The Only Thing More Powerful Than LMS Data Is LMS Data Plus Multiple Sources
“Octopus” by oatsy40 is licensed under CC BY 2.0

A new report from Blackboard’s data science team led by Dr. John Whitmer finds that “data from multiple learning tools (…) improves the accuracy of predictions.” The report is the result of a joint research project between EdTech company and “eTextbooks” provider VitalSource and the University of Maryland, Baltimore County.

Post Pages - Post Inline - WIRIS

For a sample of UMBC students enrolled on Physics, Math, or Spanish courses, each partner provided different pieces of information. Blackboard provided LMS navigation activity data by students on the cloud-based Learn environment. VitalSource contributed “telemetry event” data on their activities. UMBC added student characterization data. The data was standardized and linked at the individual level using the IMS Global Caliper specification.

Among the relevant findings:

  • Early activity is a strong predictor of a student grade. It validates the “Waterfall” model in which ongoing activity early on leads to sustained activity, a process that could also be associated with engagement.
  • Sustained activity is correlated with a higher grade, which does not seem possible to salvage by overcompensating with high level of activity later on, nor by high activity early on that stalls at some point in the course.
  • Predictions for partial, quarterly performance results showed a higher level of confidence using multiple sources than a single one.
  • Incoming GPA remains the strongest predictor of engagement.

As the technological ecosystem of learning consolidates, companies begin to embrace strategic partnerships that allow them to deliver high quality on every front, rather than building competing products that might not be part of their core expertise. Common languages and standards, such as xAPI or in this case Caliper, bolster the industry’s ability to bridge solutions and approach the vision for a true learning ecosystem. For the end learner, an exciting landscape of opportunities where taking advantage of any solution, regardless of platform, medium, or source, can still account towards their personal learning goals as long as they are capable of communicating evidence to one another. Furthermore­, researchers and decision makers gain a more encompassing dataset about the learner’s behavior and performance, which opens up groundbreaking possibilities for better instructional and institutional design.

Download the research report, “Improving Student Risk Predictions: Assessing the Impact of Learning Data Sources” at■

eThink LogoThis Moodle Practice related post is made possible by: eThink Education, a Certified Moodle Partner that provides a fully-managed Moodle experience including implementation, integration, cloud-hosting, and management services. To learn more about eThink, click here.

Previous articleMoodle 3.5 Lets You Tag Individual Questions For Faster Search, Quiz Creation
Next articleWant To Kick Your Learning Plans Up A Notch? Paradiso Tells You How


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.