Diagnosing writing ability using China's Standards of English Language Ability: Application of cognitive diagnosis models

Despite the growing interest in diagnostic assessment, few studies have applied cognitive diagnosis models (CDMs) to diagnosing writing ability. In this study, a diagnostic checklist was developed using the descriptors of China's Standards of English Language Ability (CSE) and diagnostic results generated by the linear logistic model (LLM) were examined. Based on the descriptive parameters of CSE scales, four attributes were targeted, including grammatical knowledge, textual knowledge, functional knowledge, and sociolinguistic knowledge. The relationship between the descriptors and attributes was specified in a Q-matrix by four domain experts. Writing responses from 460 undergraduates were coded by 15 raters using the diagnostic checklist. Item parameters, attribute prevalence, and common attribute mastery profiles were analyzed. Results showed that examinees' strengths and weaknesses were identified, and the diagnostic results differed across low and high proficiency levels. The practical implications and challenges of using the CDM-based checklist approach were further discussed.