college drop-outs cite unspecified "personal reasons," twice as many as cite "family" or "finances." Federal education data cannot track risk behaviors that might inhibit college graduation. This study identified factors that predict educational attainment for future in-depth investigation. Using Tinto's theory of school retention, we identified 150 factors that might predict community college graduation. We tested whether each factor predicted educational attainment among National Longitudinal Study of Adolescent Health respondents enrolled in community college in 2001 (n=1774) using a Poisson working model with educational attainment outcomes in 2008, adjusting for pre-college variables measured in 1995 (e.g., test scores, risk behavior, and parents' socioeconomic status.) In 2008, 13% had earned certificate(s), 28% earned AAs, and 20% earned a BA or above. Gay/lesbian identity and methamphetamine use predicted 40% lower chances of AA/certificate (incidence rate ratios 0.57 (0.40, 0.82), 0.59 (0.40, 0.89)) and respectively 70% and 50% (IRRs 0.28 (0.09, 0.82), 0.51 (0.23, 1.13)) lower chances of BA, and stuttering predicted 60% lower chances of BA (IRR 0.41 (0.21, 0.81)). Other factors predicting lower educational attainment included obesity, activity limits, having been in alcohol treatment, distrusting government, and living with grandparents. Factors predicting higher educational attainment included living in a dorm, voting, political party affiliation, trusting government, bank account, email, computer access, and continuous health insurance. Religiosity, police involvement, and most substance use did not predict educational attainment. Future in-depth investigations will identify mediating factors for designing interventions.