Higher Ed's Moneyball?

Our Ideas series is exploring innovation in education.

It's 20-year-old Randall Lofton's third shot at college. He's already wiped out twice. Too much partying and basketball, he says, and not enough studying. "I didn't apply myself."

Lofton is now trying to balance a full-time job with three classes at community college. He's taking a mix of online and in-class work at Valencia College in Orlando, Fla.

"So this is pretty much my last chance," he says. "This is something that I want to do, so I'm gonna work my heart off."

So far, he is. Lofton recently aced an online introductory composition class. One small thing that may have proved a big help: His professor sent several personalized messages of support. Lofton even keeps one email from his English professor, Neal Phillips, saved on his tablet.

"I screen-shot it and I saved it," Lofton says. "And he was basically saying, like, 'Don't quit, you're very hardworking in this class. The course is almost over; I appreciate your participation. I appreciate how you're very diligent and very intrigued by the work.' That touched me that somebody was paying attention. And I'm not just in somebody's class just as a name. That was cool."

The email from Phillips showed genuine care from an engaged professor. It was also, perhaps unbeknownst to Lofton, part of the community college's strategy to marry data science with interventions to improve student performance.

Call it higher ed's version of Moneyball. Only the goal isn't a World Series ring; it's to help more students stick with college, improve academically and graduate.

Fast analysis of disparate data streams, including your daily Web surfing, helps corporate America market its goods, chart growth and follow you across the Internet with annoying ads. Innovative predictive analytics are essential for businesses, especially tech companies. They've got annual conferences on the stuff.

So why aren't more of the best minds in higher education doing more to tap those data streams to improve teaching and learning?

Yes, colleges and universities have long used data to help shape policy.

But too often, critics say, the approach is retro — looking at performance data in spreadsheets that administrators funnel into sparsely read "best practices" reports for faculty. Or the information is kept in silos and used for alumni fundraising, marketing or for preparing reports for the board of directors.

It is, in effect, academic autopsy data. The information rarely has anything to do with helping improve student learning in real time.

Now, more schools such as Valencia College are dipping their toes deeper into the streams of what some have dubbed predictive and learning analytics.

Yep, more annoying ed-tech buzzwords. Basically it means big data goes to college, as software and digital platforms collect and sort academic data that teachers can then use quickly as an intervention strategy. The data collected include how engaged a student is online with course material, with discussion forums, and information on his or her academic background and other data points.

The goal is to use data science to improve learning, boost completion rates and help teachers and counselors better target academic interventions fast, with a compelling nudge, counseling or other outreach.

"It's our mission to democratize data," says Mark Milliron, co-founder and chief learning officer with the higher ed software company Civitas Learning, a leader in this field. "And to to get that data to people who need it most — the faculty, students and advisers." There are others doing similar work including Civitas competitors Starfish and Blackboard.

Valencia College is currently piloting evidence-based data-analysis tools by Civitas aimed at helping boost student engagement and performance.

It's early in the fall semester, but Phillips already sees danger signs for a handful of students in his online Intro to English Composition class. Some students haven't logged in lately or seem academically adrift.

And so, on a recent afternoon, he's at a computer on campus, crafting an email blast to students on his class roster who have a red frowny-face circle next to their name.

"Dear so and so. I don't like to say 'dear.' So I'm gonna put instead YO!" he says as he taps away at his keyboard.

He wants to reach out to them before it's too late. The emails will land in inboxes looking personalized to each student. And they are — sort of.

"Your ole professor Phillips here," he continues writing.

He opts for the more personalized "I care" tack over the vaguely creepy, "I'm watching your online activity" approach.

"I wanted to know if there's anything additional I could do to assist you in the course," he writes to half a dozen red-flagged students.

These introductory or "gateway" courses are hugely important for Valencia, as they are at many other community colleges. National and Valencia figures show that if students withdraw from or fail even one of their first five course attempts, their chance of graduating is cut in half. Fail or withdraw from two classes, and those chances are cut in half again.

But fast, early interventions might save them. "We want to help them get there as much as possible," Philips says, "because, yeah, the data is not good if they don't do well on the first try."

