Optimizing Transfer Pathways in Higher Education

Yiming Zhang; Gregory L. Heileman; Ahmad Slim; Husain Al Yusuf

Download Paper | Permalink

Conference: 2024 ASEE Annual Conference & Exposition
Location: Portland, Oregon
Publication Date: June 23, 2024
Start Date: June 23, 2024
End Date: July 12, 2024
Conference Session: DSA Technical Session 3
Tagged Topic: Data Science & Analytics Constituent Committee (DSA)
Permanent URL: https://peer.asee.org/47821

Request a correction

Paper Authors

biography

Yiming Zhang The University of Arizona

visit author page

Yiming Zhang completed his doctoral degree in Electrical and Computer Engineering from the University of Arizona in 2023. His research focuses on machine learning, data analytics, and optimization in the application of higher education.

visit author page

biography

Gregory L. Heileman The University of Arizona orcid.org/0000-0002-5221-5682

visit author page

Gregory (Greg) L. Heileman currently serves as the Vice Provost for Undergraduate Education and Professor of Electrical and Computer Engineering at the University of Arizona, where he is responsible for facilitating collaboration across campus to strategically enhance quality and institutional capacity related to undergraduate programs and academic administration. He has served in various administrative capacities in higher education since 2004.
Professor Heileman currently serves on the Executive Committee of AZTransfer, an organization that works across the system of higher education in the State of Arizona to ensure students have access to efficient, seamless, and simple ways to transfer from a community college to a university in Arizona. He serves on the board of the Association for Undergraduate Education at Research Universities, a consortium that brings together research university leaders with expertise in the theory and practice of undergraduate education and student success. In addition, he is a fellow at the John N. Gardner Institute for Excellence in Undergraduate Education.
Professor Heileman’s work on analytics related to student success has led to the development of a theory of curricular analytics that is now being used broadly across higher education in order to inform improvement efforts related to curricular efficiency, curricular equity, and student progression.

visit author page

biography

Ahmad Slim The University of Arizona

visit author page

Dr. Ahmad Slim is a PostDoc researcher at the University of Arizona, where he specializes in educational data mining and machine learning. With a Ph.D. in Computer Engineering from the University of New Mexico, he leads initiatives to develop analytics solutions that support strategic decision-making in academic and administrative domains. His work includes the creation of predictive models and data visualization tools that aim to improve student recruitment, retention, and success metrics. Dr. Slim's scholarly contributions include numerous articles on the application of data science in enhancing educational practices.

visit author page

biography

Husain Al Yusuf The University of Arizona orcid.org/0000-0002-4769-2089

visit author page

Husain Al Yusuf is a third-year PhD candidate in the Electrical and Computer Engineering Department at the University of Arizona. He is currently pursuing his PhD with a research focus on applying machine learning and data analytics to higher education, aiming to enhance student outcomes and optimize educational processes.

Husain Al Yusuf holds an M.Sc in Computer Engineering from the University of New Mexico and brings over fifteen years of professional experience as a technology engineer, including significant roles in cloud computing and infrastructure development at a big technologies company and financial services industry.

visit author page

Download Paper | Permalink

Abstract

In higher education it is common for students to transfer from one institution to another for various reasons, with the hopes that prior earned credits will be accepted at the intuitions they are transferring into. A typical scenario for transfer students involves those admitted to community colleges planning to later transfer to 4-year universities in order to pursue bachelor’s degrees. Research on the transfer process indicates that, on average, transfer students lose credit hours equivalent to one year of coursework. Given the vast number of transfer students nationwide, such significant loss of credit hours represents a significant waste of valuable educational resources that should be avoided in order to improve student success outcomes. However, finding efficient and effective transfer pathways between institutions is challenging, particularly when accounting for program requirements that are constantly changing, students changing their major plans, the creation of new courses, etc. Crafting a suitable plan for transfer students demands expert knowledge, effort, and sometimes collaboration among multiple institutions. Managing all of this complexity manually is partly accountable for the credit loss issue mentioned above. In this paper we consider the role that data and analytics can play in addressing this problem.

