Multi-Source Emotion Annotation in Children’s Language: When LLM Consensus Diverges from Human Judgment

Farida Said, Jeanne Villaneau


Abstract
Automated emotion annotation increasingly relies on inter-LLM agreement as a proxy for label quality. We test this assumption on 2,106 clause-level segments from interviews with French-speaking children (ages 6-11) about parental roles, a setting where affect is often implicit rather than lexically explicit. Using a 500-segment expert gold standard, we show that internal consensus can be seriously misleading: Dawid-Skene, a probabilistic label aggregation method, estimates GPT-5.2 valence accuracy at 90.7%, whereas evaluation against human gold yields 71.0%, revealing substantial overestimation driven by shared neutralization bias. Conversely, Dawid-Skene underestimates Claude Sonnet 4, reversing model ranking. Majority Vote, Dawid-Skene, and MACE produce near-identical consensus labels, suggesting that the main source of error lies in shared annotator bias rather than in the aggregation rule itself. We release the expert gold subset and the probabilistic corpus to support future work. Our results show that high inter-LLM agreement cannot replace external human validation for affect annotation.
Anthology ID:
2026.cas-1.11
Volume:
Proceedings of Computational Affective Science (CAS) @ LREC 2026
Month:
May
Year:
2026
Address:
Palma, Mallorca (Spain)
Editors:
Christopher Bagdon, Krishnapriya Vishnubhotla, Kristen A. Lindquist, Lyle Ungar, Roman Klinger, Saif M. Mohammad
Venues:
CAS | WS
SIG:
Publisher:
ELRA Language Resources Association (ELRA)
Note:
Pages:
125–135
Language:
External URL:
https://lrec.elra.info/lrec2026-ws-cas-11
DOI:
10.63317/39tk69v6ca4p
Bibkey:
Cite (ACL):
Farida Said and Jeanne Villaneau. 2026. Multi-Source Emotion Annotation in Children’s Language: When LLM Consensus Diverges from Human Judgment. In Proceedings of Computational Affective Science (CAS) @ LREC 2026, pages 125–135, Palma, Mallorca (Spain). ELRA Language Resources Association (ELRA).
Cite (Informal):
Multi-Source Emotion Annotation in Children’s Language: When LLM Consensus Diverges from Human Judgment (Said & Villaneau, CAS 2026)
Copy Citation: