The star-to-sentiment labeling rule agrees with human review-level labels at Cohen's kappa 0.66 with 82% accuracy on 3,000 RateMyProfessor teacher reviews¶
Claim · All claims
Evidence · 1 study · designr2·q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly
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? Star-derived English sentiment labels are a reasonable but imperfect proxy, agreeing with EduRABSA human labels at kappa 0.66 and 82% accuracy, with the neutral class weakest. → Esteban U. Vega Barajas 2026
Evidence¶
Esteban U. Vega Barajas 2026¶
Esteban U. Vega Barajas. (2026). A Durability and Cross-Language Transfer Benchmark for a Validated Teaching-Feedback Classification Protocol. https://arxiv.org/abs/2607.11873 doi:10.48550/arxiv.2607.11873
q2 · i? · design · r2
Validation of the star-to-sentiment labeling rule against EduRABSA's independently annotated review-level human sentiment labels, on the 3,000 RateMyProfessor teacher reviews within EduRABSA that share the star-rating scale of the assembled corpus. The article reports "Cohen’s kappa 0.66, with 82% accuracy" and notes "The neutral class is the weakest, because human-neutral reviews scatter across star levels".
"review-level sentiment at Cohen’s kappa 0.66, with 82% accuracy. The neutral class is the weakest, because human-neutral reviews scatter across star levels."
Discussion¶
Related Claims¶
- An open-source Llama3 model reached substantial agreement with a human coder when sentiment-labeling pre-extracted student reflection quotes — related
- Human-LLM agreement on a complex multi-label codebook falls well below human-human agreement, while LLM-LLM agreement is comparable to human-human agreement — related
- An LLM-based AI Evaluator agrees with expert human raters on collaboration transcripts at a level similar to inter-expert agreement — related
- Sentiment classification transfers to English under the unchanged protocol, with a frozen multilingual encoder the strongest arm at wF1 0.73 and LLMs matching it only with few-shot prompting
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
star-label-rule-kappa-066-82-percent |
title |
The star-to-sentiment labeling rule agrees with human review-level labels at Cohen's kappa 0.66 with 82% accuracy on 3,000 RateMyProfessor teacher reviews |
description |
The star-to-sentiment labeling rule agrees with human review-level labels at Cohen's kappa 0.66 with 82% accuracy on 3,000 RateMyProfessor teacher reviews |
status |
review |
evidence_strength |
weak |
newest_evidence |
2026 |
recheck_by |
2029 |
generated |
by process:wiki-ingest on 2026-10-11 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — esteban-u-vega-barajas-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.