Focus on KM in higher education:
Learning analytics efforts apply business intelligence to student retention.

At first, the predictive data was only given to instructors so that they could consider it along with their own observations to help decide whether they wanted to intervene with students via e-mails and phone calls. But recently, PACE has added a student-facing element that gives them a green, yellow or red light as soon as they log in, reminding them of how they are doing in the course, according to the analytic measures.

There hasn't been any negative feedback from students yet, according to Cottam. "They seem to pay attention to the stop light signals, and many are contacting their instructors when they have a red or yellow, which is a good thing. Now that we have the system in place, we can experiment with interventions to see which ones have the biggest impact," she says. 

Looking at the bigger picture

American Public University System (APUS) and Rio Salado College are working with four other members of the WICHE Cooperative for Educational Technologies (WCET) on the Predictive Analytics Reporting Framework (PAR) project, with a grant from the Bill & Melinda Gates Foundation.

PAR's goal is to identify variables that influence student retention and progression. The data will be used to explore patterns that emerge when the data from different institutions are analyzed as a single, unified sample. All of these schools had been working on retention models independently.

Phil Ice, VP of research and development at APUS, says, "We got together and decided to create a data warehouse to look across institutions to study 33 variables that relate to student success and retention." The project has 640,000 student and 3 million federated course-level records from six institutions.

"The first year we have had to work on smoothing the data and understanding the definitions each school uses and harmonizing them," says Sebastian Diaz, associate professor in the Department of Technology, Learning & Culture at West Virginia University and senior statistician for the PAR Framework project. "In the initial phase, the analysis will all be centralized, as we look for meaningful ways to view what we are seeing. But later we will provide the campuses de-identified data so they can run their own analyses." The group also hopes to have findings it can report back to the higher education market at large.

In addition, they are working on some KM models to help them understand the best ways to work together across institutions. "We had the e-mail traffic recorded on our portal and did a social network analysis of it," Diaz says. "We created hub-and-spoke diagrams to understand how the communication was happening internally so we could learn how to communicate more effectively."