Clourite: Keystroke Log Data from 309 College Students

A sample of 309 college students provided keystroke log data through the cloud-based writing platform Clourite. The data was used to extract writing process features via a Bayesian hierarchical mixture model and to identify patterns differentiating stronger and weaker writers. The dataset, authored by Tingxuan Li, was last updated on 2026-05-28.

Use Cases

• Modeling typing pause patterns based on keystroke log data
• Investigating associations between writing process features and human-assessed writing quality
• Comparing the efficiency of hierarchical versus non-hierarchical mixture models for behavioral data
• Developing features for automated writing evaluation systems based on process metrics

Strengths

• Data was collected from 309 college students, providing a substantive sample size.
• The hierarchical Bayesian model described can borrow strength across data units to improve estimation for shorter essays.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is small in scale at 865.2 KB, indicating limited scope.

Provenance

Source

Tingxuan Li via figshare

Collection Method

Empirical keystroke log data collected through the Clourite cloud-based writing platform.

Freshness

Last updated 2026-05-28 11:01:38; freshness should be verified.