Cross-Cultural Survey Guidelines

Language Harmonization and Shared Languages Guidelines PDF

VI. Translation: Language Harmonization and Shared Languages

Webpage last modified: 2008-Nov-13

Introduction

In many multinational surveys, the questionnaire is developed in one language (the source text or source instrument), then translated or adapted into other languages. One of the primary goals of translation is to ask the same question, in the sense of semantic and pragmatic content, in all languages. "Ask the same question" [1] is the most commonly used approach in cross-national and cross-cultural survey research.

Some languages, such as Russian, French, and German, are spoken as a first language by populations in a number of countries. As a result, multiple countries participating in a cross-national survey may be fielding in the same language -- that is, "sharing" a language. Standard regional variations of a given language (e.g., American English, British English, and Indian English) regularly differ in a variety of ways. Often the differences are principally in vocabulary and pronunciation, but differences in syntax and other grammatical features of the language are also possible. Since there is often more than one way to translate any given survey question, different countries sharing a language may produce different translations of the same source text. While some of these differences may relate to the given regional standard (e.g., Mexican Spanish versus Castilian Spanish), others may simply reflect the fact that there is more than one way to say, and translate, the same thing. Still others may actually reflect different interpretations of the source text. The standard current view in survey research is that unnecessary differences across versions in a shared language should be kept to a minimum. In other words, questionnaires in the same language should keep the wording constant wherever this is possible. "Language harmonization" occurs when translation teams from different locations or countries sharing a language negotiate to remove or to avoid unnecessary differences. Some studies aim to produce a single language version to be used for all the locales using that language. Requiring the same wording across countries sharing a language constitutes "strict and obligatory harmonization".

However, keeping the wording the same in a literal sense does not necessarily ensure that the same question is being asked. Translation teams must avoid producing a harmonized version that is not understood, or is understood differently from what is intended, by one or more populations. While many ongoing studies aim for a degree of harmonization, sometimes local adjustments are necessary to ensure that respondents in one or more locations do, indeed, understand the questions as intended [2]. This can mean that harmonization is obligatory only to the degree to which it optimizes comparability. Some studies recommend harmonization but do not make it an obligatory requirement.

Sometimes a single language version is produced and then "localized" as necessary. In this instance, the starting point is a common translation. The translation teams in different locations identify where they need to change the common single language version and make appropriate adaptations. We are not aware of any research into localization approaches for shared languages in survey research and are necessarily tentative here about procedures, outcomes, and lessons learned. Although we cannot say definitively whether or not the localization approach can be recommended, we have included some pointers below.

Guidelines

Harmonize the wording of questionnaires in one language whenever possible.
Rationale

All else being equal, it is preferable to keep the wording constant within a language. If no policy of harmonization is followed, unnecessary differences may proliferate. Some, of these, such as differences in translating answer scales, may affect measurement.
Procedural steps
- Decide policy on harmonization (obligatory or optional, strict or optimized).
- Decide on a harmonization procedure and the tools to be used; these should include a documentation component (see Translation Tools).
- Inform all locations sharing a language of the harmonization decisions and requirements.
- Schedule and organize translations so that harmonization is possible.
- If working from a single translated questionnaire towards localized versions, prepare and distribute the single translation.
Lessons learned
- Without clearly defined protocols, local teams may be challenged to make informed decisions about harmonization and to properly record their decisions.
- The short time often available for translation may make harmonization meetings and discussion difficult.
- When new participating countries join an ongoing study, new harmonization needs may arise in previously harmonized versions of questions.
- Content management system and localization software can aid identification of text requiring harmonization and provide a documentation option for differences retained (see Translation Tools).
- Localized versions based on a common translation may reduce differences, but are not always optimal instruments. Careful testing should be done to make sure that each population understands the questions as intended.
Only keep necessary differences.
Rationale

