By Suzanne Wilson Summers
Higher education is increasingly embracing the use of big data to increase and assess the effectiveness of institutional policies and practices and to drive needed change. The Central Florida Education Ecosystem Database (CFEED) offers one promising model for regional data-sharing agreements that can increase educational attainment.
With a shared commitment to student success and an understanding that academic progress requires joint effort across the educational pipeline, Valencia College, the University of Central Florida (UCF), and the Osceola and Orange County public school districts created CFEED in 2018. Combined, these four entities enroll approximately 400,000 students in the Orlando region.
Both located in Orlando and founded in the mid-1960s, Valencia College and UCF have grown up together, according to Valencia’s President Sanford C. “Sandy” Shugart, and the two have a long history of working closely together to promote transfer success. In 2006, they created a 2+2 transfer program called DirectConnect to UCF, which guarantees Valencia students entry into a bachelor’s program at UCF. UCF, in turn, provides staff on Valencia campuses to help DirectConnect students onsite with advising, admissions, financial aid, and academic support. Faculty from both institutions teach on each other’s campuses and work together to tune course objectives and teaching methods to ensure that common methods and course learning objectives are used.
CFEED builds on and expands the work of DirectConnect to include the two local public school districts. It is part of a broader regional effort to increase the numbers of students graduating from high school and attaining a postsecondary credential by developing actionable data that will allow the partner institutions to identify and improve policies affecting completion and transfer.
An initial grant from the Helios Education Foundation provided start-up monies to create the necessary data architecture. Given the technical complexities, part of the initial funding was used to hire an outside firm to construct the system architecture while the partner institutions fed 10 years’ worth of student information into a data lake that staff then curated to produce useable data. As intended, after the grant’s expiration in June 2020, the operating costs for CFEED shifted to the institutional budgets of the partner institutions.
An important feature of CFEED is that student-level data is individualized, meaning that students are identified in the database by name, rather than anonymized. The use of individualized data will eventually allow the creation of algorithms that make it possible for institutions to identify specific students who are struggling and to develop real-time interventions designed to keep them enrolled and on track toward completion. To address privacy and security concerns, partner institutions signed a memorandum of understanding outlining the uses of student data and commitments to follow federal and state privacy regulations.
Institutional needs shape how partner institutions use CFEED. In Florida, state policies actively promote college going by ensuring that high school students have opportunities to earn college credits prior to graduation. So, using CFEED data, the Osceola County school district found that having students complete biology in eighth grade, instead of ninth, meant that they would be able to take an additional dual enrollment course that they could apply toward a college degree. According to Osceola’s CIO Thomas Hoke and Nicole Scala, a research analyst for CFEED, the district will be using CFEED data to determine if one result of this change is that these students are more likely to declare STEM majors once they enroll at either Valencia College or UCF.
Likewise, Orange County’s director of data strategy, Thomas Hoke, told me that they are tracking the impact of course sequencing on postsecondary GPAs. Eventually, they hope to understand how pre-kindergarten and elementary educational experiences affect the likelihood of high school graduation and postsecondary enrollment. As Osceola’s Peter Thorne put it, “We know the scores the student makes. We don’t yet know what makes the student.”
As a two-year state college and the largest source of transfer students to UCF, Valencia College has an interest in smoothing both ends of the transfer pipeline. Although the state of Florida does not assess how students transferring from high school to college fare, Valencia is using National Student Clearinghouse data, along with individual student records, to allow the partner school districts to track the progress of their graduates who enroll at Valencia. For example, the school districts can see the outcomes of students who take dual enrollment courses versus those who do not and can, therefore, identify ways to enhance first-term advising. Additionally, they can monitor student progress at six, nine, and 12 credit hours to ensure momentum toward completion.
Expanding on the DirectConnect partnership, CFEED provides student-level information about how students from Valencia fare after transferring to UCF. This is especially important since Florida has added retention and graduation of transfer students as metrics in its funding formula for universities, according to Linda Sullivan, the assistant vice president for institutional knowledge management at UCF. CFEED includes information on students’ financial aid status and demographics, which may determine student outcomes.
It also provides an in-depth look at how different transfer patterns affect student outcomes. UCF, for instance, is using CFEED data to look at how attendance of high school graduates who first enroll at Valencia and then transfer to UCF compares with outcomes of students transferring from other state colleges. By comparing data of native students at UCF to those who transfer from Valencia, it can help to identify areas for improvement. Moreover, CFEED is helping to identify for DirectConnect students which specific course patterns and programs at Valencia are most predictive of successful completion after transfer to UCF.
Analyzing student data has helped Valencia and UCF identify sources of transfer shock that, in turn, allows both institutions to develop interventions or to change policies. Sullivan and Magdalena Fernández-Civil, institutional research analyst, told me that CFEED data revealed that transfer students with a GPA of less than 2.56 are most likely to experience shock. Students’ choice of majors matters too. CFEED data revealed that STEM majors are most likely to experience transfer shock, whereas for psychology majors, taking at least a few courses at Valencia enhanced the likelihood of success. Across majors, students’ exit GPA from Valencia and course history, along with the GPA earned during the first term at UCF, are critical predictors of retention.
Now that construction of the system architecture has been completed, the partners want to move to create reports that allow for real-time interventions for individual students. Although at this stage, CFEED’s analysis is performed on end-of-term data, Valencia’s Brandon McKelvey, vice president of analytics and planning, institutional research, said that the eventual goal is to perform daily downloads during the first two to three weeks of each term so that institutions can do just-in-time interventions that would keep students enrolled and on the path to completion.
Additionally, as its database grows, CFEED will incorporate predictive analytics. Currently, it tracks 500 potential predictors of student success across pre-K–16. Having evidence to suggest which are most critical to students at each part of the pipeline will allow institutions to make system-level changes.
For institutions that might be considering data sharing agreements, CFEED Director Diana Pienaar emphasizes the need for good partners with a shared commitment to ensuring student success. Few institutions have the technical expertise to set up such an effort solo, she said, so finding and hiring a strong technical partner is critical.
CFEED represents an ambitious and innovative model of how institutions across the educational pipeline can use data to work collaboratively to increase college-going and completion. Treating transfer as an ecosystem, rather than as an institutional issue, opens the door to collective action to ensure that students can move more seamlessly across the K–20 pipeline.