Use predicted hint-taking likelihood and hint effects to adaptively decide whether to withhold or provide hints¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The article proposes that a learning environment use a hint-taking prediction model to make adaptive decisions on showing hints, because learners misuse on-demand hints. It suggests the environment "can proactively suggest hints to students who are stuck with a concept and have a low likelihood of taking a hint themselves", while withholding hints from students prone to abusing them without attempting the problem.

Design Implications¶

Context¶

Requirements¶

• A trained hint-taking prediction model and knowledge tracing model over the learner's response and hint sequence.

Constraints¶

• The article notes offering hints indiscriminately can lead to poor learning outcomes, and that integrating the two task predictions into long-term outcome-optimizing strategies is left as future work.

Target Learners¶

• Students in interactive e-learning environments with on-demand hints

Target Learning Goals¶

• Regulating help-seeking behavior
• Improving learning outcomes through appropriately timed hints

Affordances¶

• Colearn Multi Task Hint Knowledge Model

Related Strategies¶

• Hint Design
• Help Seeking
• Encouraging Students to Ask for Help
• Encouraging Help-Seeking Behavior
• Use early student-model predictions to identify at-risk students and adapt pedagogical strategy during tutoring
• Use lightweight transparency cues about AI fallibility to encourage more cautious, help-seeking interaction strategies
• Design AI tutoring behaviors that withhold solutions, prompt recall, and ask critical questions

Examples¶

-

Key Sources¶

• Ritwick Chaudhry, Harvineet Singh, Pradeep Dogga, and Shiv Kumar Saini. (2017). Modeling Hint-Taking Behavior and Knowledge State of Students with Multi-Task Learning. Proceedings of the 11th International Conference on Educational Data Mining. https://educationaldatamining.org/