ASSISTments 2009-2010 Data

Source

• Dataset metadata snapshot: source.md
• Metadata JSON: metadata.json
• Canonical external source: https://sites.google.com/site/assistmentsdata/home/2009-2010-assistment-data

Action-Time-Series Notes

• Dataset type: intelligent tutoring/student interaction sequences.
• Temporal structure: student problem-solving interactions over school time.
• Actions or interventions: attempts, hints, first action type, problem assignments, and tutor interaction features depending on release.
• Suitability: useful for student knowledge-state dynamics; action granularity varies by release.

Foundation TSFM Relevance

Agenda slot	Verdict	Evidence	Missing pieces
Latent-state prediction	adjacent	Student knowledge state is the natural hidden variable behind the problem-solving event stream.	Needs richer observed state and benchmark targets beyond tutoring-specific response prediction.
Causal structure, counterfactuals, and control	adjacent	Attempts, hints, first action type, and problem assignments expose action-like tutoring events.	Needs typed intervention semantics, randomized policy traces, and counterfactual outcome labels.
Time representation and irregular event streams	partially closes	The dataset is a timestamped student-interaction stream rather than a fixed-rate sensor panel.	Needs a standard event-stream representation aligned with numeric TSFM inputs.

Links Into The Wiki

• Foundation Time-Series Model Research Agenda
• Action-Conditioned Time-Series Datasets
• World Models