The Keystroke Logging in Compositions for Knowledge Evaluation (KLiCKe) Corpus

The Keystroke Logging in Compositions for Knowledge Evaluation (KLiCKe) corpus is a large-scale dataset featuring detailed keystroke logs from ~5,000 argumentative essays written by adult English writers in the United States. KLiCKe corpus was collected from January through November 2022 via crowdsourcing. The corpus records each keystroke and mouse operation, along with corresponding time stamps and cursor position information, using a web-based keystroke logging program. Data are provided in CSV format (for custom analysis) and IDFX format (compatible with Inputlog). Holistic writing quality scores for all essays are included, derived from double-blind ratings by trained human raters using a standardized grading scale. Additionally, the corpus offers demographic details on the writers, including age, gender, native language, ethnicity/race, education level, as well as typing skills and vocabulary knowledge. As a publicly available resource, KLiCKe bridges gaps in process-oriented writing research and offers new possibilities for advancing writing assessment and instruction.

The corpus is too large for GitHub. The link for the content of the corpus is below: https://drive.google.com/drive/folders/13Q02Mvj7T-T1OpZzkEqfHwqM7oI61lLR?usp=sharing

The link provides these items:

• WritingTask folder. This folder contains ~5,000 persuasive texts, the corresponding keystroke logs in both CSV and IDFX formats, and the holistic scores of these texts, as well as the writing prompts and the holistic rating rubric.
• TypingTests folder. This folder contains the results from the Inputlog copy tasks (Van Waes et al., 2019), represented in keystroke logs in both CSV and IDFX formats.
• VocabularyKnowledgeTest folder. This folder contains the results from the Lexical Test for Advanced Learners of English (LexTALE; Lemhöfer & Broersma, 2012).
• demographic_info.csv. This spreadsheet contains the writers' demographic details.

Please refer to the paper below for a detailed description of the dataset and the context of its construction:

Tian, Y., Crossley, S. A., & Van Waes, L. (2025 - accepted for publication). The KLiCKe Corpus: Keystroke Logging in Compositions for Knowledge Evaluation. Journal of Writing Research.