Improvement on Hint and Explanation Crowdsourcing Method for an Online Learning Platform Public

Downloadable Content

open in viewer

Crowdsourcing has been used in many successful online applications such as Wikipedia and Stack Overflow. In the field of educational research, many educational platforms, such as edX, recently implemented features that improve learning by taking advantage of crowdsourcing, such as peer grading. In my previous work, I implemented a crowdsourcing feature called "TeacherASSIST" inside the ASSISTments online learning platform. TeacherASSIST allowed teachers to create hints and explanations, which would be given to students on-demand while they were working on their assignments. In that work, I used a simple aggregation method that automatically distributed such hints and explanations created by expert-selected teachers to all students inside ASSISTments. By using randomized controlled trials, I found that crowdsourced hints and explanations improved student learning with statistical significance. In this work, I improved TeacherASSIST by adding more organic approaches to aggregate hints and explanations using trusted teachers and hint/explanation ratings. The first approach was allowing teachers to designate other teachers as 'trusted.' Their students would then be able to receive hints and explanations created by the trusted teachers. The second approach was constructing global teacher scores based on teachers rating each other's hints and explanations. The aggregation based on ranking allowed the pool of globally trusted teachers to grow over time even for teachers who do not actively search for more trusted teachers. I then ran a randomized controlled trial and used linear regression models to evaluate the effectiveness of the aggregation method based on ranking. In addition, I also designed and generated prototypes of reports that would allow teachers to see how much their student supports have helped their students and, for starred teachers, other students inside ASSISTments.

  • etd-21711
Defense date
  • 2021
Date created
  • 2021-05-04
Resource type
Rights statement


In Collection:


Permanent link to this page: