- OLI Feature Analysis -
Duration: May 2020 - Nov 2020
Keywords: Learn by doing; Interactive Online Course; Learning outcome prediction
Description

This is a RA research project in LearnLab, CMU, advised by Prof. Kenneth Koedinger and Dr. Paulo Carvalho.

a) I used educational data mining techniques to analyze practice features in a graduate-level course about E-Learning Design & Methods and b) I provided practical redesign suggestions to refine and redesign the current course.

Check out platform details at >> Open Learning Initiative Website

My Role
As a data-mining researcher, I ...

Cleaned two years of log data in the E-learning course; performed mixed-effect models to detect general doer effect; did learning curve analysis to detect low-performed KCs with actionable refinement recommendations; did exploratory analyses for quiz behavior.

Synthesised the summer work, did literature review and finished paper writing with my co-authors.

Provided feasible redesign suggestions to my co-workers.
For example, (1) create new review questions for KCs that are not covered in current review practices. (2) add instructional messages, feedback and hint messages to inner-system practices (Right Figure). (3) display prompts to encourage students to finish review practices.

Related Publications and Deliverables

[1] (Under Review) Xinying Hou, Paulo F Carvalho, Kenneth R Koedinger (Under Review). Drinking Our Own Champagne: Analyzing the Impact of Learning-by-doing Resources in an E-learning Course. Submitted to the International Conference on Learning Analytics & Knowledge (LAK’ 21).

Refined practice feedback feature (Left: Correct; Right: Wrong)