LLM-enhanced feedback did not yield a statistically significant improvement in final grades compared with traditional AI feedback¶
Claim · All claims
Evidence · 1 study · causalr1·q2
Currency · newest study 2025 · checked against its sources 2026-10-28 (automated, 1 of 1 entry) · not reviewed by a person
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q2 i? Average final grades improved from 27.6/30 (AY 2022/23, schematic feedback) to 28.2/30 (AY 2023/24, combined feedback), but the difference was not statistically significant. → Letteri 2025
Evidence¶
Letteri 2025¶
Letteri, I. & Vittorini, P. (2025). Enhancing Student Feedback in Data Science Education: Harnessing the Power of AI-Generated Approaches. International Journal of Artificial Intelligence in Education. https://doi.org/10.1007/s40593-025-00492-8
q2 · quasi-experiment (two academic-year cohorts, not randomised) · i? · causal · r1
Comparison of final grades between two consecutive, non-randomised cohorts: AY 2022/23 (139 enrolled, schematic feedback from traditional AI) and AY 2023/24 (136 enrolled, combined feedback with LLMs), using independent-samples t-tests. Grades were collected only from students who attended the first exam session, so the analysed numbers are smaller than the enrolled cohorts (106 graded students in the 2023/24 cohort's exercise analysis; the analysed n for each year in the cohort comparison is given only in Table 4). The article reports the difference in average grade as not statistically significant.
"The comparison between the two A Ys 2022/23 and 2023/24 (when we introduced the feedback based on LLMs) shows an improvement in the average final grades (from 27.6/30 of the A Y 2022/23 to 28.2/30 of the A Y 2023/23), even if this difference is not statistically significant (see bottom of Table 4)."
Discussion¶
Related Claims¶
- Students who attempted more than 90 different exercises achieved significantly higher final grades than students who attempted fewer — related
- Students rated the usefulness of LLM-enhanced combined feedback significantly higher than their own prior expectations — related
- LLM-assisted self-study yields higher post-intervention exam performance than traditional resources among early-career knowledge workers — related
Page metadata
| Field | Value |
|---|---|
status |
review |
generated |
by process:wiki-ingest on 2026-10-28 |
trust |
unverified — no one has confirmed this page's content |
Frontmatter is the page's machine-readable record — see the Schema & Guide.