| Approach |
Advantages |
Constraints |
| Ask the same questions and translate (ASQT) |
If successful, questions and item scales can be compared one-by-one across data sets, and thus permit the most sophisticated analyses based on what is sometimes called full scalar equivalence (see [zotpressInText item="{2265844:BGCZP7E8}" format="%a% (%d%)"]). |
Developing for an ASQT questionnaire may result in reduced specificity of questions used and a resultant loss of saliency and fine-grained information. |
|
ASQT is the least complicated approach to organize and implement. This is not to suggest it does not involve considerable effort, as reflected in this chapter and Translation: Overview. |
Conceptual coverage for all or some of the populations in the study may thus be reduced and not comparable across populations. |
|
Researchers engaged in it and clients requesting or expecting it may feel more like they are on familiar territory with this model than with others. |
At the translation stage, those implementing the ASQT may have inappropriate goals for the translation product and produce poor target language versions. |
|
ASQT potentially permits replication of existing questions—provided their basic suitability for translation and fielding with the target populations is ensured. |
Replicated questions encourage close translation and may not be optimal for one or more target populations. |
|
|
ASQT does not work well at all for some kinds of questions (e.g., background variables such as education). |
|
|
ASQT and adapt approaches call for expertise in question development and translation in areas still requiring basic research and/or training. |
| Decentering |
Allows two questionnaires to be developed in collaboration and creates the potential for full scalar equivalence (see [zotpressInText item="{2265844:BGCZP7E8}" format="%a% (%d%)"]). |
May result in questions with low saliency for either culture, since anything that constitutes a problem in the course of development is removed or altered. This would be an obstacle to full scalar equivalence. |
|
Avoids the situation where the needs of one questionnaire and language/culture dominate. |
Decentering is not viable for projects involving more than a handful of languages. |
|
Can be useful in developing comparable questions for very disparate cultures. |
Decentering is very work-intensive, and there is little information about recent experiences using this technique. |
| Ask different questions (ADQ) approaches |
Avoids the need to base questionnaires for various cultures and languages on translation of a source questionnaire. |
Little detailed information is available about recent projects adopting an ADQ approach. Researchers have few guidelines about how to develop the quality assurance and quality control steps needed. |
|
Researchers can select the indicators and questions considered most salient for a given population, provided these produce data which can still be compared across populations. |
If different populations are only asked questions developed for them, item-by-item analyses across populations are more difficult to justify. |
|
It is easier for a group joining an ADQ-based study after other groups have developed and fielded their questionnaires to produce a suitable questionnaire for their context than it is for researchers joining an ASQT project after the source questionnaire has been finalized. |
Most researchers and clients are unfamiliar with ADQ approaches. |
| Mixed approaches combining ASQT and ADQ components |
These can combine the advantages of ASQT and ADQ. |
They increase the number and kind of procedural steps to be implemented and assessed. |
|
|
They call for expertise in areas still requiring basic methodological research. |
