This study explores the potential of adaptive experimentation in conjunction with machine learning to enhance continuous improvement in education. Traditional teaching strategies may not have a clear pathway for rapid data usage to optimize student experience. Adaptive experiments analyze data from different educational strategies as they are deployed, using machine learning to identify the most promising approaches for improving student outcomes. The study demonstrates this with a comparison of traditional and adaptive experimentation in the context of self-explanation prompts in online homework for a CS1 course.
Publication date: 23 Oct 2023
Project Page: https://doi.org/XXXXXXX.XXXXXXX
Paper: https://arxiv.org/pdf/2310.12324
