GenAI training attenuates the predictive effect of prior GenAI experience on revision performance¶

Claim · All claims
Evidence · 1 study · causal r2 · q3
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? Only the Group × GenAI experience interaction was statistically significant (β = -3.445, p = 0.034); the Group × SRL and Group × Pretest CSEM interactions were not. → Huang 2026

Evidence¶

Huang 2026¶

Huang, B., Xie, L., Liu, Y., Liu, X., Tang, H., Feng, X., Qiao, C., Guo, Q., Wu, C., & Bao, L. (2026). From answer machine to collaborative partner: impact of structured generative AI training on scientific problem solving in STEM education. International Journal of STEM Education. https://doi.org/10.1186/s40594-026-00652-9

q3 · i? · causal · r2

Robust linear interaction model (Magee's pseudo R² = 0.605) on RCT data. The significant "Group × GenAI experience interaction" (β = -3.445, p = 0.034) indicates training attenuated the prior-experience effect; the interaction plot showed a flatter slope for the experimental group (β = 0.015 vs. 3.460).

"only the Group × GenAI experience interaction was statistically significant ( β = -3.445, 95% CI = [-7.137, -0.264], p = 0.034). The Group × SRL skills interaction ( β = 1.161, 95% CI = [-3.884, 5.623], p = 0.681) and Group × Pretest CSEM score interaction ( β = -1.476, 95% CI = [-4.535, 2.812], p = 0.547) were not statistically significant"

Discussion¶

Related Claims¶

• Structured GenAI scaffolding training raises sophomore STEM students' posttest CSEM electromagnetism scores over problem-solving training (d = 0.81) — related
• SRL skills and prior knowledge predict revision performance, while SRL predicts higher and prior knowledge lower cognitive load — related
• GenAI dependence weakens the positive association between GenAI use and both student outcomes, attenuating it to statistical non-significance at high dependence — a broader claim this one bears on
• AI literacy moderates the pathways from perceived AI-adaptive environments to academic emotions, attenuating negative emotional effects among students with higher AI literacy — a broader claim this one bears on
• Transformative feedback attenuated the positive association between joy and regulatory behavior, and adaptive feedback moderation was otherwise limited, not supporting H2 — related
• GenAI training leaves total cognitive load comparable while marginally raising intrinsic cognitive load — related