Translating survey instruments

Translation is part of survey design. Each language version should ask about the same concept, population, period and response options in words people understand. Budget the work alongside questionnaire development, testing and enumerator training.

The original Indonesia materials were assembled mainly in English with some Bahasa Indonesia submissions. The historical note records that subsequent revisions left parts of the translations out of alignment. Reconcile each item against the approved source version before using those materials in fieldwork.

Define the meaning before translating

Give translators the purpose and definition of each variable, its eligible respondent, reference period, response codes and interviewer instructions. Explain terms such as household, employment, ownership and customary use. Keep the selected national survey's original wording and edition available where comparability is intended.

A literal match can still change the question. A local term for household may describe an extended family rather than the people sharing provision; a term for work may omit seasonal fishing or unpaid production. Record the intended construct and investigate these differences with speakers familiar with local practice.

Review and test each language version

Use translators who understand the language and the survey context, followed by a separate bilingual review. Resolve discrepancies with the subject specialist and document the reason for the final wording. Back-translation can help expose a difference; respondent testing establishes whether the issued question is understood as intended.

Ask pilot participants to explain what they thought the question meant, which people or period they considered, and how they chose an answer. Test instructions, response categories and sensitive terms as well as the question itself. Review the pattern of difficulties across relevant groups before release. The European Social Survey's Round 11 instructions provide a concrete model: translation, review and adjudication, followed by verification, full-questionnaire pretesting and proofreading. They retain alternative wording, adaptation decisions and test findings in the translation record. Adapt the workflow to the study's scale and languages. ESS, 2022, pp. 4–6 and 10–13

Keep versions aligned during fieldwork

Maintain one item ID across languages and a record of the source text, translated text, review decisions, test findings and approval date. A substantive change to the source question triggers review of the corresponding translations. Release forms and interviewer guidance together.

Train enumerators in the agreed terms and approved explanations. Record the interview language and any interpreter involvement. Review recurring requests for clarification through a controlled revision process so interpretation remains consistent across teams and rounds.

Continue with survey design, field form preparation and the detailed language and interpretation guidance. The methods references retain the broader reading trail.