This paper makes clear the factors that determine the research outcomes of doctoral candidates in the engineering field of the American research university. Tinto's longitudinal doctoral persistence model and Astin's I-E-O model are used as reference frameworks. The data comes from the survey conducted by the American Research Council for the doctoral candidates at the American Research University in 2006 (NCR 2006). Research outcomes of doctoral candidates are measured by the number of research presentations and refereed articles. Survey results show that financial support for research travel has a positive effect. Relationship with faculty advisors also has a positive effect to produce refereed articles. Research productivities of graduate students are not only important for their success at graduate school but also important for productive academic life through their career. Therefore, the findings of this study give suggestions for improving graduate education and their career.