LOOM^TM

It’s no secret that analytics are taking the academic world by storm.  In 2009 Purdue University launched Signals, a data mining and early alert system that consistently tells students how they’re doing, which captured national attention.  EDUCAUSE has been taking a detailed look at how universities can access and employ online data through the use of academic analytics since at least 2005.  And earlier this year, George Siemens organized the first conference of its kind dedicated solely to learning analytics, an event that is to be repeated early next year in Vancouver.

The battle to gain access to the hordes of data collected by higher education is one that has been waged since the 1980s.  Since then, hundreds of millions of dollars have been spent on multiple technologies aiming to do just that—enterprise resource planning (ERP) systems, data warehouses and data marts, to name just a few.  Much has changed in those last few decades, including the widespread adoption of the learning management system (LMS), which alone means that the amount and types of data gathered has increased substantially.  And this trend is not unique to colleges and universities: in the last three years alone, high school students have become three times more likely to have access to online learning.

So it’s clear that institutions should look at the use of academic analytics as a beneficial, even imperative, move forward.  And many are, as the speed at which they’re gaining traction and even being used in major policy decisions shows.  But what exactly are we talking about here?   What exactly are learning analytics?

Siemens recently wrote that after much debate, the team that was organizing the first annual learning analytics conference came to the conclusion  that “learning analytics is the measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimising learning and the environments in which it occurs”.  Undeniably broad, that definition hints at both the five steps of analytics, as outlined by David Campbell and Diana Oblinger, and the many purposes for which those steps might be implemented.

Those five steps are: capture, report, predict, act and refine.  John Fritz of UMBC, in a presentation that was part of an open learning analytics course, addressed a similar model, this time in stages, before asking, “Where are you at?”  His version was: extraction and reporting, analysis and monitoring, “what-if” scenarios, predictive modeling and simulation, and automatic triggers and alerts.

Where are you at, indeed.  Fritz follows up with a chart to answer that (or at least to show how it was answered in 2005, when the data was collected).  A whopping 69.9% had only reached Stage 1.  I wonder how those numbers have changed, six years on, in an atmosphere where analytics are moving closer to the forefront at the administrative level.  My guess, though, would be that they haven’t moved on by much—although Signals offers a basic level of intervention, it is still imposed primarily at the lowly course level, and institutions on the whole have largely failed to adopt any broader means by which to leverage such analytics as a resource for commercial and academic purposes.

This is our goal for Weaver, our own analytics product: that it addresses all five steps, or stages, and makes the information collected by the LMS not only available but also a driver of success for all its users, from administrators to students.  Learning analytics have the potential to be powerful in any hands, but only if a system is built for that purpose.

Harriet May hmay@loomlearning.com