GenAI-enhanced instruction is associated with substantially greater legal reasoning gains than traditional instruction in property law, though the effect size warrants cautious interpretation¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r2 · q2 · i3 large
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person
Design limits · Haifang Liu 2026: the intervention came bundled with another change; groups were assigned but students were analysed as independent · Haifang Liu 2026 (2): groups were assigned but students were analysed as independent (read from the study by two models, 2026-10-29)

Subclaims¶

q2 i3 The group-by-time interaction was statistically significant (F(1,178) = 52.73, p < .001), with the experimental group showing faster growth across the semester. → Haifang Liu 2026 q2 i3 ANCOVA controlling for pre-test scores confirmed the pattern (F(1,177) = 61.24, p < .001, partial η2 = .26). → Haifang Liu 2026 (2)

Evidence¶

Haifang Liu 2026¶

Haifang Liu. (2026). AI-enhanced legal pedagogy: a case study of generative AI integration in property law courses. Education Tech Research Dev. https://doi.org/10.1007/s11423-026-10698-9

q2 · i3 · causal · r2

Group-by-time interaction on legal reasoning scores across the 16-week semester, shown in Fig. 3 as a sustained upward trajectory for the experimental group; the article prints partial η2 = .26 for the pre-test-adjusted analysis of the same outcome, a large effect by the η² thresholds.

"the experimental group generated faster growth compared to the control group, as demonstrated by the statistically significant group-by-time interaction effect, F(1,178) = 52.73, p < .001"

Haifang Liu 2026 (2)¶

Haifang Liu. (2026). AI-enhanced legal pedagogy: a case study of generative AI integration in property law courses. Education Tech Research Dev. https://doi.org/10.1007/s11423-026-10698-9

q2 · i3 · causal · r2

ANCOVA on post-test legal reasoning with pre-test as covariate in the same quasi-experiment; the article prints partial η2 = .26, which exceeds the .14 threshold for a large effect, while noting ANCOVA "controls for initial ability levels only".

"ANCOVA controlling for pre-test scores confirmed this pattern, F(1,177) = 61.24, p < .001, partial η 2 = .26, indicating that the association between GenAI-enhanced instruction and learning gains remained substantial after accounting for pre-test differences"

Discussion¶

Related Claims¶

• The GenAI-integrated group outperformed controls across course assessment components, with Cohen's d ranging from 0.65 to 0.88, and a higher proportion earning B+ or higher — related
• Students report that GenAI's immediate feedback enabled iterative learning and low-stakes practice, the dominant qualitative themes of the intervention — related
• A subset of students showed over-reliance and algorithmic conformity patterns, consulting GenAI before independent analysis or revising to match AI output — related
• Case analysis assistance and Socratic dialogue simulation dominated GenAI usage, and dialogue-based interaction patterns were associated with better outcomes than passive use — related
• Instructors reported role transformation toward facilitation, shifted assessment toward performance-based formats, and identified professional development and integrity needs — related
• Perceived usefulness, attitude, and behavioral intention form a significant TAM pathway predicting GenAI adoption among law students — related