The seeds of an effective early warning system exist in course management systems that deliver Web-based course content and online courses. These systems collect an enormous amount of data about student activities, including: the number of times they log into a course; what they're doing inside the course (reading, taking a quiz, posting an argument on a discussion board, turning in an assignment); and how much time they are investing in it.

Purdue University has analyzed online learning management system data from 2,300 courses and has identified the strongest indicators of student failure. In response to these activity patterns, the university has developed automated triggers that successfully intervene with faculty and students long before they would otherwise sense trouble. The beauty of this framework is that all of the tools for analysis and intervention are in place; an institution needs simply to harness them.

This ground-breaking work is part of a new academic practice called academic analytics, in which learning activity data and grades are captured and analyzed to support academic assessment, improvement and accreditation. Several major institutions are embarking on academic analytics projects, and with the further evolution of e-learning and technology, the implications promise to be far-reaching.