Date of Completion


Document Type

Honors College Thesis


Business Administration

Thesis Type

Honors College

First Advisor

David Novak

Second Advisor

Thomas Chittenden


data warehouse, retention, data mining, student retention


One of the biggest concerns of universities across the United States is the student retention rate. Because it is much more cost effective to keep an existing student enrolled than to enroll a new student, improving a university’s retention rate translates to a saving in costs for that institution. UVM’s first-year retention rate is currently 85.8%, which places them above many other public universities, but below most of UVM’s aspirant schools. UVM conducted a study in 2011 in an effort to determine causes of students leaving after their first year, but retention rates since the study have only marginally increased. Some universities have been using data mining techniques to determine factors correlated with student retention, such as living off campus or an income level below the poverty line. This thesis recommends that UVM create a data warehouse aggregating all student-related data from across campus in an attempt to improve student retention. There is currently no central repository of student-related data from sources such as Residential Life, Blackboard, Student Health Services, and Undergraduate Admissions. Data mining techniques could be used with this data warehouse to discover patterns between different fields of data and a student’s likelihood to withdraw from UVM. For example, what if there is a correlation between a student’s dorm view room and their likelihood to leave UVM? How does a student’s frequency of Blackboard use impact their chance of staying enrolled? This thesis explores the technical and logistical considerations involved in a large data warehousing project. While building a data warehouse may seem operationally daunting, the insights it could generate would be very beneficial for decision support for many years.

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.