Replicate the study on your own Prolific account

This guide is for an independent validator who wants to re-run the exact study with their own participants, paid from their own Prolific account, into their own isolated data. You get the same instrument we ran, your own private results, and a clean export you can analyze yourself. Nothing you collect touches our data, and nothing we hold touches yours.

You will need a replication slug and access token from the RTHG team. Ask us and we will issue them. Everything below assumes you have both.

1. Confirm your private workspace

1. Go to /replicate on this site.

2. Enter your slug and token. You land on your console at /replicate/<your-slug>.

3. The console shows your participant entry link:

   https://rthg-study.org/study?org=<your-slug>&t=<your-token>
   

   Every response from that link is written to your own partition. You can see counts, per-wave headlines, and exports here, scoped to only your data.

2. Test it before you spend money

1. On the console, open the entry link with test mode (the "Open in test mode" button, which adds &testMode=1).
2. Click through to the end. Test responses are marked and excluded from real counts.
3. Refresh the console and confirm your test response appears, then you are ready to field.

3. Create the study in your Prolific account

1. In Prolific, create a new study and choose "I'll use an external study link."

2. Set the study URL to your entry link, and let Prolific add the participant ID. The simplest reliable setup:

   https://rthg-study.org/study?org=<your-slug>&t=<your-token>&source=prolific&PROLIFIC_PID={{%PROLIFIC_PID%}}
   

   Prolific substitutes {{%PROLIFIC_PID%}} automatically, so each response carries that participant's ID for your own dedup and approvals.

3. How participants finish: use the completion-code method. Because the entry link is shared across your participants, choose Prolific's "I'll redirect them using a completion code" option, and Prolific gives you a code. Your participants reach the debrief page, then return to Prolific and paste that code. (If you would rather use Prolific's redirect-URL method instead, tell us and we will enable a per-validator return URL on your link; the completion-code method needs no extra setup.)

4. Eligibility: set Prolific's screeners to U.S. nationality/residence, age 18+. The instrument also re-asks citizenship and will hard-stop a non-citizen, matching our own exclusion rule, so non-citizens never enter the official counts.

5. Reward: field a small pilot first (10 to 30 people) and check the median completion time on your console before scaling, so you can set fair pay against Prolific's floor.

4. Field, monitor, and approve

1. Launch to your pilot. Watch your console update as responses land.
2. Approve or reject submissions in Prolific using the participant IDs (visible on your rows).
3. When the pilot looks clean, scale to your target N.

5. Get your data out

On your console:

• Export JSONL gives you the de-identified row-level dataset for your partition.
• Export codebook gives you the machine-readable instrument (item ids, prompts, options, scale labels) so your analyst knows exactly what each field means.
• Run the published analysis exactly as we do: the same reproduce step recomputes the headline numbers from your JSONL, so your result is produced by the same code, not by our word for it.

What stays separate

• Your participants, your spend, your Prolific account, your data partition.
• Your responses are not added to our public counter unless appropriate, and never merged into our dataset.
• We cannot see participant identities in your export; it is de-identified by design.

If anything in your console looks off (a response missing, a token rejected), stop and contact us before spending more.