In 2015, one of the biggest buzzwords in education was predictive analytics, and it will certainly be a focal point for technology investment in 2016. Higher education is finding many uses for this technology. It’s ubiquitous in early alert and intervention, in prospective donor identification, and in content recommendation engines for adaptive learning.

Now enrollment management is emerging as a new area of use that is sure to impact both students and institutions. As more vendors offer integrated analytics with college planning, admissions workflow, and enrollment forecasting solutions, institutions will need to consider the ethical and legal ramifications.

Predicting Acquisitions

Since the release of our 2013 white paper (learn more in this post) on the use of predictive analytics in higher education, Eduventures has encouraged institutions to join the Predictive Analytics Reporting (PAR) Framework to benchmark their progress toward retention goals relative to peer institutions. Hobsons has been deeply involved with the PAR Framework since its inception. This month, Hobsons announced that it had acquired the organization and immediately made the largest shared pool of student outcomes data available to over 300 institutional clients through its suite of student success and retention products. This shared community dataset from PAR promises to help Hobsons build products that better identify which support structures and interventions work best for particular student populations.

Current higher education predictive analytics platforms have two key weaknesses. First, they lack large, diverse datasets on student outcomes. Second, they are not tied to specific services and interventions provided to individual students. Hobsons’ acquisition of the PAR Framework intends to address these weaknesses. In addition to the Starfish retention platform, the announcement indicates that Hobsons intends to bring PAR technology and the skills of the PAR team to bear on its existing college readiness and admissions products (called Naviance and Radius, respectively). Hobsons will attempt to push analytics earlier in the student lifecycle to inform all interactions between students, guidance counselors, and admissions staff members. The goal is to show some form of analytics output any time that choices or decisions are being made. These analytics can help assess the best fit or path using what the community as a whole believes are the best options for individual students.

Predicting Your Chances

This would be the first time a for-profit technology vendor integrates predictive analytics in a college planning or admissions solution. While other college planning platforms have used data science to recommend best-fit colleges to individual students (see my own patent application for a college recommendation engine from ten years ago), this is the first time that outcomes data on a prior class of students would be considered in an admissions decision.

It’s a natural, almost inevitable, evolution for institutions to consider available data on what has made enrolled students successful in their admissions decisions. Comparative data among peers can help institutions assess their relative success in enrolling and retaining different student populations and determine how their academic programs stack up. Predictive analytics can also dispel long-held biases by showing that recruiting students from the same few high schools may not yield the best fit or most retainable students.

It’s important to recognize that predictive models play a dual role in admissions: assessing the likelihood of acceptance—a benefit to students—and assessing the likelihood of student retention—a benefit to institutions. Each of these uses of predictive analytics should be approached with consideration for the ethical and legal ramifications of using this type of data in admissions.

Predicting Legal Challenges

Should institutions use retention outcomes data on prior classes of admitted and enrolled students to decide whether to admit an otherwise qualified prospective student? For example, should a college’s poor track record in retaining students of color be a reason to deny admissions to a current applicant? Predictive models that consider retention rates by ethnicity or race may soon become illegal, pending the SCOTUS ruling on the University of Texas at Austin’s use of race in making admissions decisions. If the final decision restricts the use of race in admissions, institutions and vendors will need to reconsider the implications of using race or ethnicity at all in predictive models.

The federal government is also paying attention to other phases of the admissions process during which predictive analytics are used. The U.S. Department of Education (ED) now prevents institutions from seeing students’ college lists when distributing financial aid. A poorly kept industry secret, it was widely reported in 2013 that institutions were using the order of the colleges listed on a student’s FAFSA in their predictive models. These models determine the amount of financial aid offered based on, in part, a student’s likelihood of accepting. ED revised the FAFSA process in its latest version so that the Institutional Student Information Record (ISIR) provided to institutions no longer names the other schools applicants include in their FAFSA.

The bottom line is that institutions, in partnership with their analytics vendors, need to consider the legality of using certain types of student data in their predictive models. This is particularly relevant for models that are used to make admissions or financial aid decisions.

Predicting Success

Hobsons is certainly not the first organization to foray into enrollment analytics. The Coalition for Access, Affordability, and Success is another platform for college admissions that is well poised to both use predictive analytics in interesting ways and fill the hole left by the recent changes to the FAFSA. When the coalition application opens in 2016, thousands of students will apply to nearly 100 institutions through a single application process.

Conceivably, the student application experience could include an estimate of the chance of admission to a Coalition member institution based on data the system has on other applicants to the same college. Similarly, the platform could provide member institutions with tools to select which students to admit based on a best-fit analysis, not unlike how medical schools pool data on applicants and choose students based on a holistic review process. Conversely, if a student is admitted to a program that is only offered by a handful of institutions, schools could reasonably predict which they may choose based on other factors.

Another vendor, DecisionDesk, is planning to use an advisory board to help define how it uses predictive analytics within its admissions workflow platform. Its product roadmap already includes a dashboard to show how an admitted class is shaping up. This dashboard can help admissions officers and VPs of enrollment achieve the right composition for the incoming class. By actively pursuing institutional partners for participation in an advisory board, DecisionDesk hopes to answer the question, “If all of your data systems could report into a single location, what would you do with this capability?” This advisory board will also take up the task of determining best practices for the design of predictive models and which data elements can and cannot be used for admissions decisions. Ideally, these institutions would take into consideration the legality of using specific data elements, such as race and ethnicity.

2016 is shaping up to be the year of analytics in higher education. Analytics is hitting the mainstream, becoming integrated into all phases of the student lifecycle, and becoming a core part of vendor marketing efforts. Eduventures continues to assess the impact and challenges of integrating analytics into institutional business processes and student engagement. If your institution is using predictive analytics in new and innovative ways, particularly within recruitment, admissions, retention, career services or alumni advancement, please reach out to us and share your stories.