Hobsons' Predictive Analytics Integration: From the PAR Framework to Holistic Tools
A Q&A with Ellen Wagner
It's been a year and a half since Hobsons acquired the PAR Framework and began integrating PAR's predictive modeling into its product offerings. Is predictive analytics well-established now, in our common workflows and tools? Here, VP for Research Ellen Wagner talks about how Hobsons' integration of predictive analytics, including the PAR Framework, has provided the education sector with tools to approach a myriad of issues more holistically — from college and career planning, to recruitment and student matching, to admissions, advising, and student success.
"We'll see analytics as more of a standard tool in everyone's workflow — a tool that ultimately produces more decision-making, and more action, at more levels. And so much will come from all that to benefit students." — Ellen Wagner
Mary Grush: I want to ask you about how we are starting to use predictive analytics today in higher education, but let's approach that using an example from your own work. Let's start with your current passion — your interest in the holistic use of predictive analytics, in areas which you describe as spanning from recruitment, through enrollment management, and on through advising and ongoing student support. Could you explain a little about your own journey that brought you to your current fascination with areas such as admissions and enrollment management?
Ellen Wagner: I've been thinking a lot about enrollment management since I joined Hobsons a year and a half ago. It's been an eye-opening experience for this former professor. You know, those of us on the academic side of the house typically didn't think much about where our students came from. We just knew they would be in our classes, and that we would be responsible for making sure they learned in our classes and programs. In my time at Hobsons, I have come to understand just how much effort goes into enrollment management, student support, and student affairs.
Hobsons is well-known for its great work with college and career readiness, and the tremendous support provided to high school counselors, students, and their families. We are also tightly connected to the world of college recruitment, enrollment management, and admissions. We've known for years that helping students and their families develop college and career aspirations, along with the ability to research options deeply, isn't something that happens in the weeks before high school graduation. The earlier we start, the greater the probabilities of giving our young people opportunities for dreaming bigger and for seeing ways to realize their dreams.
It's always striking when my higher ed colleagues speak of the moral imperative we have to support the students we admit into our higher education institutions. Given my new awareness of just how much work goes on before students ever cross the college threshold, I have been wondering: What if we considered connecting all the support those students get before they go to college with the support they get once they have arrived at college? We can do this by collaborating more actively with our K12 colleagues to build the pathways we know are essential to doing a better job of supporting students along their entire learning journey.
Grush: Where in that storyline would you especially want to do a better job of supporting students?
Wagner: There is a large ecosystem built around getting K12 students to think about colleges, including recruitment, college visits, admissions — it's an entire enrollment management industry.
In all my experience over the years in academic affairs, of course I knew what was going on there, but I don't think I actually realized just how much of a business engine enrollment management has to be for a college or a university, and what a large piece of institutional operations it is.
What I have come to understand is that so much of the work of finding students for our colleges and universities — even at very prestigious schools — has been relatively arbitrary.
For example, why is it considered good practice for institutions to purchase lists of 100,000-plus names from the College Board or from ACT? In what way is it productive to send piles and piles of mail and marketing collateral to all those names in the hope that maybe some small percentage of names will eventually make their way through that funnel?
And, there is a resulting circumstance that Hobsons is actively working to change: Why is it almost a point of pride for even prestigious institutions to lose a substantial percentage of the freshman class every year? Why do they seem almost happy to point out that so many students probably should not have been admitted in the first place?
Grush: How would you fix all that?
Wagner: Why couldn't we work up front more, to find out things that would match the student to the college as well as to look for the kinds of things that would maximize their fit with the institution — based on an interest in the culture, or the types of majors that we have, or any of the numerous things that make our institutions the kinds of unique, special places that they are.
I've been reflecting lately that the traditional college was more like a factory model, where the college would take the "raw material" of a young incoming student and mold them into the perfect alum. These days it's less a matter of stamping out alums. Today it's more a matter of maximizing the unique strengths that each individual brings to the college.
Grush: Can more of that type of change happen now?
Wagner: Now, given all the opportunities we have to offer more personalized experiences with the data we have about student interests, performances, risk profiles, strengths profiles, and the like, it does seem that we can do a better job during the first few weeks of the relationship, by using more customized enrollment management practices. We know from our PAR research that we can do a highly effective job finding students at risk, identifying variables putting them at risk, and finding interventions that have been empirically shown to improve student performance. Why wouldn't we all want to accelerate our student success efforts by starting even before students start college, by doing a better job of matching students with colleges?
Grush: So hopefully then you don't start out with something that you need to fix later?
Wagner: Yes!
Grush: And this is something you feel more hopeful about at this point in your career?
