Abstract
Lexical richness, a common analytical measure of textual analysis in EFL writing research, is measured by three indices: lexical density, sophistication, and variation. Researchers have developed variant software such as AntWordProfiler to facilitate corpus-based analysis of lexical richness in writing. However, few studies have explored the lexical richness of middle school students’ written English in China. The pre-stored corpus of word levels fails to accurately represent the lexical richness in the writings because the leveled word list is much larger than the students’ acquired vocabulary. This study examines 239 articles written by middle school students in Zhejiang Province using AntWordProfiler, employing two different sets of BASEWORD lists: the pre-stored BASEWORD lists and the rearranged BASEWORD lists in accordance with the English Curriculum Standards for Compulsory Education (2022 Edition). The pre-stored BASEWORD lists within AntWordProfiler were used to assess the vocabulary richness of the students’ articles; subsequently, rearranged BASEWORD lists were applied by AntWordProfiler. The results are presented in terms of lexical variation, lexical sophistication, lexical errors, and lexical density. A comprehensive analysis was conducted using these categories. The research found that homework writings were richer than exam writings, exhibiting higher lexical sophistication and fewer errors. However, they were similar in lexical density. The differences were particularly notable and may have implications for vocabulary teaching in middle schools.