Translate research questionnaires with AI
Keep question IDs, response options and branching intact across language versions. Prepare blinded review files and connect returned feedback to the right draft.
Translate research questionnaires with AI when a fluent sentence still needs to function as part of a form. Changing the order of two answers can send a respondent to the wrong follow-up question. Reviewers need to see the wording in context while the study team keeps track of which draft they read.
The research-translation skill separates the authoritative source, candidate translations, back-translations and returned comments. It can also organise material already translated by people. Internal provenance records stay separate from the package prepared for a blinded review.
What it does
- Preserve response options and branching instructions
- Trace each translated item to the source version
- Keep a back-translation beside the candidate it checks
- Prepare a reviewer package without exposing candidate origins
- Collate returned feedback by item and version
Install and use
- Install the complete skill folder in a compatible agent and prepare an authorised workspace. The local release checker uses Python. Record consent before exposing study material to an agent or external provider.
- Ask, “Prepare these two translations for blind review. Keep question IDs and answer order, and leave the source receipts in our internal workspace.”
- Inspect the actual archive before sending it. Have a qualified human reviewer assess meaning and suitability for the intended participants.
FAQ
Source
Made by Dominik Lukeš. Original skill: research-translation, published on GitHub under the MIT licence.
View on GitHub