Adaptive scaffolding policies (BKT and DRL) significantly improve posttest performance over a non-adaptive control in a logic ITS¶

Claim · All claims
Evidence · 1 study · causal r2 · q3
Currency · newest study 2026 · checked against its sources 2026-10-17 (automated, 1 of 1 entry) · not reviewed by a person

Subclaims¶

q3 i? Both adaptive conditions (BKT and DRL) significantly outperformed the Control condition on posttest score, with comparable effect sizes. → Tithi 2026

Evidence¶

Tithi 2026¶

Tithi, S.D., Alam, N., Yasir, T., Shi, Y., Tian, X., Chi, M., Barnes, T. (2026). Adaptive Scaffolding for Cognitive Engagement in an Intelligent Tutoring System. arXiv. https://arxiv.org/abs/2602.07308 doi:10.48550/arxiv.2602.07308

q3 · i? · causal · r2

Controlled between-subjects classroom study (N=113 undergraduates in a discrete math course) comparing Control, BKT, and DRL scaffolding conditions with stratified random assignment after pretest. A mixed-effect regression with problem ID as random intercept found a significant condition-posttest association (p <0.001); "both adaptive conditions significantly outperformed theControl" with A=.58 for BKT and A=.58 (p=.002) for DRL.

"As shown in Table 2, both adaptive conditions significantly outperformed theControl(mean = 65.7) on posttest score. TheBKTgroup (mean= 72.3) showed significant improvement overControl(A=.58, 95% CI [.52, .63],p=.005)"

Studies bearing on this claim¶

Does choosing practice or scaffolding by BKT's mastery estimate improve learning or reduce the practice needed, compared with fixed or random sequencing, and for whom? This table gathers the studies behind this claim and the 8 narrower claims that bear on it: 3 distinct studies (2 causal, 1 review), q2–q3, 0 of 3 with an effect size coded. One row per study, codes as each entry records them; no effect is averaged and no verdict is drawn. Built by scripts/synthesize_claim.py from eval/synthesis/; a row that looks wrong is fixed on the claim it comes from.

Discussion¶

Related Claims¶

• DRL scaffolding significantly benefits high prior knowledge students while BKT does not, and condition-by-prior-knowledge interactions are not significant — a narrower finding that bears on this claim
• BKT scaffolding significantly benefits low prior knowledge students, who outperformed low prior knowledge controls — a narrower finding that bears on this claim
• No significant differences in posttest rule accuracy across scaffolding conditions — related
• BKT students completed posttest problems significantly faster than Control students, with DRL showing marginal advantages in time and solution optimality — related
• The DRL policy produced a significantly different scaffolding distribution than Control and BKT, favoring Guided examples (60%) and avoiding Buggy examples (4%) — related
• All training policies narrowed the prior-knowledge achievement gap from pretest to posttest, with BKT achieving the largest reduction (77.1%) — related
• Intelligent tutoring systems improve achievement over non-individualised instruction — a broader claim this one bears on
• Students receiving Adaptive proactive hints based on HelpNeed predictions achieve significantly higher posttest optimality than Control students — related
• Structured GenAI scaffolding training raises sophomore STEM students' posttest CSEM electromagnetism scores over problem-solving training (d = 0.81) — related
• Adaptive task difficulty in an ITS did not improve second-graders' mathematics posttest performance compared with an identical non-adaptive ITS — reports the opposite
• Teacher-judged logical reasoning combined with prior knowledge to predict learning only under the control condition; scaffolding almost fully diminished the disadvantage of lower logical reasoning — related
• Scaffolded third-graders outperformed control students on reasoning based on irrelevant evidence after instruction — a narrower finding that bears on this claim
• Scaffolding mitigated the limiting effect of low prior knowledge: prior knowledge predicted posttest irrelevant-evidence reasoning only under the control condition — related