ABSTRACT
The most widespread application of learning analytics is the early identification of students predicted to be at risk of failure or withdrawal. Predictive models can be built using historical data about the activity of previous students. One of the key assumptions of 'student success', 'early alert' or 'early warning' systems, as they are variously known, is that data collected about student engagement is a proxy for how well they are learning. This usually comes from use of the LMS and other web-based applications such as library systems. Combining data about what students are doing with data about who they are and what they have done in the past can give a stronger indication of their likely academic achievement than simply looking at their current activity. Probably the most frequently cited institutional deployment of learning analytics for early alert and student success is the Signals project at Purdue University in Indiana.
