jsPsych / Behavioral experimentsRead · predict · build · inspect

Assessment and submission guide

All graded work is individual. In-class pair work and discussion are encouraged. Acknowledge borrowed code, collaboration, and substantial AI assistance. You should be able to explain and modify your submission. No marks are awarded for obtaining a significant effect.

Component Course points Default deadline Brief
Coding assignment 1 10 Before Meeting 2 First task
Coding assignment 2 10 Before Meeting 3 Conditions and data
Coding assignment 3 10 Before Meeting 4 Practice and control flow
Project proposal 5 Before Meeting 4 Proposal submission
Project pilot 5 Before Meeting 5 Pilot submission
Final project 60 14 days after Meeting 5, 23:59 Europe/Warsaw Final project
Total 100

The instructor translates these into calendar dates and posts upload links before the course begins. The three exercises collectively retain the original 30% weight; proposal/pilot retain 10%; the project retains 60%.

Grade bands apply to the unrounded total: below 50 = 2 (fail); 50 to below 60 = 3; 60 to below 70 = 3+; 70 to below 80 = 4; 80 to below 90 = 4+; 90 to 100 = 5. Attendance, reassessment, and integrity rules are in the syllabus. Do not interpret these task sheets as replacing binding programme rules.

Every coding submission

Submit an archive with runnable source, all required assets/dependencies or a complete course-relative folder structure, a short README with launch instructions, and a small synthetic/self-pilot data export. Use a student identifier only in the learning platform’s submission record or archive name, never in participant-facing experiment code. Do not submit .git, passwords, API keys, browser profiles, or other people’s raw research data.

Include a 100-200 word note: what you changed from the starter, one test you performed, one limitation, and attribution/AI assistance. “No external assistance” is acceptable. Templates are in templates.

How marks are awarded

Each rubric names observable evidence and its maximum points. Award full credit for complete correct evidence, about half for a recognizably implemented but incomplete criterion, and zero for absent or nonfunctional evidence; use intermediate points with a specific reason. Assess the submitted version rather than a verbal promise of a fix. Give one actionable next step per assignment. A single bug should not be deducted repeatedly unless it independently undermines multiple criteria (for example, wrong scoring and no pilot evidence).

Students can use the published rubric as a self-check.