Research

Making Peer Feedback Playable

Setup

To replicate the system used in this study for automating feedback classification and assigning card-based incentives, follow the steps below.

1. Make a copy of the cards and print them out (double-sided since every second page are the backs of the cards) on standard 8.5x11" paper or card stock.
2. Download the scripts that will be used in the following steps.
3. Collect Peer Feedback Using Moodle
   • Use the Workshop activity in Moodle to set up a peer review cycle.
   • Ensure that students are required to submit both their work and feedback for at least three peers.
4. Download feedback in bulk using the Moodle scraper.js script to extract all peer feedback comments into a CSV format.
5. Classify Feedback Using ChatGPT
   • Use few-shot prompting to classify each comment using the Code Review Taxonomy.
   • You can (and probably should) edit the taxonomy to fit your discipline (e.g., art critique, essay peer review, etc.).
6. Convert Classifications to Points
   • Each taxonomy category maps to a point value as described in the Methodology#Pre-Intervention Phase.
   • You can either get the AI to do this or create a formula in Excel.
7. Assign Cards Based on Points
   • Use card-assigner.js to translate points into action cards (e.g., 3 = 1 steal card).
   • Show the output of the script to your students when it's time to play so they know who gets which cards.

Game Rules

Poster

• Abstract
• Closing Statement
• Code Review Taxonomy
• Conceptual Framework
• Conclusion
• Consent Form
• Data Analysis
• Data Collection Procedure
• Discussion
• Game Intervention
• Intrinsic Motivation Inventory (IMI)
• Introduction
• Literature Review Outline
• Literature Review
• Methodology
• Plotly
• Pre-Test & Post-Test
• Problem Statement
• Few-Shot Prompt Design for LLM Classification
• Research Proposal
• Results
• Structuring Knowledge
• The Game
• Thesis