M4 Master Moodle Plugin Collecting and analyzing student feedback

Team

  • Lorenzo Battiston
  • Laura Langhauser
  • Niklas Lengert
  • Sao Chi Pham
  • Viet Anh Jimmy Tran

Supervision

  • Prof. Dr. Lucy Weggler
  • Prof. Dr. Gefei Zhang

Starting Point

The project was designed to be used in the upcoming semester by our supervisor Prof. Dr. Lucy Weggler. The creation of this plugin was based on her lacking a reliable difficulty label for her large number of math test questions. To generate difficulty labels before this project, she developed a pipeline that extracts key concepts from math questions using a language model, based on the assumption that more concepts indicate higher difficulty. However, this label shows no correlation with the statistical parameters that can be obtained from tasks in Moodle, such as success rates (correct entries).

Future Use by Supervisor

Over time, the plugins will fill this gap by generating a reliable difficulty label from the student reviews. In the future, the data collected by the plugins will be used to verify the current naive label and replace it with one derived from the student reviews. Ultimately, the data will be used to train a model to take over the complexity estimation.

General Use Cases

The plugin suite developed in this project is designed for the entire Moodle community, not just our supervisor. By enabling students to rate the difficulty of test questions, the plugins provide valuable insights into their learning experience. This helps teachers to better understand where their students are struggling or excelling. In courses with varying skill levels in particular, this feedback enables instructors to adapt their teaching focus based on actual student responses, for more effective teaching.