Graveyard Shift Digital

Home › Free guides

MTurk Shut Down Sept 30: Alternatives for Researchers (2026)

Yes — Amazon killed Mechanical Turk on September 30, 2026. If your lab recruited participants on MTurk, your subject pool is gone, your in-flight studies got cut off, and your IRB protocol names a platform that no longer exists. Here's the honest breakdown of where researchers are actually going, what each alternative costs, and what to do about the paperwork.

1. Prolific — the default replacement

Prolific is where most MTurk researchers land first, and for good reason. Participants are pre-verified, the platform enforces minimum pay rates (currently around $12/hour), and data quality is consistently better than late-era MTurk — which matters, because an EPFL study found 33–46% of MTurk submissions were AI-generated. The trade: it costs more per participant than MTurk did, and the participant pool skews toward the US, UK, and Europe. Best for: standard surveys, behavioral experiments, anything where data quality beats rock-bottom cost.

2. CloudResearch — the MTurk power-user's pick

CloudResearch (formerly TurkPrime) was built by researchers who lived on MTurk. It offers vetted participant pools, longitudinal study support, and tools MTurk never had — like approved-participant lists and serious fraud filtering. It sits between MTurk and Prolific on cost. Best for: researchers running repeated waves, diary studies, or anyone who misses MTurk's flexibility but not its bots.

3. Connect (by CloudResearch) — the budget option

Connect is CloudResearch's lower-cost panel — closer to old MTurk pricing, with basic screening built in. Data quality won't match Prolific, but it beats unfiltered crowdsourcing. Best for: pilot studies, large-N designs where cost per head dominates, and exploratory work you'll replicate on a cleaner sample later.

4. Your university SONA pool — the free option you're ignoring

If you're at a university, the psychology subject pool still exists and still works. It's undergrads, so generalizability is limited — but it's free, IRB-friendly, and fast for simple designs. Best for: pilot data, student projects, and anything where a convenience sample is defensible.

5. Custom panels — the nuclear option

For hard-to-reach populations (clinicians, specific professions, rare demographics), none of the above will cut it. Custom recruitment through professional lists, panels like Qualtrics Panels, or direct outreach costs more and takes longer — but it's the only honest answer for some designs. Best for: when your sample is the study.

What to do about your IRB protocol

Your protocol names MTurk. Since September 30, that platform doesn't exist, and continuing to run under a stale protocol is a compliance problem. You need an amendment: new platform name, new recruitment flow, new screening and payment procedures. The amendment paragraph should cover: the platform, how participants are recruited and verified, how you screen for bots and inattention, and how participants are paid.

The honest comparison

No alternative is MTurk with a new logo. Prolific costs more but the data is cleaner. CloudResearch gives you the most MTurk-like control. Connect is cheapest. Your SONA pool is free but narrow. Pick based on what your design actually needs — the shutdown happened, and the question now is where your next wave runs.

Don't have time to migrate it yourself?

The MTurk Rescue service migrates your study in 48 hours: rebuilt on Prolific (or the alternative that fits your design), bot-fraud screening wired in from day one, plus the rewritten methods paragraph and the IRB protocol-amendment text. $249 flat per study — and if we can't migrate it, you don't pay.

Rescue my study →