Increasingly, institutions are building learner record stores to more fully describe student interaction with educational technologies. However, little research has assessed how much these increases in data actually move the needle in predicting student achievement. To find out, Blackboard, VitalSource, and UMBC partnered on a research project to explore both vendors’ Caliper and LTI-enabled tools to mine UMBC students’ use of an LMS and e-textbooks. Among other things, we found the combined model was very accurate in predicting C or better final grades by week 4 of the semester. We will also discuss how these findings support other ongoing research.
- Combine UMBC learner data from Blackboard and VitalSource to predict student outcomes with greater accuracy, earlier in the semester
- Identify behavioral and demographic factors most predictive of student success
- How does faculty adoption/integration of EdTech or overall course design, affect those relationships?
- Understand the effectiveness of IMS Caliper as data protocol to answer student success research questions
“If your data isn’t actionable, what’s the point?” – John Fritz
- Combine Blackboard, VitalSource, and SIS Data with IMS Caliper.
- At week 4, evaluate relationships between activity and course grade.
- Data: 986 total students in 5 subjects/7 sections, Fall 2017 term
- On using data… getting the right message to the right students at the right time. – John Fritz
- Students who didn’t take the syllabus quiz were 4 times more likely to earn DFW.
- Encourage faculty to require and use adaptive release for a syllabus quiz.
- Students who register late (under 40 days) are much more likely not to persist.
- Students that use math tutoring reduce their changes of failing between 10 and 20%.
- Presentation: umbc.box.com/eli19