Assessing the writing quality of English research articles based on absolute and relative measures of syntactic complexity

Abstract

Prediction of writing quality based on various linguistic features has greatly supported writing assessment in academic contexts. Research article (RA) is an important academic genre widely taught and assessed in EAP (English for Academic Purposes) writing courses, but there is little research exploring the assessment of RA writing quality on the basis of linguistic features, which may have the potential to assist teachers’ evaluation of students’ RA drafts. The present study aims to explore whether features of syntactic complexity can effectively predict RA writing quality. To represent syntactic complexity as a multidimensional construct, we employed traditional absolute measures and the newly proposed relative measures related to the use of verb argument constructions (VACs). Robust predictors of writing quality were identified at both the whole RA level and the part-genre level. It is found that VAC-based measures are more useful for indexing the writing quality of RAs at whole text level, while absolute measures have a stronger predictive power at part-genre level. Our research also shows that the relationships between writing quality and syntactic complexity are subject to part-genre variation. Implications for language teaching and assessment in the EAP context are discussed.