The relationship between academic vocabulary coverage and scores on a standardized English proficiency test

• The AWL coverage in CanTEST was below the level of coverage in academic texts.
• There was variation in AWL coverage within and across the test versions.
• The AWL coverage had no correlation with the overall reading and listening scores.
• There was a small, negative correlation between AWL coverage and the cloze scores.

This study examined the occurrence of items from the Academic Word List (AWL) (Coxhead, 2000) in reading, listening and cloze texts in 12 versions of an English proficiency test used for admission purposes at Canadian universities. The relationship between the coverage of AWL in the texts and the corresponding scores of 6380 test takers was also analyzed. The results indicated that AWL coverage in the passages was consistently present and substantial although below the established level of coverage found in academic texts. Furthermore, AWL coverage had no correlation with the overall reading and listening comprehension scores, and there was a small, negative correlation with cloze test scores. The findings are discussed in light of lexical specifications of language proficiency tests used for university entry purposes.