Predicting scores in English argumentative essays by L1 Spanish and English authors: A broad exploration of writing quality indices
Predicting scores in English argumentative essays by L1 Spanish and English authors: A broad exploration of writing quality indices

Summary/Abstract: In recent years numerous studies have investigated how different writing features, such as lexical complexity and syntactic complexity, impact writing quality scores. However, most studies focus on only one set of features (e.g., lexical sophistication) to explain holistic writing quality scores. Thus, the purpose of the present paper is to use a broad set of features, covering lexical complexity, syntactic complexity, and cohesive properties, to explore which features were associated with and best predicted holistic scores of English essays among L1 Spanish and English authors. Findings from regression analyses suggest that lexical sophistication and phrasal complexity best predicted high scores in both the L1 Spanish and English groups. This finding supports previous research on lexical sophistication and syntactic/grammatical complexity (Biber et al., 2011; Kyle & Crossley, 2015). However, differences were found between L1 groups in the number of predictors, specific predictors, and amount of variance explained by each model. Furthermore, cohesion indices and multi-word units played a limited role in predicting scores in this data set. The present study can provide teachers and learners with insights into the combination of features that predict higher writing scores from a broad set of features, and thus help to focus teaching and autonomous studies.

• Details
• Contents