sample.csv contains one synthetic curriculum-design example used for smoke tests and demonstrations. It contains no learner records or personal data.
course_id: synthetic course identifiermodule_id: synthetic module or unit identifieroutcome: intended learning outcome textactivity: teaching or learning activity textassessment: assessment task text
A future research dataset may also include rubric criteria, discipline, programme level, reviewer IDs, expert alignment judgments, and reviewer notes.
The current code operates on three text fields: outcome, activity, and assessment. Production adapters should validate that these fields are present, non-empty, and mapped to the correct curriculum unit before analysis.
For empirical evaluation, keep the automated signals separate from expert judgments. Recommended labels include:
- detected or assigned cognitive level for each curriculum element
- whether an activity gives adequate practice for the stated outcome
- whether an assessment can demonstrate achievement of the outcome
- whether a flagged level difference is pedagogically intentional
- reviewer confidence and notes
Use multiple reviewers when possible and report inter-rater agreement.
Do not commit confidential curriculum documents, restricted institutional materials, private student work, accreditation evidence that cannot be redistributed, or licensed content that prohibits publication.
For any real study, document source, permission or license, educational level, discipline, inclusion criteria, preprocessing, annotation procedure, reviewer expertise, disagreement handling, known limitations, and permitted uses.