To help intervene, he's using software and a learning platform called Illume and an app called Inspire. Both are from Civitas.

Here's how it works: The system vacuums up as much data as possible on things such as student participation online in assignments and discussion boards, how often and how long a student logs in to the school's learning platform, and data on a student's academic background. It's refreshed four times a day to create a student "engagement score," among other data points.

"It's unique because it's not just grade data or did they reply to something. It's a more holistic view of the student," says Valencia mathematics professor Brian Macon. "I don't think I'd want to teach without it now that I've had it for a year."

Of course, professors have long had the ability to send a pep talk email or try to motivate and engage students with the right words after class or during office hours.

But the difference now, says Kurt Ewen, a senior administrator at Valencia, is that real-time feedback allows teachers to more quickly flag a student who's struggling and, perhaps, better pinpoint what types of interventions might be needed.

"Part of what this tool allows for us to do is to shorten the amount of time that we have to wait to see where a student is," Ewen says. "Now they can do it in a way that they don't have to wait until after midterm."

The tools also can gather and analyze data from college or university learning-management and student-data systems, among other streams. That, for example, can tell a school exactly what its GPA tipping-point for student success or failure is, or whether grades are more predictive of completion than SAT or ACT scores.

"We help the institutions, especially the faculty and advisers, turn the lights on bright so they can see what's happening and use the best of their professional skill and will to change the trajectory for that student," Milliron says. The four-year-old company recently got new investor pledges of some $60 million to expand its products.

At Valencia and other schools using predictive analytics, a semi-personalized email is just one intervention. Others might include one-on-one or group tutoring, peer-to-peer mentoring, meetings with an adviser or professor or other steps.

The job of intervening is made that much more challenging by the fact that Valencia serves more than 70,000 students in two sprawling central Florida counties. Many students have to balance school with full-time jobs, kids and other responsibilities.

Here's another reason using data for early intervention matters: Studies show that nearly 80 percent of first-year community college students need at least some remedial English or math help.

But Florida was tip of the spear in a national backlash against what many saw as ineffectual remedial education. Students at public universities in Florida are no longer required to take remedial tests or classes. It's now their choice.

So now some introductory courses act as de facto remedial classes. Students in Phillips' composition class will have to show they can handle college-level work and craft cogent arguments within well-framed essays. "They have to have minimum competencies," Phillips says. "They have to do so much writing. They have to follow through on the tasks."

No one involved in Valencia's pilot project sees these new digital platforms as silver bullets. Civitas' Milliron readily concedes the best data systems in the world "will not fix a student who is not on purpose, who is not willing to engage and doesn't have the tenacity and motivation to succeed in higher education."

His company's tools do help, he believes, those students who might be really trying hard but just aren't connecting with the material or who are stuck and just don't know whom to talk to.

Civitas says its numbers from the dozens of schools it's working with have consistently shown a boost of 3 to 7 percent in course completion and "persistence," meaning continued enrollment.

And Civitas is betting that helping improve retention rates will prove profitable. As the company's chief executive, Charles Thornburgh, has said: "Approximately $100 billion in tuition and fees are spent each year [in the U.S.] on students who will not graduate. We believe there is an opportunity to improve that number by 5 to 10 percent, which is a multibillion-dollar opportunity."

As for Valencia, it's not yet clear whether the added data and interventions are helping boost student performance or graduation rates. "It is still early to determine the impact of the analytic tools," Ewen says.

Nonetheless, Valencia professors and administrators are cautiously optimistic.

"It has the potential to be innovative depending upon the people who use it," Ewen tells me. "We believe that predictive analytics has the potential to take us to a place we haven't been before and as an enhancement to the work we've been doing already."

For student Lofton, all the nudges helped push him to participate and thrive. He got an A in Phillips' class.

"We felt like the essays were, like, so moving and touching, we was like, 'Man, we gotta discuss this!' " he says. "The course was interesting. The extra credit [for online participation] to me was just a bonus." Students should give it their all, he says, especially if a professor such as Phillips "is going all out as far as trying to help the kids pass."