The post-test advantage of generative AI-supported instruction persists after controlling for pre-test performance (ANCOVA)¶

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 After controlling for pre-test scores, instructional group had a statistically significant effect on post-test geometric reasoning and proof performance, partial η2 = .377. → Davor I 2026

Evidence¶

Davor I 2026¶

Davor I (2026) Generative AI-supported instruction as a cognitive scaffold: effects on senior high school students' geometric reasoning and proof construction. Front. Educ. 11:1869961. https://doi.org/10.3389/feduc.2026.1869961

q3 · i3 · causal · r2

ANCOVA with post-test scores as dependent variable, instructional group as fixed factor, and pre-test scores as covariate in the same quasi-experimental study. The article states the "partial η2 = .377" indicates a large effect; the overall model explained approximately 50.3% of variance (R2 = .503).

"after controlling for pre-test scores, a statistically significant effect of instructional group on post-test performance was observed, F (1, 83) = 50.30, p < .001, partial η2 = .377."

Discussion¶

Learner Variables¶

• Geometric Reasoning — outcome: instruction changes it

Related Claims¶

• Senior high school students taught with generative AI-supported instruction scored significantly higher on geometric reasoning and proof construction than students taught conventionally — possibly the same claim (merge candidate)
• Both instructional groups improved significantly from pre-test to post-test, with larger gains in the generative AI-supported group — related
• Students in the generative AI-supported condition reported favorable perceptions of the instruction, with feedback for correcting mistakes rated highest — related
• Prompted AI-supported learning raises overall post-test scores above both slide-based and unprompted AI conditions — related
• Non-layered AI feedback produced significantly higher post-test performance than layered AI feedback after controlling for pre-test — related
• Students' initial knowledge level (midterm scores) did not significantly influence improvement in lab scores, indicating it did not confound the feedback-source difference — related
• AI-supported PBL students obtain significantly higher posttest computational thinking scores than traditional-lecture students when controlling pretest scores — related