University of Roehampton leverages AI for student support and retention
Dr. Aleata Alstad-Calkins, Director of Student Support & Success at the University of Roehampton, discusses how the university is using data and AI to improve student retention and support at-risk students.
In this exclusive for ETIH, she explains how Roehampton’s predictive, proactive, and personalized approach is helping to boost continuation rates and ensure students receive the support they need to succeed.
The University of Roehampton is using data to address student retention rates by identifying and supporting those at risk of dropping out. Through advanced dashboards and targeted interventions, the university delivers personalized, preventative, proactive, and predictive student support.
Tackling retention with data-driven insights
Many first-year students returning after the winter break will successfully continue their studies, but an estimated 6% to 7% will not return in January. Financial difficulties, academic struggles, and emotional challenges are among the key reasons for withdrawal.
For higher education institutions (HEIs), student retention is a growing priority, particularly with the introduction of the B3 conditions by the Office for Students, which mandate an 80% first-to-second-year continuation rate. Addressing this challenge requires a proactive approach, and Roehampton is using data and AI to intervene early.
Using data to improve student success
Roehampton’s system tracks student engagement data from multiple sources, including application forms and pre-university surveys. This allows the university to identify students at risk before they arrive on campus. Groups more likely to face academic challenges—including first-generation students, care leavers, BAME students, and those from lower socio-economic backgrounds—receive additional support.
Outreach begins before term starts, with tailored induction programs for high-risk students. Many students, particularly those with disabilities, say Roehampton’s early contact helped them access support from day one.
The 4P approach: Predictive, Preventative, Proactive, and Personalized
Once the term begins, real-time student engagement dashboards track attendance, learning platform activity, library usage, and coursework submissions. Students are categorized into Red, Amber, or Green zones, enabling staff to identify and support those at risk of withdrawal as early as week six.
This multi-service model allows faculty and student support teams to connect students with relevant services, including disability advisors, mental health counselors, and academic support.
The School of Psychology was the first to adopt this approach. Vaithehy Shanmuganathan-Felton, Deputy Dean of the School of Psychology, noted that many students face work and caregiving responsibilities, making tailored support essential. “Many of our students struggle because they work full time or have caring responsibilities and may need a tailored support package to remain on their course. We want to provide the best support available for them to ensure that they thrive and succeed.”
Addressing mental health challenges
Mental health concerns are a key factor in student retention, with 25% of students citing it as a reason for dropping out—a 210% increase since 2009/2010. However, many do not disclose their struggles to their university.
At Roehampton, 76.4% of students said that wellbeing support was crucial to their continuation. Data-driven interventions have shown measurable results—68% of students who received counseling continued their studies after being at risk of dropping out.
Technology-driven solutions
Roehampton initially relied on manual tracking through Excel spreadsheets to monitor student engagement. Now, automation has streamlined the process.
Shanmuganathan-Felton explained how AI is enhancing student support: “Now the dashboards have all the information available to us; it’s a very quick process to mine it for information. To save time, we developed a custom ChatGPT to analyze anonymized engagement data and create a profile and tailored support plans for our most at-risk students. Now, it takes about an hour to create profiles for all Psychology undergraduates, which we can then share with them.”
The results have been significant:
55.5% of all students showed improved engagement scores.
Red-rated students saw the greatest improvement.
42% reduction in non-submissions among students receiving targeted support.
1,319 disabled students achieved, on average, 5% higher final degree results than their non-disabled peers.
Looking ahead, Roehampton is exploring machine learning and predictive analytics to forecast student needs and offer support before issues arise.
Results and sector impact
The School of Psychology’s continuation rate has risen by 6% to 92%. Across the university, 82.3% of students in July 2024 credited the support system as a key factor in their decision to stay—up from 74% in 2023.
This approach has earned the University of Roehampton a place on the shortlist for the 2024 Times Higher Education Awards for Outstanding Support for Students.
Sector-wide implications
Across higher education, dropout rates have increased by 28% since 2018, according to the Student Loans Company. Universities have access to vast amounts of student data, but many are not leveraging it effectively to improve retention.
By investing in AI-driven predictive analytics, proactive outreach, and personalized support, institutions can improve student success rates, reduce lost tuition revenue, and maintain rankings and reputation. The University of Roehampton’s model highlights how data can be used to address retention challenges and ensure more students complete their degrees.
Dr Aleata Alstad-Calkins, Director of Student Support & Success, University of Roehampton
