Linguistic complexity features show few clear progressions across ESO grades in the compiled science corpora¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2024 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? MultiAzterTest feature analysis of the corpora revealed that most indices do not increase systematically with ESO grade, suggesting limited benefit of these linguistic features for ARA classification. → Uçar 2024

Evidence¶

Uçar 2024¶

Uçar, S.-Ş., Aldabe, I., Aranberri, N., & Arruarte, A. (2024). Exploring Automatic Readability Assessment for Science Documents within a Multilingual Educational Context. Int J Artif Intell Educ. https://doi.org/10.1007/s40593-024-00393-2

q2 · i? · design · r2

Feature-trend analysis across the three corpora using MultiAzterTest (125, 141 and 163 features for Basque, Spanish and English). Word length in Basque increases with grade (6.32 to 6.96) and Spanish left embeddedness rises (2.74 to 4.12), but other indices run counter to the expected trend, e.g. rare adjectives are mostly present in ESO-1 documents.

"The linguistic analysis carried out revealed few patterns across grades. This suggests that the benefit of these particular linguistic features might be limited for the ARA models for classification."

Discussion¶

Related Claims¶

• 4-level science readability models struggle most with distinguishing adjacent grades 1 and 2 — related
• Fine-tuned transformer DL models outperform feature-based SVM models for science readability classification across all three languages and configurations — related
• Science readability models classify texts substantially more accurately in a 2-level setup than in the 4-level ESO-grade setup — related