The survey tool checklist for panel studies.

1. Automatic ID capture. The tool must read PROLIFIC_PID and study or session IDs from the link and save them with each response. Asking people to type an ID introduces errors that can make matching responses harder.

2. Completion redirect + visible code. Finish → bounce back to Prolific automatically, with the code printed on the thank-you screen as the manual fallback.

3. A separate screen-out exit. Different message, different code, different redirect for disqualified participants.

4. Duplicate blocking by participant ID. Check that the server blocks a second submission with the same participant ID, including attempts from another device.

5. Data-quality instrumentation. Look for attention checks, response timing, repeated-answer detection and exports that retain those flags.

6. Quotas and screening. Cap segments (e.g. 50 per condition per gender) and disqualify early, enforced server-side so concurrent participants cannot oversubscribe a cell.

7. A test mode. You will rehearse the study several times. Those rehearsals should never sit in your data.

The table compares the tools against that checklist, including features that need manual setup.

When choosing: Qualtrics is the incumbent for a reason — it can do nearly everything on the list if you invest in learning its flow editor and embedded-data plumbing, and if your institution pays for it, that knowledge transfers across labs. Typeform optimizes for marketing forms, not research; the gaps are structural. Google Forms is free and fine until the first time you lose an hour matching mistyped IDs. Plumeform's bet is that the whole checklist should be product defaults rather than configuration — judge that claim on the academic research page.

Whatever you choose, run this gauntlet with a real pilot before the study that matters: Can you launch, complete, and approve a submission end to end without touching a spreadsheet? Does a screened-out participant exit cleanly? If you open the study link twice as the same participant, what happens? Where do your dry runs go? And — the one people forget — can you export every column you will need for your methods section: durations, quality flags, IDs, condition assignments?

Run the pilot before opening recruitment, while you can still fix the setup without affecting participants.