There are often several ways to formulate a survey question, an explanation, or even instructions. Teams cooperating in a harmonizing effort must try to lay aside personal preferences. Differences maintained across questionnaires should be felt to be genuinely necessary — and, preferably, demonstrated through testing to be so.
Procedural steps
If harmonization takes place on the basis of national individual drafts:
- Organize templates to enable easy comparison of the pre-harmonization versions.
- Organize the harmonization meeting(s).
  - These can be face-to-face if feasible, perhaps piggy-backing on another shared meeting.
  - Webcasting, conference calls, or skyping may be the only affordable modes of meeting.
  - Share versions prior to the meeting and produce a central document aligning them side by side; use a format that also allows the user to easily see the source questions.
  - If possible, appoint someone to identify types of difference ahead of the meeting, either on the basis of past experience or the shared versions to be harmonized.
  - Appoint a meeting chair and decide how decisions will be made, ensuring fair representation.
If common wording in the form of a single translated version is the starting point:
- Organize templates to enable easy comparison of the suggested localizations.
- Have each local team propose necessary modifications to the common version.
- Organize the reconciliation meeting(s).
  - These can be face-to-face if feasible, perhaps piggy-backing on another shared meeting.
  - Webcasting, conference calls, or skyping may be the only affordable modes of meeting.
  - Define the goals of this meeting (e.g., to review changes, to find new shared versions, to share questions about the single translation).
  - Share localizations prior to the meeting and produce a central document aligning them side by side; use a format that also allows the user to easily see the source questions.
  - If possible appoint someone to identify types of localization ahead of the meeting, either on the basis of past experience or the localizations proposed.
  - Appoint a meeting chair and decide on how decisions will be made, ensuring fair representation.
Lessons learned
- Clear instructions and specifications are crucial. The WHO Mental Health Survey Initiative aimed for an optimized but maximally harmonized questionnaire. The output for Spanish in Latin America and Spain is presented in Appendix A.
- Personal language perception and usage can be mistaken for generic language usage. Some English speakers distinguish between "nowadays" and "these days", for example, but it would be incorrect to assume that this is a distinction made by all speakers of English.
- It may not serve the study’s purpose to work on the principle of a "majority" vote. The issue will ultimately be necessary difference in any given version.
- Harmonization is not limited to words or phrases; it can include decisions about structure of sentences, organization of answer scale categories, etc.
- Sometimes harmonization takes the form of adding a term or an example to whatever is common with other shared language versions. Thus if a question about tobacco use does not cover a special form that is only relevant for one population, mention of it could be added for that population alongside the shared other forms of tobacco use.
- If the localization model is used, teams may spend more time discussing the single version than their localizations.
- Suggested localizations may lead to new ideas about a possible common version or a common version with occasional "add-ons" as described above.
Schedule harmonization at an appropriate time.
Rationale

Harmonization efforts can result in changes in one or all questionnaires. The harmonization decisions need to be made when each questionnaire version is at an advanced stage of development. If a team translation procedure (translation review, adjudication, pretesting) is followed, harmonization should precede pretesting (see Translation). Pretesting can also be used to check harmonization decisions. Changes may still be needed in one or more versions after pretesting.
Procedural steps
- Identify the time at which a well developed version of each questionnaire to be harmonized will be pretested and arrange for harmonization before that time. This may be quite complicated.
Lessons learned
- If countries are fielding at different times, a late-fielding group may have trouble preparing its version for a harmonization meeting in time for groups fielding earlier. The sooner harmonization is organized and scheduled, the higher are the chances of schedule coordination between countries.
- Good and accessible documentation is important. Countries joining the study later benefit by being able to see what other teams working in the same language have done.
- In practice, optimized harmonization may be difficult to achieve. The ESS aimed for optimized harmonization and recommended it to participating countries. Since harmonization was a recommendation, not a requirement, countries were left with considerable freedom as to whether they harmonized or not. In addition, the time schedule did not easily accommodate a harmonization step. The situation was further complicated by countries joining the project at different times, so that the first Russian translation was produced for immigrants in Israel and the most recent for Russians in today’s Russia. It is fair to say that shared languages in the ESS have more differences than actually required or desired.

Appendix A [Under Construction]

Glossary [Under Construction]

Adjudication
Source document/source text: The original instrument from which other (target) instruments are translated or adapted as necessary.

References

[1] Harkness, J., Van de Vijver, F.J.R., & Johnson, T.P. (2003). Questionnaire Design in Comparative Research. In J. A. Harkness, F. J. R. Van de Vijver, & P. Ph. Mohler (Eds.), Cross-Cultural Survey Methods (pp. 19-34). New Jersey: Wiley.

[2] Harkness, J., Pennell, B.-E., & Schoua-Glusberg, A. (2004). Questionnaire Translation and Assessment. In S. Presser, J. Rothgeb, M. Couper, J. Lessler, J. Martin & E. Singer (Eds.), Methods for Testing and Evaluating Survey Questionnaires (pp. 453-473). New Jersey: Wiley.