EWCCE-DATA: English Writing Corpus Collection and Evaluation

Description

The dataset used in this study comprised 200 handwritten English writing examination scripts collected from undergraduate non-native English-speaking students who completed an English writing proficiency test. Each writing sample was digitized and stored as a PDF file containing a scanned image of the student's original handwritten essay, thereby preserving the authenticity of the writing, including grammar, vocabulary, organization, and handwriting characteristics. To establish a reliable ground truth, all essays were independently assessed by two certified English language instructors using the International English Language Testing System (IELTS) Writing Band Descriptors. The evaluation followed the four official IELTS writing criteria: Task Achievement (Task Response), Coherence and Cohesion, Lexical Resource, and Grammatical Range and Accuracy, with each criterion assigned a score according to the standard IELTS band scale. The four criterion scores were then aggregated to calculate the overall IELTS band score for each essay. Subsequently, the overall band scores were mapped to the corresponding Common European Framework of Reference for Languages (CEFR) proficiency levels (A1–C2), providing standardized proficiency labels for comparative analysis. Consequently, the final dataset includes the original handwritten essays in PDF format, individual IELTS criterion scores, the overall IELTS band score, and the corresponding CEFR proficiency level. This well-annotated dataset provides a reliable benchmark for evaluating and comparing the writing assessment performance of human raters and artificial intelligence systems.

Institutions

• Tishk International University

  Erbil, Erbil

Categories

Licence