Terracotta: A tool for conducting experimental research on student learning.

Area of Science:

• Educational Psychology
• Cognitive Science
• Learning Sciences

Background:

• Identifying effective teaching practices requires rigorous experimental evidence.
• Conducting in-situ educational experiments is challenging due to cost and complexity.
• Existing methods for educational research in classrooms are often limited.

Purpose of the Study:

• To introduce Terracotta, an open-source web application for conducting randomized controlled trials in online learning environments.
• To demonstrate Terracotta's capability in automating experimental procedures for educational research.
• To test the efficacy of retrieval practice as a teaching intervention using Terracotta.

Main Methods:

• Developed Terracotta, a web application integrating with learning management systems for automated randomization, consent, and data export.
• Conducted a preregistered replication study manipulating online review assignments.
• Students alternated weekly between retrieval practice (quizzes) and restudy (reading answers).

Main Results:

• Retrieval practice significantly improved student performance on subsequent exams compared to restudy.
• Terracotta successfully facilitated the experimental manipulation of learning activities.
• The study replicated previous findings on the benefits of retrieval practice.

Conclusions:

• Terracotta provides a viable and efficient platform for conducting educational experiments in online settings.
• Experimental manipulation of learning activities, such as retrieval practice, yields measurable improvements in student outcomes.
• The tool supports rigorous research to identify causal benefits of teaching practices.