Would you believe that one of the most difficult tasks in translation is the creation of customer surveys?

Customer surveys cannot simply be translated into another language – even if they are understood. Why? Because in addition to understanding the meaning of the survey questions in the same way as a person from the source language does, the target language respondent also needs to be able to answer in a way similar to a respondent from the source language.

This means that translating international surveys goes beyond methodology and the translation of definitions and key concepts. Surveys need to be designed in such a way that the samples, question interpretation, and analysis of answers result in equivalent data sets.

In mathematics, you would say that before you can add fractions you must have common denominators, or prior to adding weights you need common units of measure. To understand the complexity of survey translation, consider the following:

  • When conducting a survey about the purchase of dinner foods, who do you interview? In America, the purchase of the food could be made by a man, woman, husband, wife, anybody who wanted to buy food. This would not be the case in many cultures where shopping is almost always done by the wife or mother. If not properly addressed, the questions will be read with a cultural bias that is not present in the source culture. For example, when the husband responds “no” to a question, he may not be saying that he does not like the food, but only that he does not purchase it.
  • A business survey referencing “feedback” will have very different cultural implications. While most business people around the world understand the concept of feedback, their beliefs surrounding feedback are quite different. For example, American managers typically encourage with their feedback. They want to inspire. French managers tend to be harsh and pointed with their feedback. A French employee is going to read into questions about feedback from a very different perspective than an American employee would. Many Asian cultures provide feedback through indirect or even anonymous means.
  • In a paper published by Adler, Campbell and Laurent (1989), they describe a survey failure trying to “reliably describe management behavior in China.” They ran into trouble when they realized that the Chinese understanding of “truth” in business was in stark contrast to the intent of the survey. The Western ideal of truth simply does not translate into a Chinese business culture steeped in Confucian philosophy.

The language of translation must also match the culture in communication style. In one report, surveys were being sent to Japan, China, and Korea by mail. After the mailing list was prepared, they encountered a problem.

The letters were addressed in Roman script, but it was discovered that the Chinese post would only deliver mail to addresses written in Chinese characters. This threw off the entire timing of the survey as they had to find a translator to readdress all of the envelopes.

When considering the translation of a survey, it is also important to create a baseline “pre-study” to judge the question’s effectiveness in various languages. Studies have shown that people will answer the exact same question differently depending on the language they are answering the survey in.

This shows that people think differently depending on the language they are using and the cultural norms that are associated with the language. By administering pre-tests, you can adjust the questions until you find a phrasing that generates equality.

Further complicating the survey issue is the format of the survey itself. In the United States, people are used to taking surveys with answers such as “None of the Above”, “All of the Above”, “Both A and C”, and so forth. Other cultures do not have questions like this in any other area of their life so the format of the survey is completely foreign.

Despite an accurate translation, the respondents are being asked to think in a way they simply don’t know how to think. This is especially true with questions involving negation such as “None of the above” or “Which one of these is not /does not….”

Studies have also shown that the formatting of the survey makes a major difference as does the length. This creates issues when translating surveys from English to Japanese, for example. The Japanese text is often greatly expanded, leading to a much longer survey. Longer surveys are answered less frequently and less carefully than shorter surveys, on average.

In short, when creating a survey for translation, there are a number of factors to consider. During the survey design stage, it would be wise to have an established relationship with a translation service agency whose translators are intimately familiar with the target language and culture. It is also recommended that you run pre-tests until you find the right language, format, and style.

This post is written by Robert Stitt, a content writer with Ulatus.