Why adequate training of linguists is a key component of translation quality in survey translation and adaptation

Why adequate training of linguists is a key component of translation quality in survey translation and adaptation

Published in: Non classé

by Steve Dept, cApStAn CEO

All language service providers will claim that they only work with experienced translators, which is mostly true, but this does not per se ensure translation quality. Serious organisations – and there are many – will invest time in preparing their survey translation and adaptation thoroughly: this may include carrying out a translatability assessment, drawing up robust question-by-question translation and adaptation guidelines, compiling bilingual glossaries and style guides for each target language. Translation agencies will guarantee that the review of the translation is carried out by a knowledgeable expert or survey methodologist, and that may often be the case. Still, when setting up a survey translation and adaptation workflow, one needs to factor in adequate training of the linguists involved.

Why is training a linguist so important?

We have translated, adapted and verified survey questionnaires in over 80 languages for 20 years now, and experience has shown that training of linguists is critical to success in survey translation and adaptation. This is explicitly mentioned in the Translation chapter of the (part of the Comparative Survey Design and Implementation Guidelines Initiative): “Briefing translators helps them to read, understand, and translate questionnaires as instruments of measurement. Translators need to be able to recognize the design features and various components of surveys in order to handle them appropriately.

This makes a lot of sense. Survey translation is a very specific task. If a given term is used three times, say once in a question stem and twice in response options, the same term should also be used three times in the translated version of the questionnaire. However, translators are educated people who have expert knowledge of their native tongue. As a consequence, they may like to show off their knowledge and avoid the clumsy repetition they see in the source version: they might proudly offer three synonyms in the target language instead of echoing the literal match. The translation will be artfully crafted, but replacing a verbatim match with a synonymous match alters the properties of the question: it may elicit different response strategies and affect the response patterns. This needs to be explained clearly to questionnaire translators – together with many other survey-specific aspects of the translation and adaptation work.

Best approaches to training linguists

While face-to-face remains the best approach, a lot can be covered in a well-designed training webinar. The training of linguists should broach key theoretical aspects of survey translation and adaptation; there should be some examples of challenges they are likely to encounter in the questionnaire they are about to translate – ask the linguists to resolve the challenge, make the webinar interactive. The training should include a hands-on exercise. If one wants this to be effective, the actual exercise should be sent a couple of days before the training webinar. The ideal number of participants is 10-15. If there are more than 15, split the webinar, propose two sessions rather than one. It is important, of course, to record the entire training session and to make the link available immediately after the training, so that the translators can watch it again when they actually begin the work.

In survey translation and adaption, what the training of linguists adds to your budget and your timeline is negligible compared to the added value in terms of translation quality.

If you’d like to find out more about how to train translation teams before they begin translating and adapting questionnaires, if you want to know more about our training modules and formats, do e‑mail us your queries at hermes@capstan.be and we’ll get back to you as soon as we can.