Analysis of Enrollment: A Spatial-interaction Model

Enrollment levels are often considered as one of the success indicators of college programs. To design a focused and targeted incentive program for attracting students, enrollment managers need information on specific external factors that affect college choices. This paper asks whether income, distance, or geographic locations of potential students are important determinants of student migration and sheds light on the expected patterns of spatial distribution of student enrollment if these factors are controlled. The analysis employed spatial-interaction or gravity models, which treat enrollment levels as interactions of a college or university with the student population in surrounding states or counties. By comparing the model predicted enrollment estimates with the actual levels, it is possible to identify the states or counties that are under- or over-supplying students to the intended college. With this information, enrollment administration bodies can design effective promotional packages for the under-supplying regions for future enrollment enhancement. The method is illustrated with limited data for West Virginia University but can be applied to any university with appropriate data.