From left to right, Tan, Quintana, and Schulz presenting this work at the MIDAS Learning Analytics Symposium at the University of Michigan. (Academic Innovation)
Kyle Schulz, data scientist, Yuanru Tan, learning design and accessibility fellow, and Rebecca Quintana, learning experience designer, won the Academic Innovation at Michigan (AIM) Analytics Massive Open Online Course (MOOC) Dropout Prediction Challenge.
The team members were given a random subset of learner data from the first four weeks of courses and the challenge was to predict a probability of persistence for each learner. The two goals from this team were to understand what can be learned about MOOC data and course design evolution, and to develop a model to discover which learner behaviors have significant effects on course persistence rates.
Tan and Quintana compared MOOCs from 2008, 2012 and 2016 to learn how the courses have evolved since their first emergence in 2008. The main differences were the increased video production styles and the shortened commitment required by the learners.
