Skip To Content
Rapid Insight: Higher Education case study

Bunker Hill Community College (BHCC) Uses Predictive Analytics to Support Student Success Initiatives

Bunker Hill Image
  • After explosive growth in student enrollment, BHCC proactively increased student retention efforts

  • BHCC can now focus specific resources on students who are more likely to attrit by intervening with proactive conversations

  • With the help of predictive analytics, BHCC has seen a 1-2% increase in college-wide retention for the past 7+ years

Achieving the Dream Leader College and 2014 Leah Meyer Austin Award Recipient uses Predictive Analytics to Support Student Success Initiatives

Bunker Hill Community College is dedicated to innovating and refining student success strategies reflected in the role of an Achieving the Dream Leader College. The mission, vision, values, and strategic goals of the non-profit organization underline BHCC’s efforts in student success, many of which have been nationally recognized. Although predictive modeling is often presumed to be a tool used by only four-year institutions, BHCC has disproved this notion by innovatively using Rapid Insight software to boost their retention efforts around student success.

Bunker Hill Community College is the largest community college in the Commonwealth of Massachusetts. With over 14,000 students, two campuses (one in Boston, Massachusetts, and the other in Chelsea, Massachusetts), and three satellite locations to manage, BHCC innovates and partners with the communities it serves to support their students. In 2007 they joined Achieving the Dream, Inc. and are an influential partner of the national initiative’s movement focused on success strategies involving retention, student support, and college completion.

The cornerstone of the Bunker Hill’s student success agenda has been the creation of Learning Communities across the curriculum, supported by integrated student services and comprehensive student learning outcomes assessment. BHCC’s Learning Communities offer engaging and supportive environments that emphasize active and interdisciplinary learning. The College has implemented two levels of learning communities to deepen student learning and engagement: The Learning Community Seminar for First-Year Students and Learning Community Clusters.

The Seminar enables incoming students to explore a topic of interest as they orient to the college environment. Seminars feature integrated supports such as Success Coach advising and peer mentoring. Clusters enroll the same group of students in two or more courses, enabling students to learn and study together.  Faculty members have planned Clusters around common academic themes, and they offer interdisciplinary learning and hands-on activities such as field study and service learning. Clusters have also been the primary vehicle for accelerating and contextualizing developmental education pathways. Students can enroll in Clusters that allow them to take two levels of developmental math or English coursework in one semester; one level of developmental coursework and one college-level course in one semester; or one ESL course and one college-level course in one semester.

Compared with non-learning community peers, learning community students consistently persist both semester-to-semester and year-to-year at higher rates; accrue more college level credits; and complete developmental education pathways faster. The college received the 2014 Leah Meyer Austin Award “for building whole-college solutions to increase student success and achieve equity outcomes” as noted in the award sponsored by The Leona M. and Harry B. Helmsley Charitable Trust and administered by Achieving the Dream, Inc.

With such a strong focus on student success reflected in its curriculum, academic pathways, and student success strategies, BHCC looked to predictive modeling to further enhance retention and other data-driven efforts supporting these innovations. Community colleges often see a rise in enrollment during a recession. In recent years BHCC has seen an explosive growth in student enrollment. Now, they want to take additional measures focused on student retention because, as they know through national trends, the student enrollment growth would not last as the economic recession conditions improved. BHCC continues to utilize data to examine their strategies to their fullest potential and to further inform student success and proactive interventions.

Customer Testimonial

“Being able to identify the inter-workings and impacts of our strongest strategies by using data to inform where we can do better to support student success was the main appeal of why we looked at Rapid Insight’s Predict for our predictive modeling tool. Plus, the ease of use, the training, and support that we have had as part of the process has been excellent,” noted Karalynn Gau, Director of Research Strategy & Applied Analytics in the Office of Institutional Effectiveness at Bunker Hill Community College.

Since enhancing their student success strategies as part of college community’s work, BHCC has seen a percentage point or two increase in college-wide fall retention year-to-year for the past seven years. They want to keep building on that success by inquiring into how they can use data to further conversations integrating student success supports into the curriculum. According to Gau, community college enrollments are “inversely” tied to the national economy. With economic recovery here, BHCC wants to stay ahead of the enrollment decrease as they have watched the national trends concerning community college enrollment. “The model was really great because it allowed us to focus on some things we’ve already been doing as an institution,” Gau shared. “The aggregate report has been very useful and having the contact sheet at the end of the report is easy to see, on the individual student level, who is most likely to be retained.”

Gau expanded on how the institution has truly taken advantage of the attrition model they ran. Since then, they have been able to focus specific resources on students who are more likely to leave the school by intervening with proactive conversations. BHCC gathered a group of 20 callers to utilize data out of the fall-to-spring retention model. Each member of the calling team was trained by the Director of Advising and Life Map about common questions and potential issues facing students re-enrolling at the College. The group of callers included a mixture of faculty, administrators, staff and team members from advising and LifeMap. The Office of Institutional Effectiveness created a survey that each caller completed during their conversation with each student. These outreach phone calls allowed BHCC to collect formal data about why students were not re-enrolling for next semester to inform future proactive messaging to students and measure the impact of the call campaign. Almost 600 phone calls were made. “Being able to say we have 5,000 students left to reregister and here are 600 that we are going to target for re-enrollment based on predictive analytics was really helpful for us,” confirmed Gau. “Being able to be more proactive with strategies based on data and triaging resources accordingly has been really helpful as well.”

As for the acceptance of predictive analytics on campus, Gau thinks that it has been very well received by the community. “Retention has always been a big focus of our student success strategies, but now we’re even more focused on it given national decline in community college enrollment. I think folks are open to understanding how data can help us further our different student success agendas. Digesting and talking about it has led to generation of other ideas for models and ways we can look at applying predictive analytics at the college,” noted Gau. For instance, they found that being awarded unsubsidized loans was a surprising significant attributor to the retention model across student subgroups examined while the model confirmed descriptive analyses showing Pell Grant recipients are more likely to be retained.

Bunker Hill Community College is paving a new path for community colleges by implementing predictive modeling into their research strategies. Although community college student data is often different from what is available at four-year institutions, open access colleges like BHCC can translate variables and use the data-points that make the most sense for them and the communities they serve.

“Predictive analytics have helped us see things in our data that we knew were there, but we might not have been able to get at using just descriptive analyses alone. Seeing it and seeing the complexity of it and the different points where we clean data or filter data has really increased the confidence level of what data we’re using and how we’re applying it to enhance student success,” concluded Gau.