An Analysis Of Graduation Rates At Research Universities

Many large research universities (i.e. with doctoral programs) have high undergraduate retention rates as measured by their 6-year graduation rates. The variation in graduation rates between schools and the relationship between graduation rates and student preparation levels were examined for these universities. A university was defined as successful with respect to its graduation rate if it had a 6-year graduation rate of 75% of more. With a preference for universities with strong technical programs such as engineering, universities were further selected with a high percentage (25% or more) of students pursuing a STEM major. The Education Trust database system was used to collect these data. Data of particular interest were the overall graduation rates, Under-Represented Minority (URM) graduation rates and the difference in graduation rates between the overall student population and the URM student population. In this research, a high priority was given to assessing whether there was a strong relationship between this difference in graduation rates and the preparation level of the student body as measured by the median SAT and the amount of money that flowed into each program (denoted as student related expenditures in the Education Trust database). A strong relationship between the differences in graduation rates was observed using multiple regression analysis. A significant difference in this relationship was found between private and public universities. In addition, Data Envelopment Analysis was used to determine which universities were “benchmark” universities or universities that were best in class relative to having a minimal gap between the six year graduation rates of the overall student population and the URM student population. Data Envelopment Analysis is a linear programming technique that establishes an “efficiency frontier” subset of points with maximum output values relative to the input values. This presentation will demonstrate the techniques that can be used by other researchers to benchmark a group of universities relative to graduation rates. If graduation rates of colleges of engineering were available, the same techniques could be applied. With the use of the Data Envelopment Analysis, an engineering university can compare itself to an appropriate benchmark university and use this information to improve its processes for achieving a higher retention rate of all students. The Need for a National Database on Graduation Rates for Engineering Colleges This research supports the need for a database to define the relationship between the graduation rates of engineering colleges and college preparation level as measured by the SAT or ACT scores. For continuous improvement there is a need to benchmark engineering colleges by graduation rates. There is no recognized national database on graduation rates of engineering colleges. Discussions with ASEE support staff indicated that ASEE does not request graduation rates in their college database. Other efforts at identifying a national engineering college database also were futile. P ge 11172.2 The most prominent university level aggregate database found was the Education Trust Database. This paper explores relationships observable between graduation rates and other relevant variables, such as college preparedness, using the Education Trust Database... The analysis includes large research universities with a high percent of STEM majors. A limitation of this analysis is that, for large research universities, the results represent experience of the entire university not just the engineering college or STEM disciplines. However, the research documented in this paper shows the benefits of such an analysis and demonstrates the potential of a national engineering college database, if it existed.