Wagner: Yes. This has been my big "ah-ha" at Hobsons — that you really can, and we do, look holistically at a lifetime of learning. And, especially using predictive analytics, we can address important questions much earlier in the student's learning lifecycle. Hobsons' most recent product release, called Intersect, is a matching platform that allows colleges and universities to connect with students who are a "best fit" for their institutions and encourages high school counselors and college admissions counselors to join forces and work together to help students.
Grush: Have you — through your work at Hobsons — started to work directly with K12 on some of these issues?
Wagner: Yes. Actually, my immersion into the world of middle school and high school has been one of the other huge learning curves I've enjoyed in my time at Hobsons. We have a product called Naviance, a college and career readiness platform, that is currently used in 40 percent of schools featuring, in the U.S., grades 6-12. We have been working with the AASA, the superintendent's association, to support their initiative called "Redefining Ready," which is rethinking the whole notion of what it means to be ready for college.
Grush: What are some of the things that type of research would capture? Would you create predictive models for all this?
Wagner: In some cases we will create predictive modeling; in other cases it might be descriptive research. In certain cases we would likely use inferential statistical techniques. It really depends on the types of research questions to be answered.
Here are a few examples: We are looking into data reflecting students' authentic learning outcomes, identifying and measuring high-impact practices in K12 education, and working with school districts to learn about how extracurricular activities have been integrated into classroom activities to establish connections between what students are learning in the classroom and how they are gaining the skills they need to apply that knowledge to what is going on in the world around them.
Last year we completed some survey research with AASA and AACC on the topic of dual credit programs. We also completed a similar survey with AACRAO, to explore dual credit from the perspective of registrars. In April, we finished a research report with the American Council on Education, exploring what can be done to improve the odds of upward transfer. We used a nationally representative data source and a multivariate, multilevel model which took into consideration the nesting of students within high schools, and we tested the influence of a series of independent variables on upward transfer. Dual enrollment turns out to be one of the indicators to predict success in transfers from community colleges to four-year institutions.
The more time that I spend in the K12 world the more it feels as if higher education is "last to the party" when it comes to understanding that education is a holistic enterprise; that is, that it's not just up to what happens at the college. Practitioners across the entire education ecosystem need to understand that it's up to all of us as education stakeholders to make all the pieces fit, throughout the education lifecycle.
Think about all the energy that so many of us in higher education have poured into predictive analytics to help find our own students at risk. Why couldn't these predictives have come much earlier? Perhaps more to the point, why couldn't we use a common framework for guiding progress, outcomes attainment, and persistence goals so that students would not find themselves forced to start over every time they graduated and moved on to the next phase of their education experience?
If we start to look at all the different points in the lifecycle of students throughout the education system, we can do a lot to expand the conversations that help make students successful — before they become high risk. And of course there is a natural role for predictive analytics there.
Grush: Before Hobsons, you were concentrating on building the PAR framework, as Chief Research and Strategy Officer for PAR. Then Hobsons acquired PAR in 2016. What is the role in all of this now, for PAR?
Wagner: PAR is now an integral part of Hobsons' new, improved Starfish student success solution. You won't find PAR as a standalone product going forward, although you will certainly still find PAR methodologies, data, and research agendas driving conversations and research efforts. Considering that PAR is a framework — a set of methodologies and common definitions used for predictive analytics reporting — this makes total sense. Hobsons has successfully taken PAR's predictive analytics reporting and its intervention measurement work and integrated that inside Starfish. This is a big step toward making predictive analytics an important part of our everyday, common decision workflow.
Once you finally see PAR working inside Starfish, you begin to think about how predictive modeling could work with other products. It's all based on helping institutions to leverage models from their existing data; creating projections from their past data. We're finding out that models are working their way into just about everything. This is really a great outcome and future for all the work we did on the PAR Framework.
Grush: So all the work on PAR has promoted the use of predictive analytics more generally, rather than carving out a specific niche for PAR as a product?
Wagner: Yes. Analytics are tools — holistic tools — the means for bringing order and meaning to data, the end of trying to see patterns that would not otherwise be visible. Hobsons is not an analytics company as much as we are a student success company. We're not defined by the tools we use, so much as by the problems we are trying to help solve.
And because the original work of the PAR team was so concentrated on the student success applications of the Framework, we have since been able to move on to focus on related things that matter — things like college and career readiness and the important goals supported by enrollment management.
There's an even wider range of issues that Hobsons products address, and I'm sure we'll see more predictive analytics integration going forward. And we'll see analytics as more of a standard tool in everyone's workflow — a tool that ultimately produces more decision-making, and more action, at more levels. And so much will come from all that to benefit students.
This will ultimately be a very holistic approach to predictive analytics that we'll all rely on for everyday tactical decisions — not just occasional strategic positioning.
To me, where PAR is now, is a very successful fulfillment of what we had always been working toward when we developed the PAR Framework in the first place. It's all about improving student success. It really feels like we've come full circle; now it's up to us to keep building on that Framework.