Hint Factory MDP-generated hints gave correct next-step hints over 80% of the time and a pilot showed students solved more logic proofs with hints¶

Claim · All claims
Evidence · 1 study · review r2 · q2
Currency · newest study 2013 · checked against its sources 2026-10-20 (automated, 1 of 1 entry) · not reviewed by a person

Subclaims¶

q2 i? As reported in this review of Barnes and Stamper (2008), extracted MDPs from historical student data provided correct next-step hints over 80% of the time, and augmenting a propositional logic tutor with the Hint Factory increased the number of proof problems students solved. → Kenneth R. Koedinger 2013

Evidence¶

Kenneth R. Koedinger 2013¶

Koedinger, K. R., Brunskill, E., Baker, R. S. J. d., McLaughlin, E. A., & Stamper, J. (2013). New potentials for data-driven intelligent tutoring system development and optimization. AI Magazine, 34(3), 27–41. doi:10.1609/aimag.v34i3.2484

q2 · i? · review · r2

Secondary report: this review article describes Barnes and Stamper (2008), whose feasibility demonstration on historical student attempt data found that extracted MDPs "could provide correct next-step hints towards the problem solution over 80% of the time". In a pilot study, also by Barnes and Stamper, augmenting a logic tutor with the Hint Factory let students solve more proof problems when hints were included. This article gives no figures for the pilot.

"Barnes and Stamper (2008) demonstrated the feasibility of this approach on historical data, showing that extracted MDPs with the proposed hint-generating functions could provide correct next-step hints towards the problem solution over 80% of the time."

Discussion¶

Related Claims¶

• Mindful-condition students completed fewer problem-solving steps and requested fewer hints per step than Cognitive-condition students — related
• Error rates per completed problem-solving step did not significantly differ between conditions — related
• In peg-solitaire simulations with known true abilities, the RLMM attains lower RMSE of log βj than the MDP-MM on all four original boards — related
• The RLMM recovers ability parameters on peg-solitaire boards with up to more than 23 million states, where tabular MDP-MM estimation is impractical — related
• RLMM runtime stays within approximately 15–20 seconds across the four peg-solitaire boards while MDP-MM runtime grows with task complexity, reaching a 19.1× speedup on the diamond board — related
• Influence-function diagnostics identify steps 3–8 of the Big-L peg-solitaire board as the most consequential region, with the lowest-β participants showing substantially larger absolute influence values there — related
• K-12 learners preferred incremental hints and multiple attempts over immediately revealed solutions — related