Investigating the scopes of textual metrics for learner level discrimination and learner analytics - Archive ouverte HAL Accéder directement au contenu
Chapitre D'ouvrage Année : 2022

Investigating the scopes of textual metrics for learner level discrimination and learner analytics

Résumé

This chapter investigates the linguistic interpretation of complexity metrics in L2 proficiency assessment. By analysing 84 formulas of metrics linked to lexical diversity, readability and syntactic complexity, we identify a taxonomy of their underlying linguistic scopes. These metrics are classified according to text, sentence, clause, phrase and word scopes with attributes and methods. Homogeneity of scopes was evaluated by applying a mixed clustering PCA approach to metrics computed for 329 L2 texts. Discriminative power was evaluated with a random forest approach on the same dataset including the CEFR levels. Results show that metrics are diversely clustered but they also suggest in-cluster homogeneity. The CEFR classification shows mixed results suggesting that diversity, repetition and size in word and text scopes are significant.
Fichier non déposé

Dates et versions

hal-03670165 , version 1 (17-05-2022)

Identifiants

  • HAL Id : hal-03670165 , version 1

Citer

Thomas Gaillat. Investigating the scopes of textual metrics for learner level discrimination and learner analytics. Agnieszka Leńko-Szymańska; Sandra Götz-Lehmann. Complexity, Accuracy and Fluency in Learner Corpus Research, John Benjamins, In press. ⟨hal-03670165⟩
19 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More