A comparative study of text characteristics of CET-6, IELTS, and TOEFL reading passages based on computational tools

This study conducts a multidimensional analysis of reading texts from three major English proficiency tests--CET-6, IELTS, and TOEFL--using computational tools such as Coh-Metrix, VocabProfile, and WordSmith. Based on a corpus of 60 reading passages from each test, the study evaluates text characteristics across five dimensions: vocabulary, syntax, textbase, situation model, and readability. The results reveal significant variations across multiple indices: (a) CET-6 texts demonstrate lower lexical and syntactic complexity but weaker cohesion at the textbase and situation model levels, while TOEFL passages exhibit the highest levels of lexical and syntactic sophistication; (b) IELTS texts occupy an intermediate position, with greater variability in text length; (c) no significant differences were observed across the three tests in terms of L2 Readability, suggesting that these exams may be more comparable in this aspect than previously assumed.

The findings highlight the interplay between text characteristics and cognitive demands, offering empirical insights for both test development and language education. This study underscores the necessity of adopting a multidimensional approach to assess text complexity, ultimately assisting learners in optimizing test preparation and language acquisition.