% correct per question
Mean time per question (s)
Per question
| Q# | Type | Bloom | Question | % correct | Disc. | Mean time (s) | Mean rereads | Answers A–D |
|---|
Group overview
Score distribution
Time per question vs score
Off-screen gaze vs score (webcam quality check)
Mean rereads per student
Students (click a header to sort, click a row for details)
Reading metrics
Open answers
| Student | Answer | Grade | Graded by | Note |
|---|---|---|---|---|
| Loading… | ||||
Attitudes & system perception
1 = strongly disagree · 5 = strongly agree
Perceived helpfulness vs. score
English comfort vs. score
Language-vs-concept lens.
Background & study situation
What would improve these exercises? (optional)
Local LLM analysis
Checking LLM status…
Analyzing — a small local LLM may take 1–3 minutes on CPU…
Past reports
Export
Download section JSON Research CSV (one row per student × question) Survey CSV Longitudinal CSV (all your sections) Grading fine-tune JSONL Download fine-tune JSONL
The research CSV is a flat table — one row per student × question with answers, grades, Bloom level and gaze metrics — ready for analysis in R or pandas.
The section JSON is a full portable dump (quiz, students, answers, events, gaze).
The entire database is the single file data/eyerec.db — copy that file
to move or back up everything.
The fine-tune export contains every saved (features → report) pair as chat-format
JSONL, ready for Unsloth fine-tuning. Run at least one AI analysis first.