Running Large Studies

Pairit handles about 150 participants active at the same time. Total participants per study are not limited. What counts is how many people are mid-experiment at once.

Above that, pages may load slowly. Matchmaking and chat are the most sensitive to it. If you plan a study with more concurrent participants, contact the Pairit team before you launch.

Limit how many people arrive at once

• On Prolific, release places in batches instead of all at once. For example, release 50, then the next 50 when most of the first batch have finished.
• Avoid sending one link to a large class or panel at a set start time. Everyone starts together, and that is the heaviest load.

Matchmaking studies

• Set timeoutSeconds to how quickly participants actually arrive. If you release 20 places for a 3-person group, people may wait several minutes for the third member.
• Always set a timeoutTarget, so participants who are not matched still get a solo page or an exit survey.
• Pilot with a small batch, then check how long people waited using wait_duration_seconds on the onMatchFound and onTimeout events.

Tell the Pairit team when a study is live

Updates to Pairit can briefly interrupt matchmaking and chat. Let the team know when a multi-participant study (matchmaking, chat, or workspace) is running, so updates wait until it ends.