Challenges in translating surveys

The art of translation is never more under a microscope than when attempting to convey the contents of a survey in a target language. Notwithstanding the fact that revisions, back-translations and pre-testing on a sample population throw up a Pandora’s Box of confusion and mayhem, the mere act of translating the text from one language into another is a poser in itself. Here are some of the reasons why:

1) Lack of context

Out of context, even intra-lingual translation isn’t in its natural habitat as it were. Without a background to use as reference, translating a word into another in the same language is hard at best; and when that exercise involves two languages, it is nothing less than a Freddy Krueger horror story. Imagine having to translate an answer to a question that you haven’t even seen. It’s almost like working on your car in the night with a blindfold. Accidents will happen!

For example, in Romanian, there are at least 5 words that mean “all”; so, if you have “All” as one of the answers in the survey, do you translate it as ‘tot’,’ totul’, ‘toată’, ‘toți’ or ‘toate’?

In other cases, the answer may be translated too closely, giving more importance to the words themselves rather than the associated meaning. Without context, you cannot even be sure that the translated answer corresponds to the question asked.

These and many other problems plague the translator when working out of context. Yes, it is possible to have the questions made available when translating, but not all clients are aware of how important essential this is.

2) Regional or cultural nuances

This problem is all-pervasive in the world of translation. For surveys, however, this can create a dead-end and bring the work to a complete stop. Worse yet, it could lead to horrendous results and skew the very intent of the survey. A good example would be the sharing of personal information in a survey. In some cultures, sharing personal information such as income, age and so on is considered taboo. Since surveys depend on this for demographic profiling, missing information may render a survey unusable.

In other cases, the use of slang or colloquialisms may change the entire meaning of a sentence, or make it meaningless in the target language. In such cases, unless the translator is aware of the cultural context of the answer, this can very easily happen.

3) Diacritics and homonyms

Diacritics play an integral role in European and Asian languages and must be given importance to. Since the scope of a diacritic extends to the entire word rather than the letter it is associated with, a wrongly understood or missing diacritic will likely change the meaning of the source word, not to mention the target word. For example, in Romanian, if you see ‘mana´ (assuming it may or may not have diacritics), it could mean: honeydew, hand, drive, prod….

As with the confusion around diacritics, there are also words that are written the same way, but have different meanings – or homonyms. For example, the word ‘run’ in English has no less than 645 meanings – and that’s just the verb form of the word! With such ambiguities being abundant in many languages, a misunderstood word can easily slip in to the translation of a survey, changing the entire meaning of the answer and rendering the purpose of the survey ineffective.

4) Variations in concepts

In certain cases, a particular word has no standard equivalent in another language. In fact, the same word can have slightly different meanings or refer to slightly different concepts depending on where it is being used. For example, the concept of equality/égalité is different in the U.S., English-speaking Canada and French-speaking Canada.

In addition to words, there are other concepts such as currency that carry a lot of ambiguity. For example, the currency symbol may appear before or after the number, with or without a space, with a comma or a period indicating thousands and millions and so on. These can easily pose technical difficulties when attempting to interpret or aggregate data.

Gender distinction is another area that may cause trouble. In English, there’s no male or female form associated with nouns. In many languages, there is.

Names can pose problems, too. In some cultures, the surname comes first; in others, it comes at the end; in yet others, rather than a surname, a person carries the name of his or her village of origin. Though none of these are translated, when the data is collated, it can cause a huge amount of confusion.

Conclusion

All of these challenges are present in just one phase of translating a survey – the initial translation of responses. In order to appreciate the true magnitude of error that can come into play, we also need to consider the other stages, such as formulating the questions to avoid ambiguity, being “culturally aware” of other languages that the survey may be translated into, vetting the technical aspects of the survey, back-translating to ensure that the integrity of the survey questions are preserved, and much more. In any of these stages, the slightest error can snowball into an incomprehensible set of responses.

The presumption of all surveys is that there is a measure of uniformity in the responses. When these responses themselves are subject to ambiguity of this magnitude, the integrity of the data can often come into question. And this is exactly the problem in a nutshell. With so many variations and possibilities for error, the translation of survey responses must be carried out carefully, with the appropriate context made available and with all the care of handling a fragile object that can crack on the slightest impact.

Other articles in the series:
Challenges with Translating Animal Metaphors
Challenges in Translating Legal Documentation
Challenges in Translating Literature
Challenges in Translating for the Advertising Industry
Challenges in Translating Poetry
The Challenges of Consecutive Interpreting
Challenges in Film Translation
Why Is Simultaneous Interpreting Challenging?