To gain a deeper understanding of this challenge, we first formally define the Optimal Transfer Pathway (OTP) problem, which involves finding a two-year to four-year degree plan that can be used to satisfy the degree requirements from both a community college and a 4-year university using a minimum number of credit hours. We consider the significant data requirements necessary to solve the OTP problem. These include collecting the Boolean formulas that describe all degree requirements, the courses that may be used to satisfy these requirements, as well as the transfer equivalencies that exist between institutions. The combinatorics associated with finding degree pathways between any associates degree and any bachelor’s degree make this problem exceedingly difficult, and a proof of the NP-Completeness of the OTP problem is provided. Thus, solving this problem through an exhaustive search in a reasonable amount of time is computationally infeasible. To address this issue, we treat the OTP problem as an assignment problem that seeks a feasible course-to-degree requirements assignment. In particular, we describe a 0-1 integer quadratic programming algorithm for the OTP problem that returns near optimal transfer plans in a reasonable timeframe. Experiments with this algorithm, using real degree requirement data from two Arizona institutions, have yielded insightful results regarding degree completion plans. The solution was created using the JuMP mathematical optimization modeling language, implemented in the Julia programming language, and is solved using a commercial optimizer. The analytical results returned by this system allow students to clearly understand how each course is used to meet specific degree requirements, which courses are transferable or not, and the reasons for their transferability. Additionally, it facilitates the consideration of multiple completion plans by advisors, which is beneficial for future degree requirement designs. We conclude with a discussion on leveraging this algorithm to meet the more tailored requirements of individual transfer students.

Citation
Format

Zhang, Y., & Heileman, G. L., & Slim, A., & Al Yusuf, H. (2024, June), Optimizing Transfer Pathways in Higher Education Paper presented at 2024 ASEE Annual Conference & Exposition, Portland, Oregon. https://peer.asee.org/47821

TY - CPAPER
AB - In higher education it is common for students to transfer from one institution to another for various reasons, with the hopes that prior earned credits will be accepted at the intuitions they are transferring into. A typical scenario for transfer students involves those admitted to community colleges planning to later transfer to 4-year universities in order to pursue bachelor’s degrees. Research on the transfer process indicates that, on average, transfer students lose credit hours equivalent to one year of coursework. Given the vast number of transfer students nationwide, such significant loss of credit hours represents a significant waste of valuable educational resources that should be avoided in order to improve student success outcomes. However, finding efficient and effective transfer pathways between institutions is challenging, particularly when accounting for program requirements that are constantly changing, students changing their major plans, the creation of new courses, etc. Crafting a suitable plan for transfer students demands expert knowledge, effort, and sometimes collaboration among multiple institutions. Managing all of this complexity manually is partly accountable for the credit loss issue mentioned above. In this paper we consider the role that data and analytics can play in addressing this problem.

To gain a deeper understanding of this challenge, we first formally define the Optimal Transfer Pathway (OTP) problem, which involves finding a two-year to four-year degree plan that can be used to satisfy the degree requirements from both a community college and a 4-year university using a minimum number of credit hours. We consider the significant data requirements necessary to solve the OTP problem. These include collecting the Boolean formulas that describe all degree requirements, the courses that may be used to satisfy these requirements, as well as the transfer equivalencies that exist between institutions. The combinatorics associated with finding degree pathways between any associates degree and any bachelor’s degree make this problem exceedingly difficult, and a proof of the NP-Completeness of the OTP problem is provided. Thus, solving this problem through an exhaustive search in a reasonable amount of time is computationally infeasible. To address this issue, we treat the OTP problem as an assignment problem that seeks a feasible course-to-degree requirements assignment. In particular, we describe a 0-1 integer quadratic programming algorithm for the OTP problem that returns near optimal transfer plans in a reasonable timeframe. Experiments with this algorithm, using real degree requirement data from two Arizona institutions, have yielded insightful results regarding degree completion plans. The solution was created using the JuMP mathematical optimization modeling language, implemented in the Julia programming language, and is solved using a commercial optimizer. The analytical results returned by this system allow students to clearly understand how each course is used to meet specific degree requirements, which courses are transferable or not, and the reasons for their transferability. Additionally, it facilitates the consideration of multiple completion plans by advisors, which is beneficial for future degree requirement designs. We conclude with a discussion on leveraging this algorithm to meet the more tailored requirements of individual transfer students.
AU - Yiming Zhang
AU - Gregory L. Heileman
AU - Ahmad Slim
AU - Husain Al Yusuf
CY - Portland, Oregon
DA - 2024/06/23
PB - ASEE Conferences
TI - Optimizing Transfer Pathways in Higher Education
UR - https://peer.asee.org/47821
ER -