Class 5 — Pilot Testing & Data Generation

Part I
No new syntax today — finish and debug your own task, agree on a shared file-naming convention as a class, then pilot a partner’s experiment and export real data.
Published

October 6, 2026

PART I · CLASS 5

There’s no new Builder mechanic or Python concept today. Instead, the whole class is about making your own task actually work end to end and proving it by running it on a real person. You’ll work through a debugging clinic covering the most common Builder/Coder pitfalls, agree with the class on a shared file-naming and participant-ID convention, then pair up to pilot each other’s tasks and sanity-check the exported CSVs — the same files Part II of the course will read directly into R.

Reading due: PsychoPy docs; Think Python ch. 8 (optional) · Due today: HW2 due

What you’ll be able to do

  1. Diagnose and fix the most common Builder/Coder pitfalls that surface only when running a full experiment end to end (missing files, unwired routines, code-component tab errors, condition-file mismatches).
  2. Run your own PsychoPy task as the experimenter for a real (peer) participant, and sit as a participant for a partner’s task.
  3. Export and locate a clean CSV of pilot data, and briefly sanity-check it (right number of rows, expected columns, no obviously broken values).
  4. Apply a shared, agreed-upon file-naming and participant-ID convention, and explain why consistency here matters once Class 6 reads everyone’s CSVs directly into R.
  5. Identify what still needs fixing in your own task before HW3’s final pilot submission.
Back to top