Language choice and AI translation are different features.
The language settings contain 140 language and regional-variant options. AI translation is available for the supported options marked in that screen. You can also choose an option without AI translation and write that language version yourself.
Review the translated draft before publishing. Local partners should check that a term, example or response scale means the same thing to the people who will answer it. Available language options are not a promise of validated translations.
Give participants the appropriate language link.
A language-specific link opens the same study in that language. The study identity stays the same, while the language code changes. Responses from the different versions enter a common table.
The dataset records the language selected when a response is submitted. Language is useful context, but it is not the same as country, nationality or recruitment source. Record other variables separately when the research requires them.
One table does not make every comparison valid.
Use the same underlying questions, options and codes across language versions. Pilot each version, especially items about local services or concepts that may not translate directly.
Differences between language groups may reflect recruitment, country, age or experience rather than the language itself. Explain those limits in the analysis and live report.
Try the language links of a real assessment.
The following links open three language versions of the same real Nuvopoll assessment. It is a separate application from The Distance Between Us research study.
