ChatGPT-assisted students scored substantially higher on the coding assessment than No AI students (89% vs 69%)¶

Claim · All claims
Evidence · 1 study · causal r2 · q3 · i3 large
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i3 In a between-subjects experiment, AI-Assisted students achieved higher coding assessment scores than No AI students (M=0.89 vs M=0.69, Hedges' g = -1.24). → Bergh 2026

Evidence¶

Bergh 2026¶

Bergh, C., Tag, B., Vassar, A., & Renzella, J. (2026). Your Programming Students'Cognition with ChatGPT: Higher Performance, Lower Retention, and Reduced Ownership. https://arxiv.org/abs/2609.21194 doi:10.48550/arxiv.2609.21194

q3 · i3 · causal · r2

Controlled between-subjects experiment with 55 retained undergraduate CS students completing three C programming tasks graded by automated unit tests in EdStem. The AI-Assisted group (ChatGPT-4.5) outscored the No AI group (web search) with a large printed effect size, Hedges' g = -1.24, surviving Holm correction.

"The AI-Assisted group scored substantially higher on the coding assessment (AI-Assisted𝑀= 0.89,𝑆𝐷= 0.12) than the No AI group (𝑀= 0.69,𝑆𝐷= 0.19), Welch’s𝑡( 42.0)=− 4.57,𝑝<. 001, Hedges’ 𝑔=− 1.24, 𝑝Holm <. 001"

Discussion¶

Related Claims¶

• ChatGPT-assisted students reported substantially lower ownership of their submitted code across self-attributed ownership, psychological ownership, and authorship attribution — related
• The retention gap persisted at 48 hours with no differential forgetting between groups (39% vs 52%) — related
• No AI participants spent roughly 1.5 times as long on the assessment as AI-Assisted participants, descriptively reported as a marker of offloading — related
• A meta-analysis of 69 studies reports generative AI improves academic performance with g = 0.7, but the article argues this likely reflects immediate task success rather than learning — related