This study presents a unique synthesized set of data for community college students entering the university with the intention of earning a degree in engineering. Several cohorts of longitudinal data were combined with transcript-level data from both the community college and the university to measure graduation rates in engineering. The emphasis of the study is to determine academic variables that had significant correlations with graduation in engineering, and levels of these academic variables. The article also examines the utility of data mining methods for understanding the academic variables related to achievement in science, technology, engineering, and mathematics. The practical purpose of each model is to develop a useful strategy for policy, based on success variables, that relates to the preparation and achievement of this important group of students as they move through the community college pathway.