jsPsych / Behavioral experimentsRead · predict · build · inspect

Meeting 3: a local JATOS study, export, and import

Meetings 1-2 use ordinary local files. Now reuse the working Meeting 2 Stroop task to learn the platform without debugging a new experiment at the same time. Allow 20-30 minutes for preparation before class and 30 minutes for the guided activity. No Docker, API credentials, or server administration is required.

Before class: start JATOS locally

  1. Download the instructor-tested JATOS release for your OS. This edition is prepared against 3.11.3. Prefer the bundle with Java; choose the matching Mac architecture where applicable. Official releases.
  2. Extract it to a normal local folder, outside Nextcloud, OneDrive, Dropbox, or another synchronized folder. JATOS includes a running database; do not put that database in a sync directory.
  3. Windows: open loader.bat. macOS/Linux: open a terminal in the extracted folder and run ./loader.sh start; if needed, first run chmod u+x loader.sh.
  4. Visit http://localhost:9000/. The fresh-install login is admin / admin; change the default password in your local installation. Keep this as a local development instance, not a public server.
  5. Confirm that you see the dashboard, then stop there. On Windows, close the loader command window to stop it; on macOS/Linux use ./loader.sh stop.

Follow the official installation instructions for OS permission prompts or Java issues. If setup fails, record the message and use a partner’s or the instructor’s demonstration; do not spend the coding lesson repairing an operating-system setup.

In class: import, run, inspect (0-15 minutes)

  1. Start JATOS and check that the address bar says localhost:9000.
  2. In the Studies sidebar, choose + → Import study. Select the working Stroop study directly. Do not unzip it first.
  3. Open the imported study and use Run. Complete the eight-item task.
  4. Open Results, inspect that run, and export/download its result data. Find eight Stroop rows, four per condition, and the final summary row. One complete task has 18 jsPsych rows: instructions + eight fixations + eight decisions + summary.
  5. Locate the study’s asset folder from the local dashboard. Open app-jatos.js: the experiment is familiar. The platform-specific changes are jatos.onLoad at startup and jatos.startNextComponent(...) in on_finish.

The GUI workflow is described in JATOS Get started. Use normal Run rather than Preview for the data-saving check.

Export a study and move it (15-25 minutes)

  1. Give your local study a recognizable classroom title; avoid modifying a study currently collecting research data.
  2. Choose Export in the study toolbar and save the resulting .jzip.
  3. Import that archive into a different local instance (a partner’s is sufficient), or the instructor’s classroom server using your assigned account. Check the destination URL before importing. If an import reports an existing study, do not overwrite someone else’s work; use a fresh destination or the instructor’s prepared copy.
  4. Run the imported study and retrieve one new pilot result there. Confirm that the stimuli, mapping, and item counts match the original.

The archive transfers study configuration and assets, not previous results. The destination run creates new results. The participant link belongs to the destination instance; your localhost link will not work on someone else’s computer. Official transfer workflow.

Name the three different files (25-30 minutes)

File What it is for Does it contain participant results?
Course source .zip Editable HTML/JS/CSS/assets; unzip to work on files No
JATOS study .jzip Import a configured study into JATOS No
Results export Inspect/analyze the recorded runs Yes, for the runs exported

Do not rename a source ZIP to .jzip; the latter includes study metadata. In this course, the instructor’s packaging script creates both formats, and students use JATOS’s own Export/Import controls.

Exit check and fallback

Show where your study lives, where its results live, and which file would move the study elsewhere. If you used a partner/demo because installation failed, complete the same observations and export exercise together; the technical setup issue does not lower the coding-assignment mark.

Cognition.run is the alternative platform: open its prepared task in the browser, run it, and retrieve results. It does not require a local JATOS installation and does not use JATOS .jzip imports. Follow the platform guide. Meeting 5 revisits these steps with your project.