Every time a student interacts with an online course, that’s data. Every time a student uses an ID card to swipe into a building, that’s data. Every time a student pays a parking fine, registers for a course, or opens an app on their phone, that can be data too.
A college campus is an unending font of student information, and increasingly, institutions are collecting it, accumulating what is potentially a very valuable asset to improve teaching and learning. However, while the data easy to gather, not every institution has been able to take advantage of it.
The number of colleges and universities mining student data is on the rise. In September, CIO magazine reported that 30 percent of colleges and universities will be using predictive analytics by 2018. That leaves another 70 percent of schools that very likely own large data assets but don’t have a near-term plan to use the best tools to leverage it. What’s holding those institutions back?
In a recent white paper, The Data-Driven Campus — Using Learning Analytics to Optimize Teaching, Learning, and Student Persistence, to illustrate this challenge, we used the legend of the white elephant — a gift that too precious to dispose of but the care of feeding of which is so costly that it bankrupts the kingdom. Similarly many campuses may be coming to grips with questions about resources, technical limitations, human expertise, and student privacy. They may be intimidated by the simple enormity of campus data.
But, as the thought leaders interviewed for the white paper show, with effective tools, any campus can initiate the steps to put data resources to work in ways that improve student outcomes.
Who is parsing student data?
Instructors are, and always have been, analysts on a small scale, using insights gleaned from their LMS, to better understand their own students’ progress and engagement.
It can be more difficult, however, for a college administrator attempting to draw inferences about student behavior from a campus that collects data from almost every student activity.
Robbie K. Melton, associate vice chancellor for mobile and emerging technologies at the Tennessee Board of Regents, was interviewed by Campus Technology in September about data collection. He talked about the Internet of Everything on campuses; information gleaned from the Internet of Things, teamed with big data from other sources. Some of the devices he mentioned included smart wearables, smart floors, cameras, even a smart basketball.
Administrators often find themselves “drowning in data, but lacking insights,” according to Ray Henderson, managing partner of Lessons Learned Ventures and formerly president and chief technical officer at Blackboard. Henderson, who recently discussed the growing interest in adaptive learning, believes the market itself will make data analysis simpler for administrators.
He said, “I think we’ve got the right conditions to see independent learning management platforms that can blend the information you need about course performance and match that to the demographic, the financial aid, the information about the physical location of that student on campus and their habits to create richer portrayals of students and their progress.”
According to CIO, Purdue University is doing just that. Information Technology at Purdue (ITaP) has developed an app they refer to as “FitBit for academics.”
The app, Forecast, collects data on student behaviors and displays it back to the student to demonstrate how their choices affect their grades. Among the data collected by Forecast is adding classes on time, engagement in campus activities and their GPAs, as compared to similar students. Much of the data is information the school is collecting anyhow: registration data and grades, for example.
Not all institutions of higher learning are capable of this kind of initiative, however.
In the white paper mentioned earlier, Jessie Brown, an analyst of Ithaka S+R, a not-for-profit organization that works to improve learning outcomes using digital technologies, said, “There are a number of technical challenges. Do institutions have the infrastructure to check what they need to check? If they do, do they have the technical capacity to integrate this system?”
In many cases, the answer is no.
According to CIO magazine, college IT departments lack the financial resources of the private world; they’re maintaining older systems, working on a limited budget, dealing with years of technical debt, and they’re tasked with maintaining security in educational organizations, which typically prize open access.
Simply put, many colleges find their IT departments are spread thin.
The third party and student privacy
The simplest way around technical challenges on campus is to use a third party to analyze student data. Several third-party analytics tools exist for institutions that don’t have the technical, intellectual, or resource capacity to build their own programs for tracking new student data.
According to Brown, a growing number of institutions are turning to third-party tools, but the decision to turn student data over to outside companies can be a controversial one.
“Certainly, there is issues of privacy that come up if you’re sharing institutional data with a third party,” Brown said. “And I think one of the bigger issues that has come up lately with institutions using third party analytics tools is oftentimes it’s unclear to institutions what exactly the algorithms are that are used to predict success, which raises ethical issues.”
Some campuses are addressing this by allowing student to opt into data collection. The Purdue students using Forecast, for example, have chosen to use the app.
Where universities can start with data
It can be tempting for a school to delay analysis of information until they feel the perfect system for data analysis is in place, but John Kelly, managing director and a leader of Berkeley Research Group’s Predictive Analytics practice, suggests that campuses get started by simply diving in.
“It’s kind of the MacGyver approach,” he said.
He suggests that administrators find the most casual and impactful data relationships on campus first. In some cases, that might mean the most obvious data relationships, but that shouldn’t matter. Any success with analysis is a gain.
“It’s creating results to inspire further pursuit and experimentation and investment,” he said. “Not to get it perfect the first time around.”
A.J. O’Connell is a freelance journalist specializing in education reporting and a former college journalism teacher.