Translation vs. Meaning

The picture (above) is from a very well-attended poster presentation at the 2024 American Public Health Association's annual conference (#APHA2024).

Questionnaires used in clinical practice and research are regularly translated into multiple languages. These translations often assume that the meaning of phrases and words is preserved by identifying similar words.

Yet intuitively, we know that meaning is not always retained (for example, the phrase "significant other" can take very different meanings across languages and cultures), even when the words are translated correctly. We can also observe the potential influence of this problem when factor-based assessments of validity suggest different factors depending on the language in which a scale is administered.

So, how can we better approach translating questionnaires?

Dr. Tapati Dutta PhD, MCHES, MPS, MA, PGDHRM, Certified Public Health Storyteller, TEDx speaker proposed an innovative process that involves focus groups of bilingual speakers collaboratively discussing the meaning of questionnaire items to achieve translation that preserves the meaning of each item. In this case, we translated the Multidimensional Scale of Perceived Social Support (MSPSS) and we were fortunate to have the support and participation of Dr. Gregory Zimet who originally published the scale in 1988.

Our translation into Diné Bizaad (Navajo) was made possible by the enthusiastic and instrumental help of Camille Keith and other members of the Navajo community. The next steps will involve conducting standard quantitative validity and reliability assessments and comparing those data and outcomes with English-language findings as well as (hopefully) a direct (non-harmonized) translation of the instrument to assess the measured difference(s) when administering the collaboratively translated tool.

Our paper that includes substantially more information about this process is in the final stages of review and publication is anticipated within the next several months.

_________________________________________________________________________________________________

PS: Interestingly, the issue of the ambiguous meaning of words also arises in natural language processing (NLP) work, even within the same language, in cases where meaning is determined not only by textual context but also by meta-context (e.g., who might be the expected recipient of the words). See a few examples here and here.