Date of Conferral







Alicd Eichholz


Community colleges are the largest segment of higher education institutions in the United States providing access to historically underserved populations and growing numbers of first generation college students. Increasing college degree attainment is a national priority with new expectations of accountability. Despite decades of educational research, community colleges have startling low completion rates. Within the framework of Tinto's theory of retention, a predictive analytics model could provide community colleges the opportunity to drive custom intervention and support services to students. The purpose of this study was to explore the utility of Biglan's taxonomy for categorizing courses for potential use in a data analytics model to identify students at risk of failure to complete. The quantitative census study used archival data from 1,759 students. Log-linear analysis was used to test the key research question as to whether there is a predictive relationship between type of course failed, as cross-categorized by the dimensions in Biglan's taxonomy, in the first term and failure to complete a degree or certificate within 6 years. The analysis showed that a more parsimonious model, based on the interaction term for the life/nonlife and pure/applied Biglan categories, appeared related to completion, although no standardized residual was significant. A larger and more diverse sample may be necessary to determine the true effectiveness of Biglan's taxonomy as a classification schema in a predictive analytics model of degree completion. Based on these results, first term course failure appears to be a logical point for programmatic support that could lead to higher levels of associate degree completion opening doors of employment opportunity through education, thus supporting social change.